university of groningen the knowledge dynamics of ... · ter verkrijging van het doctoraat in de...

University of Groningen

The knowledge dynamics of organizational innovationSjarbaini, Vivyane Larissa Ratna Nirma

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2009

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Sjarbaini, V. L. R. N. (2009). The knowledge dynamics of organizational innovation: understanding theimplementation of decision support for planners. Enschede: PrintPartners Ipskamp B.V., Enschede, TheNetherlands.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 26-02-2019

https://www.rug.nl/research/portal/en/publications/the-knowledge-dynamics-of-organizational-innovation(8d3390ff-d08c-4ff8-9c6e-9f8d87302964).html

The Knowledge Dynamics of Organizational Innovation

Understanding the implementation of decision support for planners

Larissa Sjarbaini

Published by the University of Groningen, Groningen

Printed by: Ipskamp Drukkers, Enschede

ISBN 978-90-367-3714-2

Copyright © 2009 by V. L. R. N. Sjarbaini

All rights reserved. No part of this publication may be reprinted or utilized in

any form or by any electronic, mechanical or other means, now known or

hereafter invented, including photocopying and recording, or in any

information storage or retrieval system, without written permission from the

copyright owner.

The Knowledge Dynamics of Organizational InnovationUnderstanding the implementation of decision support for planners

Proefschrift

ter verkrijging van het doctoraat in de

Economie en Bedrijfskunde

aan de Rijksuniversiteit Groningen

op het gezag van de

Rector Magnificus, dr. F. Zwarts,

in het openbaar te verdedigen op

donderdag 26 maart 2009

om 14.45 uur

door

Vivyane Larissa Ratna Nirma Sjarbaini

geboren op 26 mei 1968

te Amsterdam

Promotor: Prof. dr. R.J. Jorna

Copromotor: Dr. D.J. Kiewiet

Beoordelingscommissie: Prof. dr. T. van Engers

Prof. dr. W. van Rossum

Prof. dr. J. Wilson

v

Voorwoord

Dit proefschrift is een resultaat van mijn promotieonderzoek. En ik ben er blij mee; het is het eerste boek dat ik ooit heb geschreven en dat is bijzonder. Het is een tastbaar resultaat van mijn ontwikkeling als onderzoeker. En, zoals ik in dit proefschrift betoog, een ontwikkeling wordt pas in gang gezet als je een verschilwaarneemt, een verschil tussen wat je al ‘weet’ en wat je waarneemt. Kort door de bocht, je gaat je pas ontwikkelen op het moment dat je ziet dat het ook anders kan. En zo’n moment doet zich vaak [pas] voor als je in contact komt met andere mensen. Een aantal van deze mensen wil ik hier in het bijzonder noemen. Allereerst mijn begeleiding RDJ.

Beste René, enigszins verbaasd was ik toen ik jou voor het eerst aan de telefoon had. In de vacature werd een sociale wetenschapper gevraagd. Destijds associeerde ik Bedrijfskunde meer met Economie en geld. Als psycholoog wilde ik dan ook weten of het wel zinnig was om te solliciteren op een vacature bij Bedrijfskunde. Jij zei dat ik juist kon solliciteren. Ik had het idee dat je me niet helemaal had begrepen, maar besloot toch te solliciteren. Het is grappig om hier aan terug te denken, wetende wat ik nu weet. Maar het is ook veelzeggend. Er zijn niet veel psychologen die zich buiten de Psychologie wagen. Jij hebt dat wel gedaan. Jij vindt dat de verschillende disciplines binnen de wetenschap veel van elkaar kunnen leren. Jij bent dan ook bereid om buiten de gebaande paden te treden. Dit brengt met zich mee dat je goed moet weten waar je mee bezig bent, maar dat je af en toe ook heel praktisch moet zijn en tevreden met ‘tussenresultaten’. Voortschrijdend inzicht is een term die jij vaak gebruikt. Dank voor je toewijding en enthousiasme, door jou heb ik veel geleerd.

Beste Derk Jan, jij hebt bijzondere didactische kwaliteiten. Jouw precisie en betrokkenheid breng je mee in ieder gesprek. Door jouw aansluiting kregen de besprekingen meer dynamiek. Dank voor het delen van jouw inzichten, het doelgericht meedenken en natuurlijk het schoon houden van mijn nieuwe Mac.

Dan wil ik graag nog een paar andere collega’s noemen. Barend van Heusden, een zeer creatieve en gedreven wetenschapper die mij heeft geënthousiasmeerd en op weg heeft geholpen in de eerste jaren. Luc, misschien wel mijn leukste collega! Dank voor je collegialiteit. Op afstand een fantastische ondersteuning. En verder anderen die mijn promotietijd op verschillende manieren hebben veraangenaamd: Martin, Andreas, Marjolein, Frank, Kees, Ruben, Hennie, Sonja,

vi

Ellen, Durkje, Kristian, Nanda, Marloes, Joost, Petra, Gerben, Frits, Rene, Arjan, Irene, Niels-Ingvar, Trudeke, Xander, Nurul en Filippo.

Ik wil ook de prettige samenwerking met de mensen uit het veld noemen, te weten Murk Hoekstra van [destijds] IKS en Jan Hendrik Kieft van Bartiméus.

Dan een aantal docenten van mijn opleiding Psychologie aan de Uva die mede mijn interesse voor de wetenschap hebben gewekt: Janet van Hell die mij, als mijn afstudeerbegeleider, de mogelijkheden van de wetenschap liet zien; Professor Elshout, die in zijn colleges op subtiele manier de ruimte liet aan studenten voor eigen ontdekkingen en Cees van Leeuwen, zo’n docent die mij is bijgebleven door zijn ongedwongen en positieve benadering van de wetenschap.

Ik wil ook nog een aantal mensen bedanken voor hun gezelschap, gastvrijheid en behulpzaamheid waarmee ze mij hebben geholpen in zeer praktische zin. Doppio e.v.a.z.: Ruben, 8.00? Voor mijn overnachtingen in Groningen: Andreas & Cecile, Gea & Gerben, Petra & Jan, Ronald. Voor mijn werkplek: Jet & Sander, Angélique & Maarten. Voor de secretariële ondersteuning: Durkje, Ellen, Hennie en op de valreep Anja. En in het bijzonder natuurlijk Sonja. Dank voor je soepele en effectieve ondersteuning op vooral lange afstand. Voor de sammendrag: Karin en voor het proeflezen: Dieuwertje.

Niels en Lily, wat fijn dat jullie mijn paranimfen zijn.

Verder dank ik mijn familie, Ima en in het bijzonder mijn moeder. Dat velen je kennen zegt genoeg, in Groningen, op school of bij mijn coachopleiding. Daar was je weer met een van de mannetjes, zodat ik borstvoeding kon geven. Als ‘alleskunner’ onmisbaar!

En dan natuurlijk het eggaliptische thuisfront [van Dam, 2008]. Reinier, Dex en Rafe. ‘Goninge’ was voor jullie iets mysterieus. Er gebeurde daar duidelijk iets spannends en jullie wilden daar ook wel wonen. Wat een heerlijke ventjes zijn jullie. En Peter, je bent geweldig. Onze aanpak werd, denk ik, door maar weinigen begrepen. Vele weekenden en zelfs hele vakanties naar Denemarken waarin jij in je eentje met de mannetjes op stap was. Onwaarschijnlijk. Je bent een fantastische vader en mijn man! Je hebt de benijdenswaardige gave dat jij je kunt verplaatsen in anderen zonder ervaringsdeskundige te zijn. Ik ben blij dat ik met jou oud mag worden. Dank voor je onvoorwaardelijkheid en dat je bent wie je bent. Op naar Noorwegen. Spin!

Amsterdam, 18 februari 2009

7

Table of Content

VOORWOORD..................................................................................................................... V

1 GENERAL INTRODUCTION ................................................................................15

1.1 Introduction........................................................................................................................... 15

1.2 Introduction to organizational innovation ................................................................. 15

1.2.1 A balanced view on organizational innovation ........................................................ 15

1.2.2 A typical case of organizational innovation failure................................................ 16

1.2.3 Probable causes of failure ............................................................................................... 17

1.3 Understanding organizational innovation from a knowledge perspective ... 18

1.3.1 Focus on knowledge in organizations......................................................................... 18

1.3.2 A knowledge perspective on organizational innovation...................................... 19

1.3.3 A multi-actor perspective................................................................................................. 19

1.3.4 Bridging theory and practice.......................................................................................... 20

1.3.5 The knowledge domain of planning.............................................................................. 20

1.4 General research questions ........................................................................................... 21

1.5 Outline thesis........................................................................................................................ 21

2 ORGANIZATIONAL INNOVATION....................................................................23

2.1 Introduction........................................................................................................................... 23

2.2 Defining innovation ............................................................................................................. 24

2.3 Foci in innovation research ............................................................................................. 25

2.3.1 Improvement of performance: Levels of analysis ................................................. 25

2.3.2 Types of innovation.............................................................................................................. 26

Differentiating innovations on newness: From radical to incremental................... 26

Technical versus administrative............................................................................................... 27

Product versus process............................................................................................................... 28

2.3.3 Stages in the innovation process ................................................................................. 28

The initiation stage ......................................................................................................................... 29

The implementation stage .......................................................................................................... 33

2.3.4 Related research areas.................................................................................................... 36


8

Organizational change................................................................................................................... 36

Organizational learning ................................................................................................................. 36

Absorptive capacity ........................................................................................................................ 38

2.4 A knowledge perspective on innovation..................................................................... 39

2.4.1 Introduction to the knowledge perspective.............................................................. 39

2.4.2 Knowledge dynamics as the motor of innovation ................................................. 39

Nonaka and Takeuchi: The social aspect of organizational innovation................... 39

Leonard-Barton: Knowledge as a wellspring for organizational innovation.......... 40

2.5 Characterizing the present study within the literature of organizational

innovation ........................................................................................................................................... 42

2.5.1 Improvement of individual knowledge ......................................................................... 43

2.5.2 Innovation .................................................................................................................................44

The implementation of decision support from outside .................................................. 44

Type of innovation ............................................................................................................................ 44

2.5.3 Focus on knowledge............................................................................................................ 44

3 STUDYING KNOWLEDGE DYNAMICS IN ORGANIZATIONS..............45

3.1 Introduction .......................................................................................................................... 45

3.2 General introduction to knowledge.............................................................................. 45

3.2.1 Introducing representation.............................................................................................. 46

3.2.2 Data, information, and knowledge................................................................................ 46

3.3 The I-Space: A model to understand knowledge dynamics in organizations 49

3.3.1 Introduction............................................................................................................................. 49

3.3.2 Codification.............................................................................................................................. 50

3.3.3 Abstraction.............................................................................................................................. 51

3.3.4 Diffusion ....................................................................................................................................51

3.3.5 The I-Space .............................................................................................................................. 52

3.3.6 Two ways to value knowledge......................................................................................... 54

Neoclassical or N-Learning.........................................................................................................54

Schumpeterian or S-Learning....................................................................................................55

Knowledge value of individuals...................................................................................................56

3.4 Evaluating the I-Space for studying knowledge dynamics .................................. 56

3.4.1 Recapitulating ........................................................................................................................ 56

3.4.2 Understanding knowledge dynamics during innovation using the I-Space 56

Table of Content

9

4 A COGNITIVE-SEMIOTIC MODEL TO CAPTURE KNOWLEDGE

DYNAMICS AT THE INDIVIDUAL LEVEL ..............................................................59

4.1 Introduction........................................................................................................................... 59

4.2 A cognitive-semiotic approach to knowledge........................................................... 59

4.2.1 Cognitive science ................................................................................................................. 60

4.2.2 Semiotics: A sign process of semiosis in three dimensions ........................... 62

The first semiotic dimension: Perception of difference................................................. 63

The second semiotic dimension: The need for communication................................ 65

The third semiotic dimension: Understanding the essence in relation to other

concepts .............................................................................................................................................. 67

4.3 Combining cognitive science and semiotics: Three stages in acquiring

knowledge........................................................................................................................................... 69

4.3.1 Sensory knowledge ............................................................................................................. 70

4.3.2 Coded knowledge ................................................................................................................. 72

4.3.3 Theoretical knowledge....................................................................................................... 75

4.3.4 Putting the three knowledge types into perspective........................................... 78

Polanyi’s tacit knowing versus sensory knowledge ......................................................... 78

Boisot’s codification dimension versus coded knowledge ........................................... 79

Boisot’s abstraction dimension versus theoretical knowledge ................................. 79

Schematic overview of the three knowledge types ......................................................... 80

4.4 A cognitive-semiotic model to understand the dynamics of knowledge........ 81

4.4.1 Accumulation of knowledge types................................................................................ 81

4.4.2 Knowledge change .............................................................................................................. 83

4.4.3 Knowledge conversion....................................................................................................... 84

4.4.4 Individual differences in learning................................................................................... 85

Job experience and expertise ................................................................................................... 86

Education............................................................................................................................................. 87

Age ......................................................................................................................................................... 87

Contractual hours and building routine: part time versus full time ......................... 87

4.4.5 Visualizing knowledge [dynamics] of innovation in the K-Space ..................... 88

5 PLANNING ................................................................................................................91

5.1 Introduction........................................................................................................................... 91

5.2 The knowledge domain of planning .............................................................................. 91

5.2.1 Planning in everyday life .................................................................................................... 91

5.2.2 The bridge masters example ......................................................................................... 92

5.2.3 Comparing everyday planning to organization related planning.................... 93

5.3 Planning as a problem solving task, the knowledge to make a duty roster. 94


10

5.3.1 Knowledge of the schedule itself................................................................................... 95

5.3.2 Knowledge of shift................................................................................................................ 95

5.3.3 Knowledge of personnel....................................................................................................95

5.3.4 Knowledge of constraints.................................................................................................95

5.3.5 Knowledge of goals.............................................................................................................. 97

5.4 Subtasks of planning ......................................................................................................... 97

5.4.1 Planning as a task................................................................................................................ 97

5.4.2 Gathering information........................................................................................................98

5.4.3 Scheduling ............................................................................................................................... 98

5.4.4 Negotiating.............................................................................................................................. 99

5.5 Planning with decision support software............................................................... 100

5.5.1 Computer supported tasks ..........................................................................................100

5.5.2 ZKR: planning in health care ........................................................................................101

Introduction .....................................................................................................................................101

The SEC-model ...............................................................................................................................102

5.5.3 Differences between manual planning and supported planning.................103

5.5.4 In conclusion ........................................................................................................................104

6 THE CONCEPTUAL MODEL........................................................................... 105

6.1 Introduction ....................................................................................................................... 105

6.2 Research questions........................................................................................................ 105

6.3 Integrated theoretical framework ............................................................................ 105

6.3.1 Introduction..........................................................................................................................105

6.3.2 Personal characteristics ...............................................................................................107

Education ..........................................................................................................................................107

Job experience...............................................................................................................................108

Age.......................................................................................................................................................108

Contractual hours per week....................................................................................................108

6.3.3 Subtasks ................................................................................................................................108

Gathering information ................................................................................................................108

Negotiating ......................................................................................................................................109

Scheduling........................................................................................................................................109

6.4 Conceptual model ............................................................................................................ 109

6.5 Hypotheses ........................................................................................................................ 110

7 BARTIMÉUS.......................................................................................................... 115

Table of Content

11

7.1 The organization............................................................................................................... 115

7.2 Planning at Bartiméus: A situation for innovation.............................................. 117

7.2.1 General vision on planning ............................................................................................ 117

7.2.2 Situation at Bartiméus before ZKR.......................................................................... 117

7.2.3 Planners at Bartiméus ................................................................................................... 117

7.2.4 Why innovate ...................................................................................................................... 119

7.2.5 Other people involved in planning at Bartiméus ................................................. 119

8 METHODOLOGY...................................................................................................121

8.1 Introduction........................................................................................................................ 121

8.2 Design................................................................................................................................... 121

8.3 Participants ....................................................................................................................... 123

8.4 Operationalization............................................................................................................ 123

8.4.1 Innovation.............................................................................................................................. 123

The first measurement ............................................................................................................. 124

The second measurement....................................................................................................... 124

The third measurement ............................................................................................................ 124

8.4.2 The three knowledge types .......................................................................................... 125

Sensory knowledge...................................................................................................................... 125

Coded knowledge.......................................................................................................................... 125

Theoretical knowledge ............................................................................................................... 125

8.4.3 Moderating variables: personal characteristics ................................................ 126

8.4.4 Questionnaire ..................................................................................................................... 126

8.4.5 Validity and reliability........................................................................................................ 127

8.5 Procedure ........................................................................................................................... 127

9 RESULTS AND DATA ANALYSIS.................................................................129

9.1 Introduction........................................................................................................................ 129

9.2 Planners’ profile ............................................................................................................... 129

9.3 Some preliminary data................................................................................................... 131

9.3.1 Reliability of the constructs.......................................................................................... 131

9.3.2 Means and standard deviations split on personal characteristics........... 134


12

9.4 Hypotheses testing......................................................................................................... 135

9.4.1 Hypotheses 1 ......................................................................................................................136

Hypothesis 1a ................................................................................................................................136

Hypothesis 1b ................................................................................................................................137

Hypothesis 1c.................................................................................................................................139

9.4.2 Hypothesis 2........................................................................................................................141

9.4.3 Hypotheses 3 ......................................................................................................................146

Hypothesis 3a ................................................................................................................................146

Hypothesis 3b ................................................................................................................................151

9.4.4 Hypothesis 4........................................................................................................................154

9.4.5 Hypothesis 5........................................................................................................................158

9.4.6 Hypothesis 6........................................................................................................................161

9.4.7 Preliminary conclusions .................................................................................................164

9.5 Secondary analysis ......................................................................................................... 166

9.5.1 Comparing the subtasks................................................................................................166

9.5.2 Training versus experience...........................................................................................167

9.6 Findings................................................................................................................................ 169

9.6.1 Increase of coded knowledge ......................................................................................169

9.6.2 Differences between tasks ...........................................................................................169

9.6.3 Personal characteristics ...............................................................................................170

9.6.4 Implementation stages...................................................................................................171

9.6.5 In conclusion ........................................................................................................................171

10 DISCUSSION...................................................................................................... 173

10.1 Introduction ....................................................................................................................... 173

10.2 Methodological aspects................................................................................................ 173

10.3 Theory .................................................................................................................................. 175

10.3.1 Organization......................................................................................................................175

10.3.2 Planning ..............................................................................................................................176

10.3.3 Knowledge .........................................................................................................................176

10.3.4 Innovation...........................................................................................................................178

10.4 Elaboration of thoughts ................................................................................................ 180

10.4.1 The I-Space model and te cognitive semiotic model......................................180

10.4.2 Other types of innovation............................................................................................181

REFERENCES................................................................................................................ 183

Table of Content

13

APPENDIX: QUESTIONNAIRE ................................................................................197

SUMMARY.....................................................................................................................205

SAMENVATTING..........................................................................................................208

SAMMENDRAG ...........................................................................................................211


14

15

Chapter 1

General Introduction

1.1 Introduction

Innovation is often accompanied by problems [e.g. van de Ven 1986; Leonard-Barton 1988/1995; Geerts 1999; Laudon & Laudon 2000/2002; van Stijn 2006]. We will argue that knowledge is a crucial factor in understanding organizational innovation and consequently the typical problems that are related to knowledge in one way or another.

1.2 Introduction to organizational innovation

1.2.1 A balanced view on organizational innovation

Organizational innovation processes are viewed as important [Strambach 2002; Subramanian & Nilakanta 1996; Duh, Chow, & Chen 2006]; some even say that they are crucial for the modern organization in that it enlarges the competitive advantage [e.g. Legge 1992; Wolfe 1994: 405; Nonaka & Takeuchi 1995; Leonard-Barton 1995]. Organizational innovation often involves information technology [Greif and Keller 1990], such as ERP – a standard software package to support decision making [Klaus, Rosemann, & Gable 2000] or other Management Information Systems [Laudon & Laudon 2006].

Although organizational innovation has great potential, a primary focus on the advantages nevertheless has a danger to it; other aspects to this change process and its consequences can easily be overlooked, such as making an investment in time and money, commitment and acknowledging the complexity of the process [e.g. Linton 2002]. Neglecting these aspects can and often does result in an unsuccessful innovation process as the failure rates of innovation processes show [see e.g. Laudon & Laudon 2002; van Stijn 2006]. It has been estimated that an alarming 50 to 75 percent of all implementation processes of organizational innovation, such as the implementation of software tools organization wide, result in some sort of failure [Majchrzak 1988; Geerts 1999]. Moreover, even when an innovation process is successful, this ironically does not imply that the initial expected benefits of the innovation have been met [Linton 2002].


16

To illustrate the complex nature of organizational innovation processes we discuss a case, based on an example by Laudon and Laudon [2002, 366-367], which highlights some of the typical issues and problems that are often involved. This case concerns the implementation of a decision support system [DSS] organization wide.

1.2.2 A typical case of organizational innovation failure

During the mid 90s of the last century an American hotel chain had to deal with a strong decrease in their profits. Management decided that introducing a ‘new customer recognition system’ should be able to face this problem. They expected this system to attract more revisiting customers. The system should provide all kinds of helpful information to improve customer service. For instance, revisiting guests would have advantages such as quick check-in and upgrades.

The hotel chain management took this operation seriously, so they called for a meeting with all the hotel chain managers to discuss the implementation of this customer recognition system. However, most managers sent their assistants due to lack of time and other priorities. In doing this, these managers showed that they did not top prioritize the innovation process. Then, the system was designed without input of the business units and the bonus system that was introduced for revisiting guests had no clear protocol. When the prototype of this system was demonstrated, arguing started about criteria the system had to meet. For instance, no consensus could be reached on when and how the revisiting customers should profit. Furthermore, there had been no communication between marketing and the CEO’s who had different expectations of the system to be developed. Also the communication between middle management and the board was insufficient, which was caused mainly by the high job rotation over the last years.

Pilot studies showed technical problems, which resulted in difficult and time-consuming procedures. By the time the system was ready to be introduced most parties involved felt a disliking to it, resulting in for instance managers who did not want to spend their budget on expensive presents for their customers. Furthermore, the system was slow and difficult to use. Personnel did not receive any training or even documentation. After six months the system was abandoned. The result of this organizational innovation: no system, a lot of money spent and many frustrated people.

Chapter 1 General Introduction

17

1.2.3 Probable causes of failure

The above example illustrates typical problems that obstruct innovation processes. Innovation processes are often underestimated [managers sending their assistants instead of coming themselves] and it is tempting to focus on the benefits [more customers] while overlooking the path that realizes these benefits, such as the cooperation of all parties involved, before as well as during the innovation process.

Laudon and Laudon [2002: 368] pose a key question to be considered before initiating an innovation process: Does the implementation of information technology within the organization really create an advantage? Answering this question asks for understanding and supervision of the process.

Typical problem fields are design, data, costs and the use of ICT-tools or the DSS

itself. Problems vary from immature technical design to incompatibility with operator incentives [Leonard-Barton 1995]. Typical factors of either success or failure of implementation are involvement and influence of users, support and involvement of management, complexity and risks, and supervision of the implementation process [Laudon & Laudon 2002]. In understanding these problems Leonard-Barton [1988] names transferability, organizational complexity, and divisibility as important parameters for the implementation of a technological innovation, such as a software implementation.

To prevent the problems that typically accompany organizational innovation processes from occurring, Fichman and Moses [1999] stress the integration of what is [already existing organizational processes] with what is to come [the software configuration]. They argue that it is important not only to focus on the consistency of the software configuration, but also to make it fit in the organizational processes, structures and of course the company policy. They suggest an incremental implementation strategy.

In sum, innovation processes are complex as there are many different factors and many different parties involved, often with different goals and gains. Therefore, goal and probable gain of the innovation process should be understood and agreed upon, and so should the way to reach that goal. In understanding organizational innovation, the need for implementation research – as part of the innovation process – has been stressed [e.g. Kimberly & Evanisko 1981; Voss1988; Klein & Sorra 1996]. The next section will elaborate on a knowledge perspective to organizational innovation. We take this perspective to understand the implementation process. The knowledge perspective is often forgotten and


18

this omission may lead to non-optimal solutions or adjustments of the innovation process.

1.3 Understanding organizational innovation from a knowledge perspective

1.3.1 Focus on knowledge in organizations

Over the last decades the focus on knowledge has gained interest in studying organizations [see e.g. Lammers 2003]. Perhaps partly inspired by the words of Francis Bacon over five centuries ago – knowledge is power – combined with increasing technology, around the mid 70s of the last century knowledge management was regarded as a new concern and of central importance to the public administration [see e.g. Henry 1974; Carroll & Henry 1975; Goerl 1975]. But of course the roots of knowledge management and the focus on knowledge can be traced back much further depending on one’s focus.

Today knowledge management [KM] is an established concept within the field of management and organization. An impetus for the literature on knowledge management came with Hedlund [1994]. Organizations increasingly regard knowledge as a factor for generating added value [as is innovation], which resulted in a focus on KM in organizations [e.g. Jorna 2001; Lammers 2003; Gazendam et al. 2005; Jorna 2007]. This focus on knowledge as a factor to optimize profits has led to a great deal of theorizing and research on knowledge in relation to organizational behavior. Lammers [2003] points to a range of approaches in the academic debate on knowledge management from information and library services and information systems and information technology [e.g. Newell & Simon 1972] to organizational learning and learning organizations [e.g. Argyris & Schön1978; Senge 1990; Huysman 1996; Argyris 1999], and strategic management [e.g. Nelson & Winter 1982; Cohen & Levinthal 1990]. Knowledge aspects that have been tackled range from the distributed nature of knowledge [put forward by Tsoukas 1996] and knowledge as resource in the firm [put forward by Grant 1996, and Prusak 1997] to knowledge creation and individual versus social knowledge aspects [put forward by Spender 1996]. Majchrzak, Neece and Cooper [2004] condense the focus of KM in capturing, transferring, and reusing knowledge in organizations.

Thus, knowledge is increasingly regarded as a factor to be considered in organization research. It is a crucial factor, not to be overlooked.


19

1.3.2 A knowledge perspective on organizational innovation

The above indicates that knowledge is considered to be an important factor in organization research. Not surprisingly, this awareness has also reached the re-search on innovation. For example, Sørensen and Lundh-Snis [2001] focus on the codification process of knowledge in the process of innovation. Cohen and Levinthal [1990] name prior related knowledge within a firm, the absorptive capacity, as the important indicator for the innovative capabilities of an organization. Nonaka and Takeuchi [1995] suggest knowledge creation to be at the heart of innovation processes and also Leonard-Barton [1995] regards knowledge as the main building blocks for sustaining innovation. Glynn [1996] views organizational innovation as fundamentally cognitive and Jorna [2006] even claims know-ledge to be the starting point of innovation. On top of this – and probably consequently – both innovation and knowledge share the consideration that they influence competitive advantage [e.g. Nelson & Winter 1982]. Greif and Keller [1990] link the understanding of innovation to knowledge, with a special focus on individuals.

Zooming in on knowledge related studies in innovation literature many of them emphasize the role of knowledge as a constraint to start the innovation process. We acknowledge this. Our interest, however, is in understanding the knowledge dynamics during the innovation process. We want to understand what we assume to be the core of the innovation process itself, namely the changing nature of the knowledge itself.

1.3.3 A multi-actor perspective

The traditional focus of management and organization research either takes the group level as the unit of analysis or the organizational processes and interactions. This main approach in management and organization research becomes particu-larly clear when we scan the topics discussed in the IEBM Handbook of Organiza-tional Behavior [Sorge & Warner 1997]. Ranging from markets, hierarchies, processes of organizing and strategic choice as theoretical approaches and paradigms to corporate governance, organization culture, leadership, and power, only two of the 48 topics in the IEBM Handbook directly focus on people: human relations and organizational populations. None of the topics take the perspective of the individual. Our perspective on the individual stands in a different tradition.

In studying the innovation process from a knowledge perspective we take a multi-actor approach [see e.g. Gazendam & Jorna 1998], which takes the individual as a


20

starting point to study groups, such as organizations but also to study artificial groups as in AI research [e.g. Helmhout 2006]. Our main consideration to take this perspective is that individuals have knowledge, not organizations. So, in studying knowledge within organizations it is crucial to focus on the level of the individual. This individual perspective on organizations stands in the tradition of Simon [1947] and March [March & Simon 1958] who studied organizational behavior with a focus on the individual. In line with Glynn [1996], who conceptualizes individual and organizational intelligences as being functionally similar [i.e., as purposeful information processing that enables adaptation to environmental demands], our focus in studying knowledge in organizational innovation is on the individual in an organization.

1.3.4 Bridging theory and practice

Another main issue in our study of innovation from a knowledge perspective is to bridge theory and practice. We want to see how knowledge functions in practice. The practice will be found in the domain of planning. However, we also want to focus on a theoretical framework that is solidly founded in both cognitive science and semiotics. Semiotics provides the foundation to understand the nature of the constituting elements of knowledge, that is to say representations or signs. In addition, cognitive science provides a theory of mental processes and processing and a methodology to conduct empirical research. The combination of cognitive science and semiotics thus provides the foundation to empirically test our hypotheses.

1.3.5 The knowledge domain of planning

Choosing the domain of planning to study the relation between knowledge and innovation suits our purpose in three different ways. First of all, planning is a highly cognitive activity, which involves a broad range of knowledge. Our aim is to study the dynamics of knowledge within the whole range. Secondly, the process of planning has been shown to be an important influence on innovation, especially in relation to the development of ICT. Among others, Leonard-Barton [1995] points to problem solving as an important aspect of innovation; problem solving is also a main activity within planning and supporting this planning with ICT-tools is a trigger for innovation. And thirdly, in a time frame of more than fifteen years, our faculty of Economics and Business has developed expertise on planning and planning support. In 1989, our faculty initiated the DISKUS-project, which stands for Dynamic Interactive Scheduling and Knowledge Utilization System. This


21

resulted in a tool called ZKR, which became a commercial product in 1995 and was successful in more than 50 hospitals. The present study will in particular focus on the implementation of ZKR in relation to the knowledge dynamics during organizational innovation.

1.4 General research questions

In theory, innovation is considered a key factor for organizations to survive. In practice, the innovation process is often accompanied by many problems, putting the estimated failure rate at 50 to 75 percent. Knowledge is considered a crucial factor in the innovation process. While literature at this point mainly focuses on why knowledge is a crucial factor, it remains unclear what actually happens to this knowledge. Therefore, this study aims to understand the role of knowledge during the innovation process in terms of knowledge types. To better understand this, we pose a preliminary research question, which we will use as a reference point for our review of relevant literature. In chapter 6 we will sharpen our research questions, based on our literature review, and we will construct a conceptual model to empirically address our research interest.

This study poses the following questions:

1. What happens to the knowledge of individuals during organizational innovation?

And in line with this main research question we wonder

2. Which kind of knowledge [content and type] is important during organizational innovation?

And, if knowledge changes

3. In what sense does knowledge change?

4. Which factors influence the knowledge dynamics during organizational innovation?

1.5 Outline thesis

To sharpen our focus we start with investigating the three theoretical pillars to this study, namely innovation, knowledge and the task domain of planning [see also figure 1.1]. Chapter 2 discusses the relevant innovation literature to under-stand the innovation process and its relation to knowledge. We start with


22

exploring the concept of innovation in general, and then move to exploring the innovation niche within knowledge management. We conclude the innovation chapter with the positioning of our study in this literature. The second and third theoretical pillars both focus on two aspects of knowledge; whereas the know-ledge pillar focuses on knowledge type, the planning pillar focuses on knowledge domain. Chapter 3 then discusses the study of knowledge dynamics in organizations, for which we introduce knowledge related concepts and we discuss the Information Space by Boisot [1995]. Chapter 4 presents our suggestion to bridge theory and practice in studying the relationship between knowledge and innovation, the cognitive-semiotic model. Chapter 5 deals with the knowledge domain of our study, planning. So the chapters 2, 3, 4 and 5 form the theoretical foundation for the present study; these chapters sharpen our research questions to com-pose our conceptual model from which we deduct our hypotheses in chapter 6.

The chapters 7, 8, and 9 focus on the empirical side of this study. Chapter 7discusses the empirical setting of Bartiméus, a care institution for visually impaired and blind people; we describe the organization in general and in particular the situation at Bartiméus before the innovation. Chapter 8 discusses the methodological issues, such as the subjects involved, research design, method, operationalization of the theoretical concepts and procedure. Chapter 9 analyzes the data through hypotheses testing and secondary analysis. And finally, chapter 10 draws conclusions bridging the research questions to the data analysis; we conclude reflecting on the outcomes of this study and speculating on its implications.

Figure 1.1: Schematic outline of thesis

23

Chapter 2

Organizational Innovation

2.1 Introduction

There is strong agreement among organizational researchers on the importance of innovations [e.g. Wolfe 1994]. In making this point, innovation has, for instance, been related to organizational competitiveness and effectiveness [Wolfe 1994], competitive advantage [e.g. named by du Plessis 2007, Leonard-Barton 1995], long-term survival of organizations [Ancona & Caldwell 1987 in Scott & Bruce 1994] and as the ‘key means of adapting to change’ [Gopalakrishnan & Damanpour 1997: 15].

Perhaps partly because of the agreement on its importance, the literature on innovation is wide and diverse. The range within innovation literature stretches from the disciplines of social and occupational psychologists, personality theorists, sociologists, management scientists and organizational behaviorists [named by King 1990] to anthropology [named by Subramanian and Nilakanta 1996], engineering, economics, marketing [named by Gophalakrishnan & Damanpour 1997] and education [Heywood 1965]. The foci vary from stages within the innovation process [Pavia 1991; Song, Thieme, & Xie 1998], levels of aggregation [Waalkens 2006; Jorna 2006], and key factors of innovation such as creativity [e.g. Borghini 2005; Kratzer, Leenders, & van Engelen 2004], elite values [Hage & Dewar 1973] and tacit knowledge [e.g. Senker 1995; Lam 1998] to the complexity of the division of labor [Hage 1999], interfirm relationships [Knudsen 2007] and managing knowledge [Leiponen 2006]. More theory based foci concern the differentiation in the different types of research questions [Wolfe 1994], perspectives on innovation [Slappendel 1996] and comparing academic inventors to their corporate counterparts [Golish, Besterfield-Sacre, & Shuman 2008]. To better value the content of individual innovation studies within this abundance of literature the call for comparability emerges. For instance, Garcia and Calantone [2002] urge the methodizing and classifying of the use of terms to facilitate common understanding.

The first aim of this chapter is to discuss relevant organizational innovation literature to create a general understanding of the innovation process. Secondly,


24

we zoom in on the knowledge management literature related to innovation research. And thirdly, we position the present study within the innovation research literature, in particular related to knowledge management. In order to reach these aims we define innovation [2.1] and we review the innovation litera-ture on different foci [2.2]. Then, we discuss two influential studies on knowledge creation within organizations [2.3] and we position the present study [2.4].

2.2 Defining innovation

Innovation, rooted in the Latin verb innovare, originally implies ‘to make something new’. Tidd, Bessant, and Pavitt [1997] point out that this original implication refers to invention rather than to innovation; to get from invention to innovation takes more than just a good idea, it takes an integrated approach [Khilji, Mroczkowski, & Bernstein 2006]. To illustrate the difference between these two concepts Tidd et al. [1997] refer to the introduction of the modern telegraph by Morse. Although Morse did not invent the modern telegraph – he only invented the code – he has been widely credited for it. Rather, Morse envisioned the possibilities of this new tool and he realized spreading of this new concept of long distance communi-cation. Morse turned the invention of the modern telegraph into an innovation.

So, innovation goes further than invention. This reflects in the definitions on innovation. Many definitions refer to the application of something new. And then, additional aspects are emphasized. Sundbo [1998: 19] and Satchell [1998: 18] respectively talk about ‘renewal’ and ‘conversion of an idea into an outcome’, stressing the stirring up that the innovation brings about. Satchel also refers to ‘continuous rearrangement’, stressing the incremental aspect and ‘financial return’, and emphasizing the economic aspect. Tidd et al. [1997: 24] focus on turning an ‘opportunity into new ideas’, which suggests the role of choice and focus. These approaches to innovation show similarities with the more economic and well-known view of Schumpeter, which combines the elements of creation and destruction [McCraw 2007]. Definitions more focused on organizational processes include concepts as ‘marketing’ [Maguire, Kazlauskas, & Weir 1994] and ‘perceived as new’ [Zaltman, Duncan, & Holbek 1973; Rogers 1995]. West and Farr [1990] have a somewhat technical definition of innovation, but it is precise and most suited for our purpose; we will use their stipulation on innovation:

The intentional introduction and application within a role, group or organization of ideas, processes, products or procedures, new to the relevant unit of adoption, designed to significantly

Chapter 2 Organizational Innovation

25

benefit the individual, the group, organization or wider society [p.9]

The above definition on innovation grasps the essence of ‘applying something new’ in combination with intention, relevance and benefit. Furthermore, it emphasizes newness in relation to the unit of adoption. Making newness dependent on an adoption unit puts the impact of innovation into perspective as we will see in the following section.

2.3 Foci in innovation research

Gophalakrishnan and Damanpour [1997] distinguish three pragmatic dimensions to frame the research on innovation, namely the dimensions of 1] levels of analysis, 2] types of innovation and 3] stages within the innovation process. Although they focus on research within the disciplines of economics, organizational sociology and technology management, this framework fits our purpose to review relevant innovation research to gain a general understanding of innovation and to position the present study within this field.

2.3.1 Improvement of performance: Levels of analysis

Jorna and Waalkens [2006] relate the level of analysis, or level of aggregation as they called it, to one of their three key elements of innovation, ‘improvement of performance’ [the other two being ‘change’ and ‘newness’]. Hence, an innovation aims to improve and improvement of performance implies measuring and comparing something at time A with that same something at time B to establish whether or not an improvement has taken place. That something refers to the level of analysis, which we choose. For instance, we could focus on changing economies but we could also focus on R&D research units within a particular organization. Economics and R&D-units are objects of study at different levels of aggregation.

We can roughly distinguish five levels. The highest level is probably at the level of countries. This macro economic research is mainly measured in terms of numbers, for example graduates at technical universities, innovation projects or comparisons of innovativeness between countries [van Ark & Pilat 1993].

Research at the industry level focuses on aspects, such as the chain of value, strategy and network management. For instance, O'Mahony & van Ark [2003] focus on productivity and competitiveness, with a taxonomy on innovation, Khilji


26

et al. [2006] present an integrated innovation model for biotech firms and Pavia [1991] focuses on the early stages of new product development in entrepreneurial high-tech firms.

One level lower the organization comes into sight. Research at this level studies improvement of organizational performance [Gopalakrishnan & Damanpour 1997]. For instance, Zack [1999] proposes a knowledge management architecture for performance improvement; through facilitating knowledge integration across different contexts this architecture provides new insights and increases the scope and value of the knowledge. At this level of research we can find studies on organizational size [van de Ven, Polley, Garud, & Venkataraman 1999], organizational structures [Sapolsky 1967; Kanter 1983 in Jorna & Waalkens 2006], resources such as absorptive capacity [Cohen & Levinthal 1990], organizational performance [Damanpour 1990], organizational strategy [Duh, Chow, & Chen 2006], organizational climate or culture [Nicholson 1990; Burnside 1990] and extra-organizational factors, such as competition, and environmental factors [Tushman & Anderson 1986].

One level deeper, innovation researchers analyze subunits within the organization, for instance groups or departments and the interaction between the group members in relation to the performance improvement. Gopalakrishnan and Damanpour [1997] name R&D departments as the most widely studied organizational subunit.

Analysis at the lowest level focuses on the individual within the organization. Traditionally the research on individuals involved in innovation focuses on creativity [King 1990]. However, other foci include the work role of individuals [Farr & Ford 1990], leadership and problem solving [Scott & Bruce 1994].

2.3.2 Types of innovation

A second dimension in innovation research is concerned with the different types of innovation. Gopalakrishnan and Damanpour [1997] distinguish three sets of contrasting types, namely 1] radical versus incremental, 2] technical versus administrative, and 3] product versus process.

Differentiating innovations on newness: From radical to incremental

Gopalakrishnan and Damanpour [1997] contrast radical innovation to incremental innovation. Nuances to this two-split have been made, in which these two contrasting types are more like two points on a continuum. Jorna and Waalkens


27

[2006] come up with five types [three additional types], based on the typology by Garcia and Calantone [2002] for identifying technological innovations, which they place on a continuum of newness. We then can distinguish: 1] radical innovations, 2] really new innovations, 3] discontinuous innovations or game changers, 4] incremental innovations, and 5] imitations. This typology stresses the importance of whose perspective is used to evaluate the innovativeness [Gopalakrishnan & Damanpour 1997] or newness [Jorna & Waalkens 2006]. Thus, newness is directly related to the adoption unit. Slappendel [1996] also points to this aspect of innovation. She names newness as a widely accepted key distinguishing feature of innovation and adds that the ‘perception of newness also serves to differentiate innovation from change’. The aspect of perception was already introduced in 1973 by Zaltman et al., who define innovation as ‘any idea, practice, or material artifact perceived to be new by the relevant unit of adoption’. Rogers [1962] also emphasizes perception in defining innovation as ‘an idea perceived as new by the individual’ [13].

Radical innovation is very far-reaching. When these innovations hit the market they trigger many new innovations, for instance the introduction of integrated circuits, which had a great impact on electronics technology [Jorna & Waalkens 2006]. Then comes the really new innovation, which does not catalyze new innovations in the sense that radical innovation does. This innovation comes in the form of a new product or service that leads to a discontinuity and newness in the market. For this type of innovation there should be newness in either technological or marketing sense, for instance the introduction of the cell phone. Then comes the discontinuous innovation. These innovations offer improvement of performance, reduction of costs or introduce an existing item with completely new characteristics. An example is the introduction of broad screen televisions. Then, fourthly, incremental innovations build further on existing markets and technology, which it provides with new features or improvements. Key issues for this type of innovation are adaptation, refinement and enhancement and they have an iterative nature, for instance updates on software tools. Then, opposite to radical innovations stand imitative innovations. These innovations are new to a particular organization, but not new in terms of product or process, for example the copying of Western goods by Asian producers.

Technical versus administrative

The contrast between technical innovations and administrative innovations reflects ‘a more general distinction between social structure and technology’ [see


28

also Damanpour & Evan 1984]. Damanpour [1987] notes that this distinction is founded in nature of these two innovation types and together they complement each other. Simply put, distinguishing technical and administrative innovations makes it easier to understand the organizational differences in response to these different types of innovation.

Technological innovations on the one hand change organizations, in that changes in technology are made; technology can be for example a tool, a technique, or physical equipment. Technological innovations ‘produce changes in products or services, or in the way those products are produced or services are rendered’ [Damanpour 1987: 677]. Administrative innovations on the other hand change the structure or the administrative process of an organization. In this way they indirectly change the basic work activity and more directly to the management. So, an administrative innovation is ‘the implementation of an idea for a new policy pertaining to the recruitment of personnel, the allocation of resources, the struc-turing of tasks, of authority, of rewards’ [Evan 1966: 51 in Damanpour 1987: 677].

Product versus process

Contrasting product innovation to process innovation is based on the differences in effects of these innovations on areas and activities [Gopalakrishnan & Damanpour 1997]. Process innovation involves ‘tools, devices, and knowledge in throughput technology that mediate between inputs and outputs and are new to an industry, organization, or subunit’ [p.18]. Product innovations involve new ‘out-puts or services that are introduced for the benefit of customers or clients’ [p.18].

We will follow up on the distinction between product innovation and process innovation in the next section that focuses on the third dimension in innovation research, stages in the innovation process.

2.3.3 Stages in the innovation process

Distinguishing stages within the innovation process is widely used [e.g. Hage & Aiken 1970; Zaltman et al. 1973; Hage 1980; Legge 1984]. This line of research decomposes the innovation process into stages and focuses on ‘the sequential nature of precursor events and on their determinants’ [Wolfe 1994: 411]. To Wolfe [1994: 409] this research falls in the research stream of Process Theory Models, the main aim of which is to understand the nature of the innovation process and it therefore questions of how and why innovations emerge, develop, grow, and terminate.


29

Although very workable, some researchers find this stage approach to innovation to be too simple. They say innovations have an iterative nature, more complex and uncertain than the stage type models suggest [e.g. van de Ven, Polley, Garud, and Venkataraman, 1999]. Van de Ven and Polley [1992] emphasize this iterative nature in their adaptive model leaving an important role for trial-and-error.

Wolfe [1994: 411] discerns a general pattern in the variety of research on innovation stages, which he sums as follows

� a decision-making unit becomes aware of an innovation’s existence,

� a problem or opportunity is matched to the innovation,

� the innovation’s costs and benefits are appraised,

� sources of support and/or opposition attempt to influencethe process,

� a decision is made to adopt(reject) the innovation,

� the innovation is implemented,

� the innovation decision is reviewed and confirmed (reversed),

� the innovation becomes accepted as routine, and

� the innovation is infused, i.e. is applied to its fullest potential

The above suggests at least nine stages to be studied within the innovation research. However, a rough distinction – widely used – that covers all aspects named in the above is one between the generation or initiation for the first part of the innovation process and adoption or implementation for the second part. We will use this two-split to discuss the stages of the innovation process in more detail, as this rough distinction avoids discussions on particular sub stages and its sequential or parallel occurrence.

The initiation stage

The initiation stage can be interpreted either from a product innovation perspec-tive or from an organizational innovation perspective [see also figure 2.1]. These two process lines come together in the implementation stage of the innovation process. Although we distinguish two paths to the implementation process within an organization, both the initiation stage of product innovation as well as the initiation stage of organizational innovation start with the awareness of a discre-pancy between the present situation and the desired situation [Legge 1984]. This


30

awareness triggers the desire to innovate. As this study takes on an organizational perspective we will start our elaboration on the organizational innovation. Then we will briefly discuss the product innovation side of the initiation stage.

Figure 2.1: Two separate initiation stages leading to the implementation stage of organizational innovation

Organizational innovation

Zaltman, Duncan, and Holbek [1973] refer to the discrepancy between ‘what is’ and ‘what is desired’ as the “performance gap”. This gap can emerge for a number of reasons. For instance, changes in technology, external influences, and experience are among some of them. Also, an organization may want a better financial position or an amelioration of the position and performance within the market. But prior to the awareness of the performance gap the organization first needs to detect this gap. This awareness is mostly caused by dissatisfaction on the current state of affairs. Dawson [1995] remarks that the desire to change is not always rational. For instance, it can have somewhat irrational causes, such as wanting to be fashionable. Interestingly, besides the perception, the measurement


31

of the performance gap seems to be important as well, as the more difficult the measurement the less likely will the performance gap be detected. This can result in a performance gap that will grow even bigger [Legge 1984]. Research at this stage of the innovation process deals with problem solving and decision-making and it has been studied intensively in relation to creativity [see e.g. West & Farr 1990] and R&D research [e.g. Song, Thieme, & Xie 1998]. Note that the foundations for the failure of an innovation can be traced back to this early stage of the innovation process. As we recall the example about the hotel chain described in chapter 1, most managers did not see the problem that the CEO’s

seemed to detect.

Once there is [some sort of] consensus on the size and quality of the perfor-mance gap or the problem, the innovation becomes more tangible and the pro-blem can be operationalized and worked on. Legge [1984] distinguishes this part of the innovation process as a separate stage within the innovation process. She remarks that organizations at this [sub] stage should not confuse innovation as a means with innovation as a goal itself. Without naming the goal that the organiza-tion wants to achieve with the innovation process it will become very difficult if not impossible. Differences in [expected] goals can lead to serious miscommuni-cation [Bosch-Sijtsema 2001], frustration, and even failure. Consultancy firms [e.g. Twynstra Gudde, KPMG, Ernst & Young, etc.] that focus on knowledge manage-ment distinguish the higher goal of innovation – to eventually work more efficiently – from its sub goals. These sub goals in turn can be operationalized. For instance, a higher goal such as integrating practice and education to optimally train qualified people through learning by doing can be defined into goals, primary goals, and efforts [e.g. Wijnen & Kor 1996]. This practical approach emphasizes the importance of operationalizing the [abstract] higher goal or main goal into [concrete] sub goals that can directly be linked to tangible incorporated tasks. Especially at this conceptual stage of the innovation process management should not underestimate the innovation process [Schwab 1999]; the expectations, for instance about the possibilities that a software innovation tool provides, should be realistic. High expectations can lead to great disappointment. Underestimation can be tackled by anticipation of the consequences that the innovation will have for the organization [Dawson 1995]. This can simply be done through estimating and predicting positive as well as negative outcomes that will result from this innovation. For instance, when an organization expects that new knowledge needs to be acquired for an implementation of decision support software, training can be provided for the people involved.


32

Product innovation

The focus of the present research is on organizational innovation from outside in; something ‘new’ is implemented into the organization. This new thing is something that has been developed through innovation as well. Although the research on product innovation does not fall within the scope of the present study we do want to note the link with the present study research. Hofstede’s [1992] quality chain explicitly links the design and development of innovation products to its use within an organization [see figure 2.2]. He argues

The relationship between the links of this chain is that the qua-lity of the right-hand link is one of the determinants of the qua-lity of the left-hand link … The DSS is central in the present context. The quality chain is also, in a way, a process – product chain. Starting at the right-hand side, the process of DSS deve-lopments results in the product ‘DSS’. With this DSS, a decision maker can carry out a process of problem solving resulting in the product ‘decision’. This decision then co-determinates the processes that constitute the functioning of the organization. High-quality processes of DSS development and of problem solving are crucial if DSS are to be really used. [p.108]

Figure 2.2: The quality chain of DSS [Hofstede 1992 in Mietus 1994: 10]

So, Hofstede relates the quality of the product to the involvement of its users. This involvement of users in the development of products should be considered with caution, however. Song, Tieme, and Xie [1998] promote the involvement of ‘functional areas’ in the process of product development to be restricted to specific stages to avoid counterproductive results, for instance idiosyncratic feedback [Leonard-Barton 1995]. The end of the initiation stages of both the product innovation and the organizational innovation form the beginning of the implementation stage; from the pro-duct innovation side the implementation


33

stage is the extension of the product development process and from the organizational innovation side the implementation stage is the extension of the ideas that have been formulated to change the organization.

The implementation stage

When the performance gap has been detected and initiation has taken place –that is, a concrete plan has been made of how to bridge the gap – then the more tangible part of the innovation process will start; this is the major part of the innovation process. We do note that a proper and well thought-out initiation stage will speed up the implementation stage considerably. The ideas generated in the initiation stage have to be implemented and therefore ‘important others’ need to be convinced [Jorna & Waalkens 2006]. This stage involves ‘a process of organizational change which directly affects the technical and social systems of an organization [Kimberly 1981; Damanpour & Evan 1984].

During the implementation stage a new process, product or idea will be implemented within the organization. Thus, at this point in the innovation process the innovation can be seen as coming from outside and it is implemented in the organization. In fact, as we saw at the beginning of this chapter, the implementation stage is what discriminates invention from innovation. Whereas invention does not involve the actual use of the new product or idea, innovation does. The applicability and actual application of the new concept is essential to the innovation process. Thus, innovation is partly defined in terms of its actual use and application for the target group or market. Elaboration on this thought can cause confusion, as at the starting point of the innovation process, it is not clear whether the process will be successful or not. However, it is perfectly well accepted to talk about an innovation process that resulted in failure, thus without actual adoption.

Technology often plays an essential part in the implementation stage of the innovation process [e.g. Leonard-Barton 1985, 1988, 1995; Majchrzak 1988; Verjans 1999]. Even though technology such as a software tool may be characterized as standardized, the implementation of a standardized software tool remains a custom made process [Geerts 1999] and it is often characterized by problems [e.g. Laudon & Laudon 2002]. The degree to which a software tool is standardized does vary; for example ERP [enterprise resource planning], a software tool widely used, can be implemented in three different variations, generic, preconfigured and installed [Klaus, Rosemann, and Gable 2000]. Kræmmerand, Møller, & Boer [2003] underline the importance of a clear plan


34

with goals and strategy for the implementation of a technological software tool such as ERP. They argue that an implementation is rather a journey than a single action. Furthermore, an implementation triggers different kinds of learning, exploitative and explorative; this ‘results in radical and incremental changes and outcomes that are partly predictable and partly unpredictable’ [p.343].

The success of the implementation process [and consequently the innovation process as a whole] will be determined by the people who will be working with the new technique [the end-users or the adoption unit in more technical terms]. Therefore, it is essential that these end-users are given enough time and space to adjust to the new way of working [Geerts 1999]. Typical problems occur when a software tool implemented in one part of the organization generates information for other parts of that organization. Without proper communication between these parties a chain reaction can be set, in particular when a sequential party has to rely on the information supply generated by a first party to which that information is seemingly worthless. So end users do not necessarily directly benefit from the innovation tool even though they might acknowledge that the implementation of this software tool has great benefits for other units within the organization, for instance the personnel department that save a great amount of time when information is supplied digitally.

One of the ways to aid the end-users is to offer them training [Geerts 1999] to learn to work in a new way. However, the relationship between training and lear-ning is not always that clear. Antonacopoulou [2001] notes not to underestimate the paradoxical nature of the relationship between training and learning. There is a tension field between the individual and the organization as a crucial factor in the relation between learning and training. Antonacopoulou [2001] concludes

… training cannot be assumed to produce learning, nor is learning always an integral part of training, partly because even when training may result in some learning, the organization may not provide the necessary infrastructure to support such learning after the training has been completed. … individuals have different expectations from training interventions … and are frequently unable to utilize the knowledge gained from training courses due to factors such as the relevance to their current job, the extent to which training is perceived to be relevant or practical, and whether it is provided at the right time and through the right methods [p.331]


35

So the effect of training is dependent on a number of factors, among which are expectations, relevancy and timing of the training. Particularly important is a situation that supports the use of the knowledge acquired at the training to enable and facilitate learning. This tension between training and learning shows parallels to the oxymoron that Weick and Westley [1996] sense in the term organizational learning; they point out that organizing focuses on control and learning focuses on the opposite, to unravel and disorganize.

The final part of the implementation stage can be characterized as routinization [Legge 1984]. Nelson and Winter [1982] point out that routines are a vital part of an organization. Coriat and Dosi [1998] explore the origins and roles of different organizational routines that are important to sustain diverse organizational structures. They argue that a good part of explaining why apparently 'superior' organizational forms diffuse very slowly lies in the fact that organizations are crucial [although not exclusive] repositories of knowledge; this is where the routines come in. Organizations are modified through time by their 'higher level' rules of behavior and strategies. Thus, routinization could be seen as a marker point that the innovation process has been completed in that the innovation has ‘conquered’ a place in the organizational structure.

Although of a somewhat different nature, according to Wolfe [1994], we do want to mention the research on the diffusion of innovation as related to the implementation stage of innovation. Research on the diffusion of innovation concerns differences in success of innovations in terms of their diffusion within a population. What distinguishes successful innovations from the innovations that are forgotten and thus diffuse badly? Rogers [1995] emphasizes the importance of diffusion for innovations; he argues that one can invest great sums of money in developing innovation, but still the overall goal is for the intended audience to actually use the innovation. Therefore, ‘the purpose of many diffusion studies has been to determine methods by which diffusion can be hastened’ [p.2/3]. Typical subjects to be studied within this field of research include that of the adopter, the environment, and the process by which the innovation is communicated.

The above shows that innovation research focuses on many and different factors. It furthermore shows that the area of innovation research is not precisely defined. It shows parallels to and overlap with other fields of research, because of change, newness and improvement [Jorna & Waalkens 2006]. We therefore briefly elaborate on three related research areas, namely organizational change, organizational learning and absorptive capacity.


36

2.3.4 Related research areas

Organizational change

The element of change is prominent in innovation literature [e.g. Hage 1999; Strambach 2002] and can be considered as one of three key elements to define innovation, next to newness and improvement of performance [Jorna & Waalkens 2006].

Literature on organizations and change focuses on a variety of factors. For example, at the individual level change in organizations has been related to sources of resistance to change [George & Jones 2001]. These sources point to ‘the way individuals construct and make sense of the social and organizational world’. Löwstedt [1993] studied the impact of cognitive structures and their importance for organizational change and the level of the individual. He found support for the hypothesis that human cognitive structures affect `the structuring of organizations. At a higher aggregation level literature on organizational change studies for example ‘institutionalized routines and practices embedded in organizational structure and culture’ [Hannan & Freeman 1984]. Nelson and Winter [1982] also name organizational routines as one of the three important concepts for economic change, next to ‘search’ [organizational activities which are associated with the evaluation of current routines] and ‘selection environment’ [considerations which affect the well-being of an organization and hence the extent to which it expands or contracts].

In the following we focus in more detail on two popular concepts of the ‘change-literature’, namely organizational learning and absorptive capacity. We will briefly discuss these two concepts and relate them to the change element of innovation.

Organizational learning

Literature on organizational learning [OL] is strongly related to change, in particular the results that can eventually become of change. A primary criterion for effectiveness is change of behavior [Argyris 1999]. To establish effectiveness this behavior change can either be established through direct behavior modification or through understanding ‘the meanings people create when they deal with each other’. Organizational learning has to do with the latter way of changing. Argyris [1999: 68] defines OL as a process of single- and double-loop learning. He states:

Single-loop learning occurs when matches are created, or when mismatches are corrected by changing actions. Double-loop


37

learning occurs when mismatches are corrected by first examining and altering the governing variables and then the actions.

The difference between single-loop and double-loop learning is in examining of what caused the mismatch. The matches and mismatches refer to the desired outcome of a situation that an individual has in comparison to the actual outcome. Thus, of vital importance to organizational learning is an understanding of the discrepancy between what is and what is desired, this parallels the initiation stage of innovation. Argyris does not directly link OL to innovation. He does however indirectly link OL to innovation as he discusses Van de Ven and Polley [1992] as being a part of OL literature. Interestingly, Argyris specifically states that higher educated people are less capable of double loop learning and therefore in understanding the discrepancy. According to Argyris, these people rarely experience failure, and therefore they have no experience in learning from failure. On these grounds we might expect that higher educated people will have a tougher time during innovation processes than people with a lower education do, in particular when the innovation is not directly content related. We will elaborate on this issue in chapter 6.

However, shared understanding about the meaning of organizational learning is not yet established [Huysman 1996]. For instance, Weick and Westley [1996] reveal a more skeptical view on OL; they see a natural contradiction between organizing and learning. Whereas learning implies to ‘disorganize and increase variety’, organizing on the other hand implies to ‘forget and reduce variety’ [p. 440]. They state that this important tension is frequently forgotten in OL

literature. The contradiction named by Weick and Westley pinpoints the problems and difficulties to be expected when OL is intentionally introduced in an organization. However OL is not considered to be impossible. In fact, a combination of order and disorder is important. To trigger this process Weick and Westley [1996] suggest to use humor and improvisation on routine.

Waalkens [2006: 19] links OL to innovation in that

Innovation involves a special kind of organizational learning concerned with information and knowledge that is new to the whole firm. The sources of new knowledge can be internal or external.


38

Thus, OL shows a natural overlap with innovation processes and takes the individual as unit of analysis. Also, the importance of knowledge is suggested for these processes.

Absorptive capacity

A third adjacent research field to innovation is that of absorptive capacity [AC]. With the element of change central to it Cohen and Levinthal [1990] present ‘absorptive capacity’ as a new perspective on learning and innovation in the organizational context. AC refers the capacity to recognize, acquire and assimilate as well as to exploit information from outside the organization within the organization; it involves prior related knowledge. Thus, the focus is on information from outside the firm and the ability of the firm to understand the importance of this external information with the internal knowledge that they have. Waalkens [2006] explicitly links AC to innovation and notes that AC ‘entails external learning with the aim of innovation’ [p. 19]. The AC can be exploited by the organization through ‘transfers of knowledge across and within subunits that may be quite removed from the original point of entry’ [p. 131-132]. Thus, although a central point to AC is the focus of knowledge outside the firm, the enabling of this process lies in the knowledge and communication from inside the firm. This means that communication and communication structure are important in considering the organizations’ AC. Cohen and Levinthal use R&D expenditure asthe explanatory variable, making AC tangible. They emphasize the role of basic research in firm learning in stating that ‘R&D creates a capacity to assimilate and exploit new knowledge’ [p. 148]. The adoption and diffusion of innovations are also considered to be affected by AC. As AC is a function of the firm’s level of prior related knowledge, AC directly influences the ease of learning; this is the foundation for [technology] adoption. Thus, the more AC the faster innovations can be adopted and diffused. Leonard-Barton [1995] explicitly includes the process of absorbing external knowledge as one of the innovating organizational activities that ensure the core capabilities [see section 2.3].

Thus, the element of change in relation to innovation research is important in the organizational learning and absorptive capacity research. Both OL and AC leave an important role for knowledge. However, whereas OL links the knowledge to the individual, AC chooses a higher level, for instance R&D expenditures or the number of innovation projects [Waalkens 2006].

This section reviewed the relevant innovation research and we discussed potential problems and difficulties. Failure appears to be a serious threat and this


39

becomes particularly apparent, not surprisingly, during the implementation stage when discrepancies in ideas and ways of working reach the surface. Moreover, even when the innovation process can be regarded as successful in the sense that the process went well, this does not automatically imply that the innovation process benefited either the organization or the individuals within that organization. That is, innovation should be useful in the first place.

2.4 A knowledge perspective on innovation

2.4.1 Introduction to the knowledge perspective

A knowledge perspective on innovation has gained interest [e.g. Swan & Newell 2000; Sørensen and Lundh-Snis 2001]. In taking a knowledge perspective to study the innovation process, and in particular the implementation stage, we can sense the causes and consequences of the tension – the discrepancy, which an innovation process holds per definition – so that we can understand what actually happens. Understanding the knowledge dynamics of end-users can help to decrease the tension and discrepancy between the first and second stage ofinnovation processes. This section focuses on the additional value of a knowledge perspective in studying innovation. In the following we discuss two influential studies that place the use knowledge within organizations at a key position. Nonaka and Takeuchi emphasize the social aspect of knowledge creation in the innovation process and Leonard-Barton emphasizes the use of knowledge as an engine to innovation activities in technological firms.

2.4.2 Knowledge dynamics as the motor of innovation

Nonaka and Takeuchi: The social aspect of organizational innovation

Nonaka and Takeuchi [1995] were among the first to directly relate the success or failure of innovation to knowledge. However, they do not say that knowledge per se is the critical factor [p.18], but rather that the creation of knowledge is the foundation for continuous innovation. They support their view relating the successes of Japanese firms to the way that they organize their organization. According to Nonaka and Takeuchi these Japanese firms are able to innovate continuously through creating knowledge at an organizational level, which leads to competitive advantage.

Their book The knowledge creating company describes how Japanese companies create the dynamics of innovation. Nonaka and Takeuchi suggest that a middle-


40

up-down management structure facilitates the creation of knowledge and that knowledge creation in turn stimulates the innovation of different products and services. In other words, knowledge creation is essential in order to stimulate and realize product and service innovation.

The total knowledge creation process consists of four phases [also known as the SECI-model]. The process starts as a creative idea at the level of the individual, and eventually this idea will find its way up to become knowledge at the organiza-tional level. In the first phase the individual expresses her [non-verbal] ideas through metaphors and drawings, the knowledge in this socialization phase [S] is tacit. The socialization phase opens the way to externalization in the second phase [E], in which the individual verbalizes the tacitly created knowledge. Therefore, communication with other people becomes possible. The externalized declarative knowledge of the second phase functions as an input to start the third phase of combination [C]; the individual will share the knowledge with other members of the organization. Finally, these other members in turn will internalize this shared knowledge in the fourth phase, internalization [I]. When these four phases [SECI] are completed the knowledge has moved upwards in the organizational spiral, and the process of knowledge creation starts again, but at a higher level. So, the idea of one individual can spread through an organization and it can develop intoknowledge at the organizational level. The SECI-model suggests that knowledge creation is crucial to the innovation process. It creates an environment in which innovative thoughts can mature into concrete innovative products. The model emphasizes the individual knowledge within an organization as a basis for success at the organizational level. It also emphasizes the function of knowledge types in innovation processes, but these differ from the types that we distinguish in the present study. We will explain this more fundamentally in chapter 4.

Thus, in short, Nonaka and Takeuchi see knowledge creation as a crucial basis for innovation, not just the mere existence of certain knowledge within an organization. Knowledge creation is not established at once, it is a process established in phases, which starts at the individual level and then works its way up to the organizational level.

Leonard-Barton: Knowledge as a wellspring for organizational innovation

In her book Wellsprings of knowledge Dorothy Leonard-Barton [1995] points to knowledge as the building blocks for sustaining innovation, in particular for organizations that compete on the basis of technological advantage [as opposed to personal services, access to natural resources, artistic talent, or distribution


41

rights]. Knowledge accumulates slowly, and it is constantly being created; knowledge can be viewed as a non-static reservoir for new ideas and corporate renewal. These building blocks can be created by linking people [and thus their capabilities] to organizational activities. Leonard-Barton reasons:

Activities … create a firm’s capabilities. … These activities have [no separate meaning] from the people who conduct them since those individuals bring to the activities a set of idiosyncratic abilities, histories, personalities. Each person or team conducts the activity in a distinct manner. Thus, knowledge building for an organization occurs by combining people’s distinct individualities with a particular set of activities. It is this combination that enables innovation, and it is this combination that managers manage. … managing the activities as a sterile process, without consideration for the innovation potential the actors bring to it, is dangerous. [p.8]

So, an organization should conduct innovating activities to stay ahead of the game. Leonard-Barton distinguishes four of such organizational activities:

1. Shared, creative problem solving

2. Implementing and integrating new methodologies and tools

3. Formal and informal experimentation

4. Pulling in expertise from outside

Problem solving

The first key innovating activity that Leonard-Barton identifies is shared problem solving. Individuals within an organization are limited by their background. This should be acknowledged to avoid core rigidities. The process of creative abrasion is a way to stay critical. As this will not happen spontaneously an organization should facilitate this process, e.g. through a mixture of skills among employees. But also, managers need to encourage ‘integrate skills among employees and develop those skills themselves’ [p.89] to ensure the creative abrasion to be fruitful rather than destructive.

Implementing and integrating new technical processes and tools

What interests us the most is the second activity, the activity of implementing and integrating new technical processes and tools. Leonard-Barton stresses that this activity should not just be handled as the execution of a plan, but it should be


42

managed as an innovation project. The users provide critical information to be integrated during design and should therefore play an active role in the co-development of the process tool. This way of working increases effectiveness as well as efficiency as a result of mutual adaptation. That is, ‘adapting both the technology to the user environment and the user environment to the technology so as to exploit its full potential’ [p.110]. Leonard-Barton notes a point of caution to user involvement. The user feedback should not be atypical as this could damage rather than facilitate the design process.

Experimenting and prototyping

Experimenting is the way to create new technological capabilities, as it introduces ‘new sources of knowledge, new channels of information, new methods for sol-ving problems’ [p.134]. Not only the successful experiments are valuable, but also the unsuccessful ones, as organizations can especially learn from their failures. Like creative abrasion learning from failure is not something that comes naturally. Lear-ning mechanisms should be ensured through feedback channels to product deve-lopers and project managers. This happens in an environment in which failures can be examined openly. Leonard-Barton distinguishes avoidable failures from intelligent failures and emphasizes that intelligent failures should be encouraged.

Importing and absorbing technological knowledge from outside de firm

The three innovating activities discussed in the above are not enough to ensure and maintain core capabilities. To fill this gap additional knowledge should be imported from outside the organization. The degree to which an organization is able to do this is called ‘absorptive capacity’ [see also the section, 2.2.4], the ability to identify, access, and use technology from a wide variety of sources. The capability gaps can exist for different reasons, for instance. a deliberate corporate policy to lessen internal research. As goes for the other innovating activities, this activity of accessing outside knowledge does not happen automatically. To ensure the absorptive process a set of skilled activities are required, among which is ‘identifying and effectively using those employees who serve as technological gatekeepers and boundary spanners’ [p.175].

2.5 Characterizing the present study within the literature of organizational innovation

The studies by Nonaka and Takeuchi and by Leonard-Barton both stress that the dynamics of knowledge lie at the heart of innovation, which in turn is crucial to


43

the survival of an organization. They name different actions that need to be considered in ensuring these knowledge dynamics. Nonaka and Takeuchi combine the processes of socialization, externalization, combination and internalization into a knowledge creating spiral to ensure knowledge creation. And Leonard-Barton focuses on the innovating activities of problem solving, implementing new tools, experimenting and pulling external knowledge to ensure the core capabilities. Although both studies hint at the dynamics in knowledge types, Nonaka and Takeuchi even explicitly do so; there is no actual research on this aspect. That is, it is not clear what actually happens to the knowledge – in terms of content and type – during innovation. Nonaka and Takeuchi do parallel their SECI model to case studies from the field, but the empirical foundation to their model remains weak. Gourlay [2006] even argues that their empirical foundation is ‘non-existent, anecdotal, or open to alternative explanations’ [p.1416]. But more importantly, the conceptual robustness of the knowledge types is considered fallible. It is important to submit these models to some form of empirical testing, as this would empower these models. This would not only give them more explanatory power, but also predictive value. This will sharpen our ideas about the role that knowledge plays in innovation processes. It can therefore help us to better understand innovation processes and, eventually, towards creating an environment in which innovation can optimally bloom.

The goal of the present study is to focus on these shortcomings in our research on knowledge dynamics of innovation and to try to overcome them. We particu-larly focus on the solid theoretical foundation of our knowledge concepts using two pillars: one in cognitive science and one in semiotics. After a general intro-duction of the knowledge discussion in chapter 3, chapter 4 presents our know-ledge model to study knowledge dynamics. A second aim of this study is to seriously and empirically embed our knowledge model, putting it to the test. The remainder of this section characterizes the present study in positioning itself in re-lation to innovation research, and in particular the research related to knowledge.

2.5.1 Improvement of individual knowledge

As we want to understand the ‘source’ of knowledge dynamics we focus on the individual as level of analysis. In studying individuals we can better understand the dynamics in knowledge involved in organizational innovation at a higher level of aggregation. We argue that studying the knowledge of the individual can function as a bridge to understand dynamics of knowledge at the organizational level as well as at the individual level.


44

2.5.2 Innovation

The implementation of decision support from outside

As stated in the introductory chapter focus on the implementation stage of the innovation process it is important to understand organizational innovation. Within this stage of the innovation process we focus on the dynamics of the individuals who are confronted with the innovation, which is not generated in the organization but comes from the outside the organization.

Type of innovation

The innovation involves a process, not a product; a new invented product [software] will be introduced which generates a change in the processes for which the new product is introduced as support. Innovation is considered as at least discontinuous and maybe even really new to the unit of adoption. Although this innovation can also be viewed as a technological innovation – the innovation of the planning support tool – this does not concern this study. The planning support tool that we will study [ZKR, see chapter 5] has already been developed and works satisfactory in many organizations. The innovation involves new ways of working with planning within the organization of Bartiméus [see chapter 7].

2.5.3 Focus on knowledge

And last but not least we want to characterize the present study in that it takes on a knowledge perspective. We specifically focus on the knowledge dynamics during the implementation stage as opposed to a focus on the knowledge capabilities to start implementation.

In the two following chapters we will elaborate on the knowledge aspect of this study. Chapter 4 presents our cognitive-semiotic model on knowledge types. We will use these knowledge types to characterize the dynamics of knowledge during innovation. Chapter 5 focuses on the content, the knowledge domain of planning. These two chapters will be introduced by chapter 3, which discusses the concept of knowledge in general.

45

Chapter 3

Studying Knowledge Dynamics in Organizations

3.1 Introduction

Following up on the previous chapter, this chapter aims to set the first steps to understanding innovation from a knowledge perspective at the individual level. Therefore, we introduce the concept of representation, and we demarcate knowledge in distinguishing data, information and knowledge [3.1]. We then discuss a model that analyzes information flows within an organizational setting, the Information Space model introduced by Boisot [3.2], after which we will discuss the model by Boisot in light of the purpose of the present study [3.3]. We conclude this chapter with a summary [3.4].

3.2 General introduction to knowledge

The philosophical question about the essence of knowledge has puzzled great thinkers from all over the world for centuries. Two prominent philosophers on knowledge are Confucius and Plato. Confucius, an Eastern philosopher born 551 B.C., said that true knowledge is to know what you do not know; he focused on the practical side of life [Heijloo & Eskens 2005]. Plato, a Western philosopher born 427 B.C., takes a different stand in that unambiguous [true] knowledge exists in another world apart from ours and that we only have a dim recollection of the true knowledge from this other world [Bostock 1999]. Confucius and Plato differ in that Confucius focuses on the knowledge obtained in every day life and Plato argues for the true knowledge in a world outside of our own [see Encyclopedia of Philosophy, Edwards 1972, for a more detailed treatment of knowledge in philosophy and science]. So, what do we want to discuss here to introduce the subject of knowledge? We want to clarify how we use the concept of knowledge and our motivation to use it this way in studying knowledge dynamics during innovation.


46

3.2.1 Introducing representation

Individuals have knowledge. The way to study knowledge in humans is in cognitive science defined in terms of representation. Representation is a basic concept in the study of the mind [e.g. Posner 1993; Jorna 1990]. As the next chapter will show, cognitive science and semiotics are two disciplines that use this concept of representation to study the mind; cognitive science focuses on the nature of intelligence [Posner 1993] and semiotics focuses on the use of signs and on meaning [Jorna, van Heusden, & Posner 1993]. In general we can say that a representation is something that 'stands for something else' [e.g. Haberlandt 1994; Jorna 1990]. Johnson-Laird [1993] argues that a mental representation is essential to understanding phenomena of knowledge. Many forms of representations have been distinguished and studied varying from the propositional representation, the pictorial representation, the procedural representation and the declarative representation to representational content and representation processes [Jorna 1990]. All these representations functionally describe what goes on in our minds. For instance, we can frame a concept with a word, or an image with a picture; we can substitute a procedure for a set of rules and so on. Anderson [1990] in fact even argued that a correct representation is essential in problem solving; a representation enables us to describe the human cognition, to study the structure of knowledge, the meaning of a word or a problem that needs to be solved. Thus, representation can be seen as a crucial concept in studying knowledge and the mind in general. We do want to note that the use of representation is not undisputable. For instance, Blackler [1995] stresses that knowledge is too multi-faceted and complex and therefore unrealistic for such a simple approach.

3.2.2 Data, information, and knowledge

The two concepts of data and information are often used to demarcate the concept of knowledge. Although useful, we do not aim to precisely define either data or information. However, we do want to review some examples that illustrate the differences pointed out between the three concepts within KM

[related] literature. Boisot [1998] says that

Knowledge builds on information that is extracted from data.… whereas data can be characterized as a property of things, knowledge is a property of agents predisposing them to act in

Chapter 3 Studying Knowledge Dynamics in Organizations

47

particular circumstances. … Information … establishes a relationship between things and agents [p.12].

Boisot and Canals [2004] refine the above argument into a model in which they argue that the difference between data, information and knowledge is crucial [see also figure 3.1]. They add stimuli as the input for data; perceptual filters convert incoming stimuli to data, conceptual filters convert the data into information. The information in turn becomes knowledge inside the agent. The filters depend on the prior knowledge of the agent. They summarize their differences as follows

Information is an extraction from data that, by modifying the relevant probability distributions, has a capacity to perform useful work on an agent’s knowledge base [2004: 47]

Thus, according to Boisot and Canals, stimuli are the raw material of which data is formed and data is viewed as the raw material for information, which in turn enables knowledge. Information is the intermediary between data and knowledge.

Figure 3.1: The agent-in-the-world [taken from Boisot & Canals 2004: 48]

Zack [1999] formulates a less technical distinction than does Boisot – together with Canals – and says that ‘data represent observations or facts out of context that are, therefore, not directly meaningful. Information results from placing data within some meaningful context, often in the form of a message. Knowledge is that which we come to believe and value on the basis of the meaningfully organized accumulation of information [messages] through experience,


48

communication, or inference. Knowledge can be viewed both as a thing to be stored and manipulated and as a process of simultaneously knowing and acting –that is, applying expertise. [As a practical matter, organizations need to manage knowledge both as object and process.]. Thus, Zack also underlines the more or less cumulative relation between data, information and knowledge and he introduces the role of context. Davenport and Prusak’s [2000] approach is even more practical than that of Zack. They say that data is ‘a set of discrete, objective facts about events’. Information in turn is the ‘data that makes a difference’. And finally knowledge

… is a fluid mix of framed experience, values, contextual information, and expert insight that provides a frame work for evaluating and incorporating new experiences and information. It originates and is applied in the minds of knowers. Inorganizations, it often becomes embedded not only in documents or repositories but also in organizational routines, processes, practices, and norms. [p.5]

The above shows that data is considered the least ‘advanced’ and knowledge the most advanced of the threesome. Then, data is viewed as building blocks for information, it relates to things [rather than people], deals with facts and signs that have not been interpreted [yet]. Information in turn builds on data and bridges this data to knowledge; it has a context, makes a difference and it has been given meaning. Finally, knowledge builds on information and is linked to people, it consists of values, experiences, insights and inferences, and it has a dynamic character, involving processes.

Our approach to data, information and knowledge shows parallels to the above in that they are cumulatively related from data to information and then to know-ledge. We want to emphasize that we view knowledge as dynamic and exclusive-ly generated by people. Data and information on the other hand can be viewed as input for people. The difference between data and information is in the use of a lens. We illustrate this with an example. The actual occurrence of an accident can itself be viewed as data. The people who witnessed the accident, communica-ting what they have seen tends towards information; the ten different persons that witnessed the accident will have ten different versions of what happened and this results in ten different sets of information. Knowledge in this example is the interaction of the information input with what one already knows, the models that one has. Knowledge is a process of the interaction between information and


49

an acting person. This is in line with Boisot who says that the existence of knowledge ‘can only be inferred from the action of agents’ [1998: 12].

In the following section we elaborate on Boisot’s ideas on knowledge dynamics, which form an important inspiration for our cognitive-semiotic model and the accumulation of knowledge discussed in chapter 4.

3.3 The I-Space: A model to understand knowledge dynamics in organizations

The Information Space model or I-Space model enables analysis of information flows within an organizational setting. Therefore, this model is particularly interesting for the purpose of the present study, as our main aim is to understand what factually happens with the knowledge of individuals during organizational innovation.

3.3.1 Introduction

Fundamental in Boisot’s treatment of information [and knowledge] in the I-Space, is his debate with economists. Economists treat information as an asset comparable to other products and goods. According to Boisot this is wrong for two reasons. First, knowledge is intangible and a human [mental] construct, and second, knowledge develops and goes through certain cycles. We will not expound further on Boisot’s discussions with economists in general, but focus on how he treats the dynamics of information [and knowledge].

The growing availability of information underlines the importance of being able to find ones way in the abundance of this information flow [Boisot 1995]. Economizing on data processing becomes vital to effective communication and effective organizational processes. This implies that information needs to be structured in order to become accessible. Boisot stresses that information cannot be treated like any other kind of physical resource, it needs a custom-made economic theory, which one can say most economists do not provide. He uses an evolutionary approach to information and has ‘a primary aim […] to establish information as a resource in its own right, a resource that economists should seriously account for’ [1995]. The dimensions of codification and abstraction are considered to be the key to this information processing. Together with the dimension of diffusion, codification and abstraction form the three dimensions of the I-Space, a conceptual framework that can explore the behavior of information flows to understand the creation and diffusion of knowledge [and by implication


50

innovation] within selected populations [1998: 55]. In other words, the I-Space is a tool to understand the different flows of different kinds of information; it helps to understand the creation and the diffusion of information within groups of people. This applies to organizations, but also to societies. Note that Boisot uses the concepts of information and knowledge interchangeably at this point. However, in his publication of 2003 together with Canals he does differentiate between knowledge and information. In the following we will elaborate on [the dimensions of] the I-Space.

3.3.2 Codification

The first step to data reduction is through codification. Within the I-Space codification forms the first dimension [see also figure 3.2]. The codification process prepares the incoming information for the coding process; a more effective codification process facilitates the coding. Codification is not to be taken lightly. For instance, for this dimension it is important [how] to choose the codes or categories; choosing too many categories is not efficient and choosing too few categories goes with a loss of power in using the codes. The timing of the codification is also important, because ‘once codified, standards often create a lock-in effect that over time become irreversible’ [1998: 45]. So, once you choose it is not easy to reverse this decision. And the gain that comes with codifying in terms of data reduction is paid for with a loss of flexibility. Codification of an experience can, to a certain extent, be described as committing ‘oneself to a particular view of the world’ [Boisot 1983: 163]. Boisot also remarks about codification that it [see also figure 3.2]:

… creates … perceptual and conceptual categories that facilitate the classification of phenomena. The act of assigning phenomena to categories once these have been created is known as coding [1998: 42].

Figure 3.2: The codification process precedes the process of coding


51

In terms of complexity, Boisot defines this dimension of codification as ‘the number of bits of information required to carry out a given data-processing task’ [1998: 46]; the dimension runs from uncodified to codified.

3.3.3 Abstraction

The second dimension within the I-Space is that of abstraction. Abstraction is closely related to codification as it is an extended form of data reduction. In structuring the phenomena that have been codified, the number of categories is again reduced. However, abstraction essentially differs from codification; the process of codification gives form to phenomena, and the process of abstraction structures these phenomena. On the abstraction scale abstract opposes concrete, in which abstract stands for conceptual and non-local knowledge – abstract thought – and concrete stands for perceptual and local knowledge – highly concrete experiences.

3.3.4 Diffusion

The third dimension is covered by diffusion; it establishes ‘the availability of data and information for those who want to use it’ [1998: 52]. Diffusion ‘can be scaled to refer to the proportion of a given population of data-processing agents that can be reached with information operating at different degrees of codification and abstraction’ [1998: 52]. In other words, the diffusion expresses the ratio of a certain population compared to that part of this population that is susceptible to the way that the information is codified and abstracted. This diffusion scale ‘establishes the availability of data and information for those who want to use it. It does not measure adoption: information may be widely diffused and yet remain unused’ [1998: 52]. On the relation between diffusion and effective communication Boisot remarks:

… A shared context is essential to the formulation of meaningful messages; push the requirement of sharing too far, however, and a message becomes banal and uninformative; on the other hand make sharing too tenuous, and a message becomes meaningless. Interesting messages must navigate between intelligibility and banality and effective communication within a social group therefore depends upon a partial, rather than a total diffusion of knowledge and experience among its members [1983: 163-164].


52

So diffusion merely signifies the availability of information; diffusion precedes adoption. In this way Boisot depersonalizes the information; information can be codified and structured by ways of abstraction in which sense it becomes available to a certain population. The actual adoption of that information concerns a next step. We would like to note that, in stating that diffusion precedes adoption, we see a possible discrepancy with the statement that codification can be described as committing oneself ‘to a particular view of the world’ [1983: 163], also quoted on the previous page. We argue that the former statement about diffusion implies a depersonalization of knowledge, whereas the latter statement in particular implies a personalization of knowledge.

3.3.5 The I-Space

The three dimensions of codification, abstraction and diffusion together form the I-Space, a tool to understand the different flows of different kinds of information. Boisot hypothesizes that codification and abstraction together facilitate the diffusion of information and that they reinforce each other. This hypothesis can be visualized in the I-Space [see figure 3.3a below]. For instance, the curve in figure 3.3a moving from point A to point B indicates ‘that the more codified and abstract an item of information becomes, then, other things being equal, the larger the percentage of a given population it will be able to reach in a given period of time’ [1998: 55]. Boisot considers this curve to be static, ‘depicting a function relationship between codification, abstraction, and diffusion at a single instant in time’ [p.58] and remarks that the I-Space can also be used more dynamically.

Figure 3.3a: The diffusion curve in the I-Space


53

This happens when we consider the model to have intrinsic dynamics. In other words, the better information is codified and the more structured that information is, the higher the percentage of the population to which this information will become available. Then, formulated this way, the diffusion of the information is a resultant of the degree of codification and the abstraction of information. But the diffusion also influences the abstraction and codification. As Boisot puts it

… data is … constantly on the move in the I-Space: much uncodified data sooner or later gets codified, much concrete data gradually increases in abstraction, and data that was the proprietary possession of a few individuals gradually becomes the common possession of all [p.58]

The movement can also be in opposite directions on all three dimensions

… codified data over time gets internalized and becomes tacit, abstract data gets applied to concrete problems, and diffused data gives rise to unique insights which are appropriated by well-placed individuals [p.58]

Adding up these forces in movements, Boisot constructs a schematic sequence of movements, which he calls the social learning cycle or SLC shown in figure 3.3bbelow. The figure shows six phases. The cycle starts with scanning fuzzy data and weak signals through which it becomes the possession of individuals [1]. Subsequently these data are codified through problem solving, reducing uncertainty [2], which in turn leads to generalizing the application through abstraction, capturing its essence [3]. The newly created insights are shared with a target-population through diffusion [4]. Next these new insights are applied in different situations in a process of ‘learning-by-doing’ called absorption [5]. Finally, the abstract knowledge is embedded into concrete practices in for instanceorganizational rules, impacting [6]. There are many variations to this schematic knowledge flow.

All in all, the I-Space identifies knowledge flows of information; these ‘pathways’, as Boisot calls them, are shaped by the culture of the population in question, for instance the organization or the industry. The other way around is also possible; the pathways shape the culture of the organization.


54

Figure 3.3b: Movement of knowledge in the I-Space: The social learning cycle [SLC]1: Scanning – 2: Problem solving – 3: Abstraction

4: Diffusion – 5: Absorption – 6: Impacting

3.3.6 Two ways to value knowledge

Boisot distinguishes two perspectives on knowledge: 1] knowledge is cumulative in an absolute sense, and 2] knowledge is cumulative, but confined in a paradigm. In relation to learning and knowledge dynamics the first perspective is referred to as Neoclassical learning or N-learning, and the second perspective is referred to as Schumperterian learning or S-learning.

Neoclassical or N-Learning

The N-learning strategy implies that knowledge is cumulative, gradually building up to a better quality of knowledge. This is accomplished through the elimination of errors. Therefore, knowledge should be codified and abstracted in a precise way, so that a hierarchical structure can be built [p.93].

An implication of this view is that knowledge is valuable in an absolute sense. Sharing knowledge more or less equals giving away what you have without something in return. Boisot hypothesizes that the N-learning perspective will evoke hoarding strategies. That is, the diffusion of knowledge will be obstructed. In the I-Space model this is visualized in that barriers are set to prevent the knowledge flow from ‘B’ to ‘C’ [see figure 3.3c]. N-learning typically holds the view that learning comes to an equilibrium at point ‘C’ in the I-Space, a neoclassical perspective in economics.


55

Figure 3.3c: The four regions activated by an SLC

Schumpeterian or S-Learning

The S-learning strategy also perceives knowledge to be an accumulation, but in a relative way; knowledge building occurs within a paradigm. So, this perspective allows alternative networks of knowledge to either ‘collaborate or compete’ [p.93], a view introduced by Kuhn. This view has an important implication on the essence of knowledge. Boisot [1998] captures this thought in the following citation

These networks are, in effect, patterns that we ourselves impose on the data, and in most cases the data turns out to be consistent with a potentially infinite number of patterns. Networks or paradigms, then, are not inherent in the data and just waiting to be discovered. They are free constructions of the human mind. And just as one can change one’s mind, one can modify the constructions that one overlays on the data of experience. [p.93]

In contrast to the N-learning perspective, Boisot [1998] hypothesizes that S-learning will facilitate a strategy of knowledge sharing rather than a hoarding strategy. In terms of the I-Space model S-learning sees

… the SLC as continuing its course beyond region C in the I-Space, and moving down once more into those uncodified and

A

B

C

D


56

highly local concrete regions … in which disequilibrating discontinuities originate. [p.99]

Knowledge value of individuals

The two perspectives on knowledge can also be distinguished at the level of the individual. That is, whether or not an individual perceives his or her knowledge as absolute will affect the knowledge dynamics of this individual. For instance, when a person acknowledges that new wisdoms can be obtained in a certain knowledge domain then this person has the potential to change her knowledge. Whereas when a person perceives her obtained knowledge as absolute and, moreover, at the end of the development chain, then this person will very likely not have potential to change her knowledge [in this particular domain].

In other words, the perspective that an individual holds on its own knowledge is an important indicator to the potential knowledge dynamics of this person.

3.4 Evaluating the I-Space for studying knowledge dynamics

3.4.1 Recapitulating

The I-Space can be used dynamically to show movement of knowledge flows. These knowledge flows can be captured in a social learning cycle [SLC] and they are an indication of the culture of the population of which the knowledge flows are represented. The I-Space can also be used more statically as a tool to investigate knowledge assets. This can be done through the scaling of the knowledge of for instance a unit or the units within an organization, on the three dimensions of the I-Space. For example, low codification is hard to articulate, high abstraction is generally applicable and medium diffusion is characterized in that the knowledge is only available to a selected group within the whole population.

3.4.2 Understanding knowledge dynamics during innovation using the I-Space

The I-Space model by Boisot emphasizes the unique nature of information – and knowledge – in comparison to other economic goods that should be considered in its own right. It presents an insightful way to communicate the dynamics and value of information. This makes the I-Space model a powerful tool, which shows the dynamics of information at different organizing levels. When we evaluate the I-Space model for our own purpose – to understand the dynamics of knowledge during innovation – we keep in mind the following six points


57

� We study knowledge

� We want to establish the dynamics of knowledge

� We want a tool to study knowledge at the level of the individual

� We want to be able to relate the knowledge and its dynamics to specific tasks

� We want to relate the dynamics of knowledge to the dynamics of the innovation process

� We want a solid foundation for our knowledge perspective, preferably in cognitive science and semiotics to secure a focus on theory and empirical study

When we take these points into account as formulated above and use them to evaluate the I-Space model of Boisot, we come to the following evaluation [see also table 3.1].

The I-Space of Boisot is particularly interesting for our purposes as the I-Space enables exploration of knowledge dynamics within organizations, of whichorganizational innovation is an example. Also, the particular kinds of knowledge can be directly related to specific tasks that are performed within the organization. The I-Space can reveal dynamic as well as static knowledge, although we note that Boisot uses the concept of information and knowledge interchangeably. The possibilities of our model to measure knowledge [dynamics] equal those of the I-Space model, although Boisot does not empirically test his model.

We do however miss an important ingredient in the I-Space model. The aim of the present study is not to understand how knowledge moves within a certain population. Rather, we are interested to know what actually happens to the knowledge within a person who undergoes organizational change in the form of organizational innovation in relation to the type of knowledge; codification and abstraction can be viewed as knowledge types, but diffusion cannot. As Boisot pointed out, codifying involves mastery of codifying skills including ‘an apprecia-tion of how they [codifications] attach to a specific and narrow range of expe-riences’ [1983: 166, between brackets not in original]. Boisot uses one scale for codification, moving from non-codified at one end to codified at the other end. We see the process of codification, from a semiotic perspective which the next chapter will elaborate on, as essentially different from the non-codified skills. We are also interested in the process of non-coding to coding. The I-Space model of


58

Boisot therefore is not suited to fit our purpose in all respects. The next chapter presents our cognitive-semiotic model as an alternative to study organizational innovation from a knowledge perspective focusing on the individual level. So, what is missing is the possibility to study knowledge at the individual level.

Table 3.1: Evaluation of the I-Space model by Boisot in comparison with the adjusted cognitive-semiotic model in terms of the aim of our study

Aim Boisot Adjusted c-s model

Knowledge dynamics - / + +

Measurable + +

Empirical - / + +

Individual level - +

59

Chapter 4

A Cognitive-Semiotic Model to Capture Knowledge

Dynamics at the Individual Level

4.1 Introduction

The previous chapter discussed the I-Space of Boisot in relation to the aim of the present study to understand what actually happens to the knowledge of people who undergo organizational change. Although the I-Space model puts forward some intriguing tools to identify and understand knowledge dynamics within organizations, we argued that, for our purpose, the I-Space did not suffice [see also table 3.1]. The I-Space takes on an organizational perspective and we argue for an individual, cognitive-semiotic perspective for understanding knowledge dynamics. Whereas Boisot discusses the potential of the diffusion of certain knowledge the model that we will present aims to capture the essence of knowledge itself and the dynamics of that knowledge. In other words, Boisot focuses on the dynamics of certain knowledge within a population, whereas the present study focuses on the dynamics of knowledge within a person – or an organization when the knowledge of the individuals is added up – changing the knowledge, not its degree of diffusion. As we shall see in the present chapter, this results in swapping the diffusion dimension for a dimension on sensory knowledge. This choice is founded in the cognitive semiotic pillars of our model.

Leading up to the presentation of our cognitive-semiotic model, we discuss its two foundations in cognitive science and in semiotics [4.1]. We present the three dimensions of our model, namely the sensory dimension, the coded dimension and the theoretical dimension [4.2]. We then add the dynamic aspect to these three dimensions resulting in a new space, the knowledge space or K-space [4.3].

4.2 A cognitive-semiotic approach to knowledge

In understanding the knowledge dynamics of organizational innovation, in particular at the level of the individual, it seems reasonable to take a cognitive perspective; cognition is often described as information processing. But cognition does not take into account the nature of the signs and symbols that underlie


60

discussions on mental representation, while semiotics, as the study of symbol and sign structures and interpretation, does. In combining cognitive science and semiotics – a combination to which some refer as psychosemiotics [Posner, Robering, & Sebeok 2005] – a very powerful team arises in understanding knowledge dynamics [van Heusden & Jorna 2001], reasoning [Shank 1998], learning of humans [Cunningham 1998; Smith 2000] and studying the mind in general [Smythe & Jorna 1998; Bouissac 1998]. This team is no coincidence. In fact, they share historic roots [Bouissac 1998].

Whereas the cognitive side focuses more on the mental processes and structures, enabling solid empirical research, the semiotic side is fundamentally based on the symbol and sign structure approach with an emphasis on semiosis [sign understanding, the concept coined by Peirce]. The common interest that these two disciplines share is that both have signals, signs and symbols as a central concern [Michon, Jackson, and Jorna 2003]; they differ in the fact that semiotics does, and cognitive science does not actually explain the nature of the sign. For pragmatic reasons we will refer to this common interest as representations [Jorna, 1990; van Heusden 1994: 30]. The combination of a cognitive approach and a semiotic approach to representations forms the foundation of our cognitive-semiotic model. We argue that our model can account for the dynamics of knowledge in innovation processes, in particular the dynamics in knowledge that the individual within the innovation process undergoes. In the following section we elaborate on the two fields of cognitive science and semiotics separately.

4.2.1 Cognitive science

Cognitive science is a relatively new discipline; it is the factor that connects disciplines that deal with cognition. Besides cognitive psychology and experimental psychology the roots of cognitive science can be found in linguistics, artificial intelligence, neuroscience, epistemology, logic [Michon, Jackson, & Jorna 2003], cognitive neurology, and philosophy [Nersessian 1992]. One of the typical issues within cognitive science in concerned with how knowledge is represented [Simon and Kaplan 1993], for example procedural knowledge and declarative knowledge [Jorna 1990]. Among the methods that are used in cognitive science to gather data are protocol analysis, content analysis, AI and computer simulation, meta-analysis, philosophy, logic, and semantics, but also experiments and physiology. Pylyshyn [1993] argues that the key characteristic in which cognitive science differs from the more traditional cognitive psychology relates to its influences ‘by both the ideas and the techniques of computing’ [1993: 51], which are used for

Chapter 4 A Cognitive-Semiotic Model to Capture Knowledge Dynamics

61

the ‘imitation of certain unobservable internal processes’ [1993: 53]. Simon and Kaplan describe cognitive science as ‘the study of intelligence and intelligent systems, with particular reference to intelligent behavior as computation’ [1993:1]. Representations can play a key role; cognitive science studies how they are formed and changed.

Simon and Kaplan [1993] appoint intelligent behavior to be the core subject of cognitive science; it ‘is adaptive and hence must take on strikingly different forms in different environments’ [Simon 1980 in Simon and Kaplan 1993: 38]. Thus, Simon argues that human cognition is dynamic and when used this cognition changes. Simon continues to say that

On a longer time scale intelligent systems make adaptations that are preserved and remain available for meeting new situations. They learn. There are many forms of learning. One important form is the accumulation of information in memories and the acquisition of access routes for retrieving it. Learning changes systems semi-permanently and hence increases the difficulty of searching out invariants [Simon 1980 in Simon and Kaplan 1993: 38; italics not in original].

The above implies that the ability to adapt is an important aspect of intelligence and human cognition, accumulation being an important form of adaptation. This accumulation enables more complex forms of cognitive activity.

In trying to explain and predict human intelligent behavior three levels [stances, Dennett 1971/1978] can be distinguished: the design stance, the physical stance and the intentional stance. Dennett [1971] illustrates these three stances in using the example of a chess-playing computer. When we play a chess-playing computer we can adopt different strategies to predict the moves of our opponent. The first strategy involves the design stance. In using this strategy we rely on our knowledge of how the computer is designed, including how it is programmed. Knowing this we will be able to foresee the responses to any move that we make just by following the computational instructions of the program. This stance relies on the notion of function. That is, ‘a design of a system breaks it up into larger or smaller functional parts, and design-stance predictions are generated by assuming that each functional part will function properly’ [p.88]. Dennett notes that design-stance predictions can be made at several different levels of abstraction. Essentially, using the design-stance we ‘make predictions solely from knowledge or assumptions about the system’s functional design,


62

irrespective of the physical constitution or condition of the innards of the particular object’ [p.88].

The second stance is the physical stance. This stance bases its predictions on ‘the actual state of the particular object’ [p.88]. This is worked out by ‘applying whatever knowledge we have of the laws of nature’ [p.88]. This stance can predict malfunctions of the system.

The third stance is the intentional stance. This strategy is used when the computer is viewed as an intentional system. In this case ‘one predicts behavior in such a case by ascribing to the system the possession of certain information and by supposing it to be directed by certain goals, and then be working out the most reasonable or appropriate action on the basis of these ascriptions and suppositions’ [p.90, italics in original]. In explaining and predicting cognitive behavior our interest is at the level of the design stance. In the following section, which discusses the semiotic foundation to this study, the functional level of behavior will be examined, as we will see, distinguishing three types of knowledge.

In sum, cognitive science has its foundations in a variety of disciplines. It deals with intelligent behavior, such as setting adequate goals in a coherent and appropriate way and choosing relevant actions accordingly. Essentially cognitive science can be said to be the study of representations and its underlying neuronal and physiological structure. Important is the characteristic of adaptation and the accumulation principle, which is the study of the changing representations. Within cognitive science three strategy levels to explain and predict behavior can be distinguished, the design level, the physical level and the intentional level. And last but not least, cognitive science has a number of methods available to do empirical research. However, while cognitive science does not concern itself with the nature of the representation of intelligent behavior, the discipline of our second foundation, semiotics, does.

4.2.2 Semiotics: A sign process of semiosis in three dimensions

Semiotics has been described as ‘the study of the innate capacity of human beings to produce and understand [the nature of] signs of all kinds’ [between brackets not in original], but also as “the study of ‘semiosis’ or sign action, … how signs are interpreted and used to effect various ends” [Sebeok 1994 in Smythe and Chow 1998: 783]. In short, semiotics is the theory of signs. The American Peirce and the Swiss de Saussure laid the foundations for it. As modern semiotics is mainly influenced by Peirce [van der Lubbe & van Zoest 1997] we take his work as a


63

starting point to introduce the discipline of semiotics. Peirce argued that all knowledge is mediated, that is, there is no direct relationship between the object and the knowledge of this object [Peirce 1868 in Schuyt 1997: 156]. For instance, there is no direct relation between a book and what we know about this book. These two are mediated by a sign, the sign that we have for the book. The process of knowledge – just as every form of thinking and every form of communication – involves signs. Peirce explains this mediated relation by distinguishing three related elements in the process of thinking [van der Lubbe & van Zoest 1997]

1. the sign [or representamen van Heusden 1994] that can be perceived [the word ‘book’]

2. that to which the sign refers, the object or referent [the book itself] and

3. what is evoked in the person when the sign is perceived, the interpretant [what we know about the concept of book] – we will refer to this element as meaning

In other words, the process of thinking essentially distinguishes these three elements and together these three elements form the essence of the semiotic process. As we will see every element that is represented in our cognitive system – in the form of a semiotic sign – is made up of these three semiotic elements and each element adds a dimension to the semiotic sign. In using an example, based on van Heusden and Jorna [2001], the added value of each semiotic element becomes more tangible. This example shows how new experiences are perceived and represented in the mind, from a semiotic perspective.

The first semiotic dimension: Perception of difference

The semiotic process starts with perceiving a difference. But how does this perception come about? When we do not perceive a difference, what we experience [in actuality] is precisely the same as the experiences that we have had before [stored in our memory]. In other words, the things that we perceive are already known to us, e.g. we see a book and we have seen books before. Thus, we constantly compare what we perceive, the world around us, to what we already have experienced, in our mind. When this comparison does not add up to a difference we are still in a non-semiotic situation; memory and actuality are one and the same [Jorna & van Heusden 1999]. But now we come across something that we have not experienced [perceived] before, so our experience


64

and perception [actuality] differ from our memory [what is already represented in our minds], the semiotic process is triggered [van Heusden & Jorna 2001].

To illustrate the process just described, let us assume that we are at a Christmas party eating a dish consisting of different kinds of fruits. While tasting the different pieces of fruit that this new dish contains, we notice a piece of fruit with a taste unknown to us. So, we perceive a difference between what we experience [tasting the piece of fruit unknown to us] and our memory [all the types of fruit that we have tasted before in our lives]. This discrepancy is what makes us notice the taste of the new fruit. This means that, although the taste of this new fruit will always be experienced in comparison to the fruits that we experienced before, to start the semiotic process it is crucial that we not only perceive differences, but that we also perceive similarities between actuality and memory. This combina-tion of differences and similarities is what sets off the semiotic process. Thus, we search our memory for experiences to compare to the new experience; these experiences form a framework or reference point for the new experience. Hence, when there is no common ground between the new experience and the 'old' experiences, then it becomes impossible to give this new experience a perspective making it impossible to comprehend or grasp. This becomes particularly apparent when you enter an unfamiliar discipline. For example unfamiliar with physics you come across a very complicated formula; then, you do perceive differences, but there are too many to start the semiotic process.

Returning to the Christmas party, we compare the new fruit to a different, but in a way, similar fruit that we know. Let us assume that the new fruit in fact is a piece of mango. Therefore, a similar fruit to which we can compare this new piece of fruit could be a pear. Similarities between our experiences with pears and our new experience with the new fruit [e.g. both soft and juicy fruits] together form the basis of a concept for the new fruit. The semiotic process is triggered and therefore the creation of new knowledge is started.

In perceiving the new piece of fruit [mango] in comparison to a familiar piece of fruit [pear], the familiar pear functions as a reference point to the new mango. Thus, the new experience is perceived and interpreted based on similarities between the two fruits, the old fruit [the pear] and the new fruit [the mango]. However, perceiving difference is what starts the most primitive form of the semiotic sign, the perceptual, one-dimensional sign is formed. At this first stage in the semiotic process the newly formed concept of the mango is still dependent on the presence of a context; we need the mango to be present to be able to apply our knowledge of this new concept.


65

From a semiotic perspective, using the three elements of sign, object and meaning, the above example of encountering a new fruit can be captured in terms of the physical, tangible mango to which the word 'mango' refers [object], the word 'mango' [sign], and our interpretation, the essence of the mango fruit [meaning]. In the first stage of the semiotic process these three aspects object, sign, and meaning are still one-dimensional, they are one and the same [see figure 4.1a]. That is, the new knowledge of the mango can only be of use in the presence the new piece of fruit [object]. The mango fruit does not have a sign yet [e.g. the word mango]. Furthermore, we cannot yet verbalize or make our interpretation of this new fruit tangible, that which is specific for the mango fruit. We cannot verbalize this because we do not [yet] know in what way the new fruit essentially differs from other fruit [meaning]; we cannot pinpoint the difference. Thus, both meaning and sign are still directly related to the object of the new fruit; the three aspects of the semiotic sign are one. As stated in the above, as no earlier experience with this type of fruit is available, no representation for this new fruit is available; the object has no representation yet.

= MANGO =What there is to know

about a mango

Objectthe actual mango

= Signthe word mango

= Meaningthe meaning of the mango

concept

Figure 4.1a: The first semiotic dimension in which object, sign and meaning are one

The second semiotic dimension: The need for communication

Returning once again to the Christmas party, the person to whom you are talking happens to know the new fruit that you have just tasted; he says that this fruit is called 'mango'. In naming the new fruit we have taken the second step in the semiotic process; the semiotic sign has become a two-dimensional sign. That is, object [the actual mango] has become separated from the sign [the word mango that we use to refer to the object mango]. Now that we have named the new fruit, this second stage in the semiotic process enables reference to the new fruit without its actual presence. Hence, we can talk about a mango without the


66

mango being present. The sign has become a code that we use to refer to the object; the word mango functions as a code to refer to the object mango. This second dimension opens up possibilities for communication. We can replace the experiences with the new fruit mango with a code. The code is an arbitrary phrase, word or sound for our experiences. As the code [mango] replaces our experiences with the mango fruit it is important to keep in mind that the second dimension in the semiotic process presupposes the first. We cannot use a code, without the experiences that form the foundations for that code. As different people experience the same fruit in a different ways, this implies that the [same] code that people use will be used in different ways. The more abstract the concept to which the code refers the more likely this will be the case; this is often a source of miscommunication.

Thus, in terms of the three semiotic elements of object, sign, and meaning, the second dimension separates object [actuality] from sign [code] and meaning; the sign has become two-dimensional [see figure 4.1b]. We now can refer to the object mango by using the code mango. For communication this implies that other people should also have knowledge of this code ‘mango’ [as referring to the fruit]. As the code is arbitrarily chosen it should be clear that a code is therefore restricted to a limited group of people. A code is only of use when the communicating parties use the code in the same way. For instance, even though the Swedish word barn [meaning children] and the English word barn [meaning a storage] have the same spelling and they are pronounced in the same way [the code is the same] their meanings do differ. Only people can attach meaning to a code. A code loses its meaning when it is used in communication with others of outside the group [Boisot 1983: 163].

� MANGO =What there is to know

about a mango


� Signthe word mango

= Meaningthe meaning of the mango

concept

Figure 4.1b: The second semiotic dimension, in which the code separates from actuality


67

Returning to the mango example, the people that use the word [code] mango must have experience with this particular fruit. They must know to what object the word mango refers in order to communicate effectively [without the fruit being present]. We could say that the second semiotic dimension comes into play when a need for communication concerning the experience with the new fruit mango emerges. Therefore, the essence of the second dimension partly lies in sharing and communicating. Would there be no need for communication, then neither would it be likely that a need for the second dimension would emerge.

The third semiotic dimension: Understanding the essence in relation to other

concepts

Let us return to the Christmas party for one last time. Joining the mango-group is a woman who happens to be an expert on mangos. She explains the origin of this fruit, its physical appearance and so on. She explains the essence of a mango, that what distinguishes a mango from other kinds of fruit. With this expertise it becomes possible to judge individual cases of fruit on their 'mango-ness'. This is in essence the third and last semiotic dimension. When we actually understand what it is that makes a mango a ‘mango’, the second dimension makes way for the third dimension. This last dimension separates sign, object, and meaning into three different parts of a whole; thus 'meaning' also becomes separated [see figure 4.1c]. Therefore, we now have the mango as an object, the word mango as a code to refer to the mango object, and conceptual content that comes with the code [and therefore also the object]; the sign has become three-dimensional.

� MANGO �What there is to know

about a mango


� Signthe word mango

� Meaningthe meaning of the mango

concept

Figure 4.1c: The third semiotic dimension, in which meaning separates from the code

The third dimension adds coherence and integration, it structures. Whereas the first dimension relies on a process of transformation [transforming the bundled


68

experiences with the mango fruit], and the second dimension relies on a process of substitution [replacing the actual fruit for the word mango], the third dimension relies on the analysis of relations or structures [van Heusden & Jorna, 2001]. The third dimension replaces the arbitrariness of [the code in] the second dimension. Hence, as we recall, the use of codes is restricted to a limited group of people. Therefore, to be able to enter the third dimension, one needs to be able to use the codes accurately as the third dimension basically structures the codes from the second dimension. So the third dimension is not accessible to just anyone. The third dimension is not based on convention, as is the second dimension, but it is based on coherence and structural qualities. In the mango example this third dimension reflects the structural differences between the mango and the pear. For example, whereas the pear has several pits inside the mango only has one big pit. Knowing this we no longer merely have to rely on experiences; we can now make a judgment based on logic and reason. When we can only judge based on experience, we might deduce from these experiences that the color of pears is just green and that the color of mangos is a mixture of colors such as yellow-orange-green. When we come across a green mango based on this judgment rule we should reason that this green mango is a pear. However, this is not correct as we came across a green colored mango. Therefore, only having a judgment rule at your disposal is not sufficient to correctly judge the given situation. But when we, besides the descriptive qualities, also have the structural qualities of the mango fruit at our disposal, then we can make an accurate judgment. When we know about the structural quality of the mango, for instance only one big pit, then we can reason that this alleged pear actually is a mango [note the similarity with possible idiosyncratic feedback during the innovation process as a possible threat to the innovation process, e.g. Leonard-Barton 1995].

So, the third dimension provides coherence and structure. Note that this third dimension is cumulative to the second as the second is to the first. Therefore, this third dimension also presupposes the first dimension: no structure or logic without codes or perception. The third dimension structures the codes of the second dimension. However, theoretical knowledge is not coded by definition [van Heusden & Jorna 2001].

In sum, when we experience something that we have never experienced before we perceive a difference, given certain similarities between the actual and memory; this perceived difference is what triggers the semiotic process. Then, presupposing the first, substitution takes place: the new experience is categorized


69

as a specific experience. Finally, a new type of concept is added to the categories of the second dimension; these concepts no longer refer to classes of objects, but to structural relations such as 'number' and 'cause' and 'effect'. We can now signify the semiotic process in terms object, sign, and meaning [van Heusden & Jorna 2001]. We want to note at this point that the three semiotic dimensions also parallel the differences between data, information and knowledge, in particular argued by Boisot and Canals [2004] emphasizing the role of the perceptual filter and the conceptual filter. The three semiotic dimensions, together with the empirical focus of cognitive science, will function as the foundation for our cognitive semiotic approach to knowledge.

We argue that the semiotic perspective, as discussed in the above, will function as a firm foundation in understanding organizational innovation. In distinguishing three types of knowledge parallel to the semiotic dimensions we can gain understanding of how knowledge types are mutually related and of their dynamic character.

In combining the cognitive framework with the semiotic framework, we have a solid foundation for three different types of knowledge. Van Heusden and Jorna [2001] capture this line of thought in the following way:

Our knowledge of reality is a semiotic, that is, a representational construction. Representational activities can be studied, both from a cognitive [focusing on the empirical reality of the behavior] and from an interpretative [focusing on the meaning of the behavior] perspective. As [cognitive] mental activity, this representational behavior is studied by cognitive science. As representational activity, however, the cognitive mental behavior is the object of semiotics [p.84].

Understanding knowledge from a cognitive-semiotic approach implies that the acquisition of knowledge involves three stages. In the following section we will discuss these three stages and relate them to three types of knowledge, namely sensory knowledge, coded knowledge, and theoretical knowledge.

4.3 Combining cognitive science and semiotics: Three stages in acquiring knowledge

In the following section we will present our cognitive semiotic model. Therefore, we will discuss separately each of the three dimensions of which this model


70

consists, namely sensory knowledge, coded knowledge, and theoretical knowledge. As discussed earlier, the model to be presented was greatly influenced by the I-Space model of Boisot. This model also contained three dimensions, the codification dimension, the abstraction dimension, and the diffusion dimension. The codification dimension shows great parallels to our coded knowledge dimension and the abstraction dimension shows great parallels to our theoretical knowledge dimension. However, the I-Space model differs from our cognitive semiotic model on the third dimension; whereas Boisot introduces the diffusion dimension, which is, more or less, a resultant of the codification dimension and the abstraction dimension, we introduce the sensory knowledge dimension, which precedes the coded knowledge dimension.

4.3.1 Sensory knowledge

Sensory knowledge is the first type that we distinguish in our cognitive-semiotic model, based on the first semiotic dimension. As the name reveals this type of knowledge is essentially sensory and it is based on perception. Based on perception, this type of knowledge precedes the stage of coding. Note a parallel with Boisot and Canals [2004] who refer to a perceptual filter that sorts out the incoming stimuli into data as the foundation for knowledge.

Although all knowledge is personal, sensory knowledge is personal in the sense that it can only be applied. There is no mediating sign; sensory knowledge is completely dependent on its context. Examples of this type of knowledge: are, the taste of mango, the act of actually riding a bicycle, or a sense of the existence of different roles within an organization, which could be translated to the concept of ‘hierarchical structures’; these examples can all be referred to as sensory knowledge.

As this type of knowledge is not mediated by a sign, sensory knowledge can only be transmitted through imitation; in an organizational setting this could be translated to on-the-job training. Through sharing the same experience different individuals build up the same [or at least very similar] sensory knowledge based on common experience. For instance, in teaching a novice how to negotiate within an organizational setting a person with many years of job experience can show the novice how she, as an expert, negotiates. This way the novice can experience the process of negotiating. So, the novice will share the same experience with the experienced person, or rather these two individuals will experience the same situation although both with a unique frame of reference. This does not have to be a problem per sé, just as long as these two individuals


71

agree on the experience as a reference. Boisot [1983] remarks that uncodified knowledge – which shares the ‘pre-codedness’ with sensory knowledge – is difficult to transmit. According to him

Ambiguities abound and can only be overcome when communi-cations take place in face-to-face situations. Errors of interpre-tation can then be corrected by a prompt use of personal feed-back. Consider the apprenticeship system as an example. [p. 164]

The essence of on-the-job training is the possibility of instant feedback and calibration. This opens the possibility to highlight certain aspects of the sensory knowledge, decomposing it in a sense. So in on-the-job training the expert essentially provides the novice with a model to perceive differences. But on-the-job training also helps the expert to understand the frame of reference of the novice. This is important as the frame of reference determines the vigilance of a person. A way to highlight certain aspects of the coded knowledge is the use of metaphors, analogies and examples; these can function as a mutual reference point in the abundance of stimuli. Although the use of metaphors can be very useful, one should always keep in mind that a metaphor cannot function as a substitute. For instance, an organization is not actually an organism with all the characteristics ascribed to an organism. The purpose of using a metaphor lies in communicating some sense of sensory knowledge through highlighting some of its characteristics in using a metaphor.

As sensory knowledge is transferred from person to person the diffusion of this kind of knowledge is very limited. This puts restrictions on the time span in which this knowledge can be transferred; sensory knowledge takes a great amount of time to diffuse over a whole population.

A more complex example of sensory knowledge is the sensory knowledge of a more abstract concept such as ‘hierarchy’. One might sense differences between the statuses of different persons within an organization through, for instance, the way that these persons are spoken to. This sensing precedes being able to exactly frame the differences in the way that these persons are spoken to. This sensing refers to the sensory knowledge about the concept ‘hierarchy’. Note that in using an example to explain the essence of sensory knowledge we in fact use the very same means as described above, that is, we use this example to communicate sensory knowledge [building up of a mutual frame of reference].


72

When we view sensory knowledge itself more closely we can distinguish a variation in the degree of detail, from rough sensory knowledge [e.g. general knowledge of sports, such as throwing a ball] to detailed sensory knowledge [e.g. knowledge in top sports]. We need this refinement in sensory knowledge to express the great variation between the first sense of difference and the very detailed sense of difference. The refinement of sensory knowledge is also a process that is triggered by perceiving differences. For instance, when we are refining our skills in swimming we can only do this when we perceive that a certain way of swimming is better or more comfortable, only then will we choose to stick to this particular way of swimming.

We can capture the essence of sensory knowledge by summing up its characteristics and their implications. Sensory knowledge is based on perception and it precedes the coding stage of knowledge. It cannot be verbalized, only applied. Use of metaphors, analogies and examples often captures both similarities and crucial differences in a certain situation. Transfer of this type of knowledge can only be from person to person by ways of imitation and on-the-job-training, which ensures instant feedback and opportunity for calibration of skills. Finally, sensory knowledge is very dependent on its context, it diffuses slowly and it is restricted in time. In sum, one might say that sensory knowledge is knowledge that works in practice [van Heusden & Jorna 2001].

4.3.2 Coded knowledge

The second knowledge type in our cognitive-semiotic model is based on the second semiotic dimension. We refer to this type of knowledge as coded knowledge, as this dimension introduces the use of codes to which an object or an experience refers. At this point we want to emphasize the difference withBoisot’s model of the I-Space. Boisot places both sensory knowledge and coded knowledge on one and the same scale, the dimension of codification. He emphasizes that for the process of codification

… the learning required involves a gradual mastery of the codification skills themselves and an appreciation of how they attach to a specific and narrow range of experiences [Boisot 1983: 166]

So, although Boisot acknowledges that one requires a certain skill to be able to code, he does not frame this acknowledgement by using a separate dimension. As we saw in the semiotic process, the step from the first dimension to the


73

second dimension presents a crucial difference. To emphasize this crucial difference we therefore split Boisot’s dimension of codification up into two dimensions, two different types of knowledge, based on the semiotic process.

Coded knowledge introduces the possibility to communicate knowledge without the presence of that to which this knowledge refers. For instance, we can talk about an apple without its actual presence. This essential difference with sensory knowledge lies in the fact that coded knowledge does not have to be applied –through physical action – in order to be communicated. Thus, the introduction of a code makes it possible to refer to certain knowledge. Therefore the code can be viewed as an activation key for coded knowledge. We might say that the [coded] knowledge has been tagged by a code. For instance, the object mango has been labeled as the code ‘mango’. So when we talk about a mango we expect others to know to what we refer. And the difference sensed between the people in the organization can be labeled with the code ‘hierarchy’. So when we talk about a hierarchical organization we expect people to know that this, among other things, means that the different people within that organization have a different status. Note that, although we argue that a code opens up the possibility of communication, the actual process is likely to be the other way around. That is, the need for communication forces the introduction of codes.

At this point we want to stress that the code represents certain knowledge. We make no claim that the code is the equivalent of the knowledge that it represents. Rather, the code is the key to this coded knowledge. The code can be detached from the messenger, although the knowledge for which this code stands will always need a messenger in order to have content. Therefore the knowledge is not applied through specific action, as is sensory knowledge. Rather, a person can communicate coded knowledge simply through referring to a particular code. Using codes therefore opens a great many possibilities for communication; these possibilities are absent for sensory knowledge. Boisot [1998] makes an essential remark for the procedure for codification, as he calls it, in saying that it is a risky business. He argues that if

… the strategy is faulty, the wrong data may be selected and valuable data discarded. In an organizational context, however, data discarded by the decision-making apparatus may actually be retained in an implicit form, embedded in the memory of individuals involved in the selection process’ [p. 45]


74

This is an important aspect of coded knowledge that we should always keep in mind in communication. The power that a code gains in possibilities of communication is at the same time its weakness. Especially in using complicated codes for communication we should always be aware of differences in the ‘data-reduction’ process, that is, the knowledge for which the code stands.

When we compare sensory knowledge to coded knowledge the first difference is their foundation. Whereas sensory knowledge is founded in perception [of differences], coded knowledge is founded in sensory knowledge and therefore only indirectly founded in perception. Thus coded knowledge is not personal in the way that sensory knowledge is [but we continue to stress that knowledge always has a direct link to people]. This has implications for some of the other characteristics. Whereas sensory knowledge can only be exercised by applying it, coded knowledge can be used in different ways. For instance, we can write it down. This makes coded knowledge less context dependent than sensory knowledge in two ways. Firstly, in referring to something with a code, such asreferring to the mango, we do not need the actual mango present. In this way coded knowledge is independent of the context. Secondly, in writing down the word mango in a letter, we can communicate with someone without this person being present. So coded knowledge allows knowledge transfer without either the object of reference or the party with whom we are communicating present. This makes diffusion of coded knowledge less restricted than that of sensory knowledge, and it also takes a shorter amount of time.

There are, however, still restrictions in the use of coded knowledge as the parties who communicate the coded knowledge should all be aware of the knowledge on which the codes are based. This implies that codes can only be used in a limited group, a community. But also, one and the same code can be interpreted differently within different communities. Communication is often obstructed in that the codes that are used represent different knowledge for the different people involved in the communication process. This rivalry can be a troublesome source of miscommunication. Furthermore, we expect that this rivalry could potentially be an important inhibitory factor in innovation process. So, ‘a shared context is essential to the formulation of meaningful messages [Boisot 1983: 163]. With the increased ease of communication and diffusion the chance of miscommunication also increases, especially when the coded knowledge is not embedded in sensory knowledge, but merely copied as a code. Finally the quality of coded knowledge is expressed in the degree of ambiguity – the lower the ambiguity the higher the quality of the code and the more powerful the code.


75

Codes such as the alphabet notations or musical scores are less ambiguous than codes such as icons or pictograms [see Goodman, 1981; Jorna 1990].

4.3.3 Theoretical knowledge

The third and last knowledge type that we distinguish in our cognitive-semiotic model of knowledge is based on the third semiotic dimension. Whereas sensory knowledge identifies and coded knowledge defines knowledge, this third knowledge type structures, it puts the knowledge in perspective; therefore, we will refer to this knowledge type as theoretical knowledge. We can compare this type of knowledge to the abstraction-dimension of Boisot.

Theoretical knowledge refers to knowing the essence of a concept [object, or event]. Essentially understanding a concept implies that one can also determine and analyze the relations of this concept to other concepts. This understanding is therefore more than just defining the concept, which is covered by coded knowledge. Theoretical knowledge essentially structures, it structures the coded knowledge and it therefore presupposes coded knowledge. Note that coding is not the essence of theoretical knowledge; it can be coded, but it does not have to be coded [van Heusden & Jorna 2001]. Theoretical knowledge is typically obtained in education, as education emphasizes the structure and reason. Education often highly depends on reasoning and understanding how different things are related, it offers a model to structure.

Theoretical knowledge explains; it provides the answer to the 'why-questions'. This does not mean that theoretical knowledge pretends to be true or correct. Keep in mind that argumentation valid to some may be invalid to others. McCloskey [1983] points to the use of naïve theories [of motion]. He distinguishes three ways in which everyday experience can lead to knowledge [about motion] and the naïve theory is one of these ways. He shows that these models [the naïve theories], although constructed individually, can be well articulated and show consistency over different individuals. The naïve theories studied by McCloskey related to knowledge of physics. And interestingly, although similarities were found between individuals these theories showed great inconsistency ‘with the fundamental principles of classical physics’ [p.299], showing great resemblance to the pre-Newtonian physical theory. An important aspect to note about these naïve theories is that they can be very persistent and that they can lead to misinterpretation of information. So, these naïve theories can stand in the way of forming new [coded] knowledge. In the McCloskey study the students tried to fit the information that they received into their naïve [and incorrect]


76

theory. It is less clear how these naïve theories develop or what role they play in everyday life. Clement [1983], who uses the term intuitive model instead of naïve theory, shows that mere exposure to an established theory will not eliminate the intuitive model and stresses the importance of devoting ‘more attention to conceptual primitives at the qualitative level’ [p.335]. This is in line with the foundation of our theoretical knowledge, the third semiotic dimension, namely knowing the essence of concepts leads to structure, not the other way around. Clement seeks argumentation for their persistence in the following line of thought

Preconceptions need not be viewed exclusively as obstacles to learning, however, they constitute micro-theories that students have constructed on their own, and should be respected as such. Because they ordinarily have some predictive power in certain practical situations, they can be thought of as “zeroth-order models” which the students possess. Some preconceptions can be built upon or modified by students in order to increase the precision and generality of their theories [p.335]

To the above statement we want to add that the persistence of the naïve theo-ries can be explained by the fact that they are well imbedded in the knowledge of the users. In other words, the structure of these theories fits within the know-ledge network of the users. So, not only are the intuitive theories well-structured micro-models, they also fit well into the existing knowledge of the persons.

The above citation also emphasizes the importance of having a model in perceiving and weighing information. Thus, theoretical knowledge also provides underpinnings for choosing one alternative instead of other alternatives. As aconsequence theoretical knowledge provides the possibility to take different points of view, putting things into perspective. That is, structures are multi-dimensional as opposed to coded knowledge. This means in practice that we can choose between different alternatives based on a certain coherence and reasoning pattern. In addition to the explanatory power of theoretical knowledge, it also has a predictive power. It provides a means to conduct an experiment of thought, whereas coded knowledge does not.

Boisot [1983] notes that when codification has not yet been established the process ‘remains tentative and exploratory and can only be understood on an intuitive level’ [p.166]. Johnson-Laird [2006] refers to intuitions as a possible beginning of reason: rapid and often accurate.


77

As theoretical knowledge structures and introduces coherence, it essentially cre-ates its own context. Therefore, we argue that theoretical knowledge is not de-pendent on a context [compared to sensory knowledge and coded knowledge]. This implies that the diffusion of theoretical knowledge is even easier than coded knowledge. However, the entrance level is higher than that for coded knowledge.

Like sensory knowledge and coded knowledge, theoretical knowledge can also vary in quality, namely in its degree of abstractness. Very abstract forms of theoretical knowledge tend to be universal. Concrete theoretical knowledge is what we use in our every day lives to make sense of things [compare the naïve theories]. The degree of abstractness of theoretical knowledge is therefore related to the applicability of the knowledge; the more abstract the theoretical knowledge becomes the higher the explanatory value and the higher the predictive value. Boisot [1998] refers in his dimension of abstraction to concrete perceptual and local knowledge on the one end of the scale and abstract conceptual and non-local knowledge at the other end. Boisot grasps the essence of the abstraction process, like the codifying process, in a form of economizing on data-processing [see also the previous chapter].

We conclude this section on theoretical knowledge with an interesting quote from Johnson-Laird [2006: 414] on the importance of theoretical knowledge and its practical use.

Our ability to reason is vital. The better we reason the better our lives. We are healthier, we live longer, and we are more successful in the academy and the market place. In life we make many simple inferences, and we make them almost without realizing that we are reasoning. I described what happened when I first went into an Italian coffee bar to get a cappuccino. I had to infer that one pays first then orders at the bar. This induction was supported by a deduction – that the bar wasn’t exceptional, because if it had been then other customers would have had to make the same detour as me – to the bar, to the cash register, and back to the bar.

Now that we have characterized the three knowledge types separately we will put them into perspective.


78

4.3.4 Putting the three knowledge types into perspective

Polanyi’s tacit knowing versus sensory knowledge

The sensory knowledge as characterized in the above shows strong parallels to the concepts of ‘tacit knowing’ introduced by Polanyi; he has often been used as a reference in literature on knowledge management and knowledge in relation to organizations. However there are fundamental differences between our sensory knowledge and the tacit knowing of Polanyi [1966].

Polanyi uses the concept of tacit knowing in a functional way. That is, knowledge can be tacit in one situation, but very explicit in the next. Our sensory knowledge is ‘absolute’; it remains sensory knowledge, this does not differ per situation. The following example used by Polanyi himself will show the functional use of tacit knowledge. Polanyi quotes an experiment conducted by Lazarus and McCleary in 1949, in which a person is shown different syllables. A shock is administered to this person after some of these syllables. After a few shock syllables this person will start to shows symptoms of anticipating these 'shock syllables', but without being able to name them when asked. So this person cannot explicitly state which ones are the shock syllables and which ones are not. This example, to Polanyi, illustrates the structure of tacit knowledge, that is, two terms are associated: the shock syllables and shock associations on the one side and the electric shock following on the other. According to Polanyi, this connection remains tacit, because the subject's attention is focused on the electric shock [rather than on the shock syllable]. He says: 'we attend from something [the proximal term, in the example the shock syllable] for attending to something else [the distal term, in the example the shock – original in italics]'. This functional aspect of tacit knowledge gives meaning to a situation. So tacit knowledge can be seen as the knowledge that is in use in experiencing a situation, but not in focus. To emphasize that this type of knowledge is not in focus Polanyi [1964] also refers to it as 'subsidiary awareness', playing a supporting role to add meaning to a situation [as apposed to focal awareness – the knowledge which is in focus and can be verbalized].

Polanyi’s functional use of tacit knowledge implies that the same [content of] knowledge can either be tacit and not in focus in one situation and explicit and in use [thus no longer tacit] in the next. Thus whether certain knowledge is tacit or not depends on the situation. The sensory knowledge of our cognitive-semiotic model on the other hand, will stay sensory in any given situation. For instance, the person in the shock syllables example has built up sensory knowledge about the shock syllables. However, this sensory knowledge will not become a different


79

type of knowledge given a different situation; it can only change when this person becomes aware of the sensory aspect in knowledge in that these shock syllables become apparent and this person will be able to name them. Until that happens will the knowledge remain sensory.

Thus, sensory knowledge and tacit knowledge are fundamentally different. The differences between these types of knowledge become apparent in transfer. Tacit knowing cannot be transferred per definition, as it consists of a combination of personal experiences that give meaning to a certain situation. Sensory knowledge, on the other hand, refers to perceiving differences. Even though this also depends on former experiences, sensory knowledge does not involve a complicated combination of experiences. Rather, like the mango-example, you have experienced something or you have not in which case you might be able to perceive a difference. When you do, then sensory knowledge is created. Therefore, we can transfer this type of knowledge. Another difference is in verbalization of these types of knowledge. Whereas tacit knowing is difficult to verbalize it is not impossible. In some situations tacit knowledge could be the equivalent of what we would characterize as theoretical knowledge. Sensory knowledge, on the other hand, by definition is impossible to verbalize. Similarities are founded in the fact that both are personal; this implies some mutual characteristics. Regarding dependency on context, to both knowledge types the context in which they are used is crucial to their interpretation.

Boisot’s codification dimension versus coded knowledge

The main difference between Boisot’s codification dimension and the coded knowledge dimension of our model lies in the essential difference of the conti-nuum. Whereas the codification dimension ranges from ‘uncodified’ to ‘codified’, the coded knowledge dimension ranges from weakly coded [i.e. pictograms and icons] to strongly coded [i.e. formulas and musical scores]. We argue that there is a fundamental difference between coded and uncoded knowledge, which we tackle with distinguishing sensory knowledge from coded knowledge.

Boisot’s abstraction dimension versus theoretical knowledge

The abstraction dimension of Boisot pretty much equals our dimension of theoretical knowledge; its relation is accumulative to the codified dimension as theoretical knowledge is to coded knowledge. Boisot sums up the difference between the dimensions in that the codification process ‘gives form to structures’ [p.48] and the abstraction process gives structure to the codifications. The same


80

applies to our knowledge types. Whereas coded knowledge elicits from a need to communicate – and therefore gives a form to the knowledge – theoretical knowledge puts the coded knowledge into perspective – it positions the different coded knowledge knots in relation to each other.

Schematic overview of the three knowledge types

Table 4.1 provides a schematic overview of the three knowledge types and their characteristics.

Table 4.1: Schematic overview of the characteristics of the three knowledge types

Aspect Sensory Coded Theoretical

Semiotic foundation One-dimensional sign Two-dimensional sign Three-dimensional sign

Object-sign-meaning O S M

Perception of difference

O S M

Need for coding

O S M

Essence

Transformation Substitution Structures

Context relevance Situated, can only be applied

Not situated Not situated

Coding No coding Coding Relation with other concepts

Transfer/Communication

Personal/imitation, on-the-job training, Use of examples, metaphors

Group/convention, use of jargon, no personal contact is necessary

Structure or reason

Learning On the job training,experience

Books Education, thought experiment, weighing of alternatives

Diffusion - ++ +

Quantification Degree of detail Degree of ambiguity Degree of abstraction


81

We will now discuss the dynamic aspects of our cognitive semiotic model, which hold a key position to understanding the dynamics of knowledge in the process of innovation.

4.4 A cognitive-semiotic model to understand the dynamics of knowledge

Peirce argued that ‘signs do not simply exist but also grow’ [semiosis, Smith: 5] emphasizing the dynamic property of signs. The following section will elaborate on the intrinsic dynamics of our cognitive-semiotic model.

In the above we presented three types of knowledge, which form the basic elements of our cognitive-semiotic model to be presented in this section. The essence of this model is that knowledge emerges and structurally changes during a learning process. Furthermore, this model stresses the cumulative nature of knowledge; first comes sensory knowledge, which forms the foundation for coded knowledge on which eventually theoretical knowledge is built.

As chapter 2 showed, the essence of innovation is a structural change, a discontinuity with what was before. Therefore, the cognitive-semiotic model is particularly useful to study the process of innovation in terms of knowledge, as this model presents the structural changes involved in knowledge change.

4.4.1 Accumulation of knowledge types

The nature of the three knowledge types implies that they are cumulatively related. Just as the semiotic dimensions add up, so do the knowledge types. Starting when difference is perceived [the new taste of mango], sensory knowledge emerges. This perception of difference is essential to coding; the newly perceived is coded as a need for communication [even with oneself at a later moment] emerges. Thus, coded knowledge builds on sensory knowledge. Coded knowledge in turn is fundamental to theoretical knowledge. Hence, theoretical knowledge essentially entails the deeper understanding of a code [concept], which consequently implies relating the code to other codes. Thus, theoretical knowledge is an extension of coded knowledge, which in turn is an extension of sensory knowledge. So, theoretical knowledge entails coded knowledge as well as sensory knowledge. In this way the three knowledge types are cumulatively related.

This line of reasoning can also be applied in more detail [see figure 4.2]. Coded knowledge presumes sensory knowledge. However the sensory foundation of this


82

coded knowledge was built up during a previous experience, which in content was not necessarily directly related to the coded knowledge. Perception 2, an 'old' experience with the fruit pear, does not only form the basis for sensory 1, but also for sensory 2. For the build-up of coded knowledge the same applies. Sensory 2 isthe foundation for coded 1, coded 2 and coded 3. And for theoretical knowledge the same applies, coded knowledge can function as a foundation for more than one sort of theoretical knowledge. Theoretical knowledge provides coherence and structure. So this means that coded 2 that forms the foundation for theoretical 1 can also form part of the foundation for theoretical 2.

Figure 4.2: Build-up in knowledge types of the fruit mango [start reading from the bottom to the top]

So, one individual perception is not merely linked to one specific ‘unit’ of sensory knowledge. Hence, this would mean that a certain skill from one field of work or knowledge area could not be used in another knowledge area. In fact, we argue that this cross-founding of knowledge forms the strength of a having a broad armature of skills and experiences. Knowledge, therefore, is the complex activation of a network. Note, however, that some theoretical knowledge is needed to bridge the different work fields. Without a frame of reference it is


83

impossible to use particular types of knowledge in different situations, which would leave the different knowledge fields incidental and isolated.

Now that we have established the relation between the three knowledge types we will consider their dynamics in more detail in the following section.

4.4.2 Knowledge change

The first way that knowledge can be dynamic according to our cognitive-semiotic model is through changes within one type of knowledge [see figure 4.3]. That is, the increase of quality within a knowledge type; thus, within sensory knowledge an increase in the degree of detail, within coded knowledge an increase in the degree of coding, and within theoretical knowledge an increase of the degree of abstraction. This type of dynamics within knowledge enhances the depth of knowledge. Larkin [1983, in Johnson-Laird 1993: 486] captures differences in the quality of knowledge, [the theoretical knowledge of novices compared to that of experts] in the following

… an important difference between the way a novice and an expert reason about a physical situation is that the novice's model represents objects in the world and simulates processes that occur in real time, whereas a trained scientist can construct a model that represents highly abstract relations and properties, such as forces and momenta.

Figure 4.3: Visualization of the difference between knowledge conversion and knowledge change

The above example shows that even though the novice and the scientist both clearly have a model [coherence in thinking, implying theoretical knowledge], which they use to reason and to act upon, the model used by the scientist is


84

much richer and more advanced than that of the novice. The novice uses a more concrete form of theoretical knowledge than does the expert. The fact that the model of the novice represents [physical] objects also point to this fact [compare the abstraction dimension of Boisot].

4.4.3 Knowledge conversion

Next to changes within a knowledge type knowledge can also be converted from one type into another; this is called knowledge conversion [see figure 4.3]. Knowledge conversion occurs every time that a new dimension emerges. Thus, when the need for communication emerges and codes are made, then a conversion from sensory knowledge into coded knowledge occurs. Also, when a deeper understanding of the code leads to relating this code to other codes then coded knowledge is converted into theoretical knowledge. This, however, does not imply that the sensory and coded knowledge, respectively, decrease or vanish. The conversion of knowledge adds a dimension to the knowledge rather than replacing one type of knowledge for another type of knowledge. However, we do want to note that the emergence of an additional knowledge type will open up possibilities. The fact that it might appear that the sensory knowledge decreases when coded knowledge comes into play can be explained in that knowledge is not a ‘thing’; it is a dynamic interplay and can only show its value when in use. So, when sensory knowledge is converted into coded knowledge often the coded knowledge takes over and the sensory knowledge is used less. This decrease in use then causes the sensory knowledge to decrease; so the knowledge conversion can indirectly cause a decrease.

Theoretically, three types of knowledge add up to three combinations of two knowledge types with each two variations:

1a. sensory � coded and 1b. coded � sensory

2a. coded � theoretical and 2b. theoretical � coded

3a. sensory � theoretical and 3b. theoretical � sensory

Our cognitive-semiotic model only allows two variations: 1] sensory knowledge converts into coded knowledge [1a], and 2] coded knowledge converts into theoretical knowledge [2a]. The cognitive-semiotic model argues that knowledge is accumulated from perception to sensory knowledge, to coded knowledge and then to theoretical knowledge. Thus, change of codes leads to change of theoretical knowledge, but not the other way around; theoretical knowledge


85

cannot change coded knowledge. So there is no knowledge conversion from theoretical to coded or to sensory knowledge.

Knowledge can, however, be rebuilt. Let us take the Copernicus example of refocusing from a geocentric model to a heliocentric model as an example. Copernicus suggested that the geocentric model by Ptolemy be replaced by the heliocentric model. The new theoretical knowledge represented in the heliocentric model is formed by different coded knowledge. Additionally a new structure is applied, restructuring the coded knowledge. Therefore the theoretical knowledge is rebuilt into new theoretical knowledge. The structure of the geocentric model has been proven incorrect, or rather, some of the coded knowledge did not fit the geocentric model. Therefore, Copernicus applied a new structure, which could fit the coded knowledge. This line of reasoning is in line with the intermediate position that Boisot suggests between the two polar cases of N-learning and S-learning. Boisot [1998: 94] notes

It is possible to hold a position midway between the two polar cases presented. A shift of paradigm, for example, may involve a destruction of codified and abstract knowledge embedded in organizational processes, but leave intact the tacit knowledge base from which such knowledge is derived.

Nersessian [1992] uses the concept of ‘conceptual descendants’ [p.10] to refer to the rebuilding of knowledge.

We want to note at this point that knowledge does not simply change by itself; there is always something that causes the change, which sets the knowledge dynamics in motion. Thus, in case of a misfit between data and model one could either ignore this misfit and treat the data as idiosyncratic and continue to use the [old] model [think of the naïve theories] or one could look for a new mode of coherence, a new model.

So, knowledge can be dynamic either within a knowledge type or between knowledge types; we expect that both knowledge change and knowledgeconversion are processes related to innovation.

4.4.4 Individual differences in learning

Innovation can be characterized as a learning process as the definition by Hilgard and Bower [1975: 7] shows:


86

Learning refers to the change in a subject’s behavior to a given situation brought about by his repeated experiences in that situation, provided that the behavior change cannot be explained on the basis of native response tendencies, maturation, or temporary states of the subject (e.g. fatigue, drugs, etc.)

Hilgard and Bower [1975] compare the relationship between learning and knowing to the relationship between process and result. So, the result is the newly acquired knowledge, which reveals itself in new behavior.

We know from literature that people differ greatly in their learning processes [e.g. Ackerman, Sternberg, & Glaser 1989]; this results in differences in knowledge.

Therefore we are interested in the possible differences in knowledge dynamics. In this respect these differences are the resultant of a process. However, these expected differences are based on the assumption that two [groups of] individuals each have a different starting point as well. In that respect the differences are regarded as indicators. Horn [1989] frames this difference in the following way: ‘Human abilities are simultaneously outcomes of learning and determinants of learning’ [p.61]. The implication for the present research therefore is not only to establish differences in dynamics, but we need to establish differences at the points of departure, before the implementation process has started. The points of departure are the different starting points of the individuals that learn or that have to change their knowledge.

Choosing the factors that influence the relationship between knowledge and innovation we are particularly interested in the characteristics of the individual itself, as opposed to characteristics such as the working environment of the individuals or the political preferences that individuals hold. In particular we focus on factors that are directly related to the individuals’ work. We therefore want to explore the influences of working experience and that of contractual hours on the relationship between knowledge and innovation. We also include the factors of age and education, as these factors are often posed as factors of influence in studies on individuals. In the following we will discuss these four characteristics separately.

Job experience and expertise

Job experience stands for the amount of working experience that a person has, mostly expressed in number of years. Job experience, or the degree of expertise,


87

has been shown to be of great influence on how people organize and conduct their work [e.g. Mietus 1994]. However, the effect of training is expected to be more predictable for people with less experience than it is for experts [Hunt, 1999: 22]. So, in understanding the implementation of a DSS in terms of knowledge dynamics we particularly expect a difference for coded knowledge; coded knowledge is expected to increase, as the way of working will become more standardized with the introduction of DSS.

Education

We argue that education is a strong indicator for theoretical knowledge, which implies the use of models and applying structure. So, the higher the education the more theoretical knowledge a person has and the more likely that person is to perceive differences. Therefore, higher educated people are expected to show an increase in sensory knowledge in comparison to lower educated people. But also, the higher the education the more set the use of coded knowledge is. So, for the coded knowledge we expect a difference between the higher and lower educated as well.

Age

Age related differences in acquiring knowledge have been studied frequently. We therefore wonder whether younger people show different knowledge dynamics compared to older people. We especially want to relate this factor to coded knowledge as coded knowledge is expected to be the key knowledge type to show dynamics.

Contractual hours and building routine: part time versus full time

Differences between part timers and full timers have been particularly studied in relation to job satisfaction [Eberhardt & Shani 1984; Steffy & Jones 1990], job attitudes [Levanoni & Sales 1990], status differences [Jackofsky & Peters 1987], and specifically in relation to women [e.g. Nakamura & Nakamura 1983]. Feldman [1990] questions the one-directional relation of part time work that influences workers’ attitudes; rather, he suggests an influence in the opposite direction as well. That is, ‘employees who have certain job attitudes and work behaviors may gravitate toward part-time work instead of full-time work’ [p.104].

Based on our theoretical framework we argue that it is very likely that part time workers differ from full time workers in the way that they organize and use their knowledge. Part time workers differ from full time workers in the amount of


88

shared experience that they have with other colleagues. Another difference is that part time workers in comparison to full time workers spend much more time on communication about the status of the job. That is, they need to communicate much more about what has happened and what still needs to be done. We do want to remark here that there is a difference between part time workers that perform tasks that are autarchic [= need not communicate with others, e.g. data typist] and those that work as part of a continuous process, such as nurses and secretaries. Furthermore, the type of work might differ as well between part timers and full timers, which could be an indication for the types of knowledge that one uses. We are therefore interested to explore the possible differences between these two groups.

4.4.5 Visualizing knowledge [dynamics] of innovation in the K-Space

Inspired by the I-Space we visualize our cognitive-semiotic model in three dimensions as well; we will refer to this space as the Knowledge Space or K-Space. Similar to the I-Space the K-Space enables visualization of both static situations as well as dynamic situations. Additionally, it is possible to capture both the knowledge [dynamics] of one individual as well as [a unit of] an organization –adding up the separate individuals.

Sensory knowledge is placed on the X-axis ranging from rough sensory knowledge [on the left] to detailed sensory knowledge [on the right]. Coded knowledge is placed on the Z-axis ranging from weakly coded knowledge at [at the front] to strongly coded knowledge [at the back]. And finally theoretical knowledge is placed on the Y-axis ranging from concrete theoretical knowledge [at the bottom] to abstract theoretical knowledge [at the top].

We note that, just as the I-Space, the axes are not orthogonally related as in a mathematical space. The K-Space is just as the I-Space a kind of metaphor.

Before the start of an organizational innovation process we can map the knowledge of the individual people who will be affected by the innovation. To place them in the K-Space we determine their sensory knowledge, their coded knowledge and their theoretical knowledge separately, which results in three coordinates [see figure 4.4a, each person resembling a dot]. This way we have mapped the knowledge of two persons at the beginning of the organizational innovation. Now, we can also map the desired knowledge, so can determine a possible discrepancy between the knowledge that a person has and the knowledge that this person needs to have to adequately take part in the


89

organizational innovation [see figure 4.4b]. When there is no discrepancy this means that the organization, from a knowledge type perspective, is ready to start the innovation process. If, on the other hand, there is a discrepancy, than this must first be eliminated.

Figure 4.4a [left]: The knowledge space [K-space]Figure 4.4b [right]: Discrepancies in knowledge visualized in the K-Space

In sum, this chapter presented a cognitive-semiotic model for studying dynamics of knowledge, distinguishing three types of knowledge. These knowledge types are essentially different and cumulatively related. We argued that this model is particularly interesting to use to study the knowledge during innovation processes, as it provides a tool to understand the structural changes that take place during this process. Furthermore, this tool can be used to evaluate innovation processes. The next chapter will focus on the domain of knowledge that we used in this study, the domain of planning.


90

91

Chapter 5

Planning

5.1 Introduction

Choosing the planning domain to study organizational innovation from a knowledge perspective was triggered in three ways. Firstly, the domain of planning is highly cognitive [e.g. Mietus 1994], as this chapter will show. Secondly, the process of planning has been shown to be an important influence on innovation [e.g. O’Connor 1998; Castrogiovani 1996; Avlonitits, Kovremenous, & Tzokas, 1994]. Although the first point is of more direct concern to the present study the second point underlines the power that comes with the planning domain. The third trigger, finally, concerns a practical reason; at the faculty of Economics and Business we developed knowledge and experience on planning and planning support, e.g. the planning support software ZKR. This was developed between 1989 and 1995 at the university of Groningen and became a commercial software product in 1996.

The aim of this chapter is to show some aspects of the rich and complex domain of planning. We therefore start with discussing the planning domain in general, fo-cusing on its practical value and the differences between everyday planning and organizational planning [5.1]. We will then show the great complexity of the plan-ning domain, in particular planning as a scheduling task [5.2]. Section 5.3 discusses the main subtasks of planning that are of interest to us for the purpose of our stu-dy on knowledge dynamics. The subtasks are ‘gathering information’, ‘negotiating’, and actual scheduling [5.3]. And finally the last section [5.4] is on the use of com-puter support for planning, in particular the use of ZKR to make the duty roster.

5.2 The knowledge domain of planning

5.2.1 Planning in everyday life

We use planning in organizing our everyday life [Woll 2002], especially when we want to achieve something which requires several parallel and serial actions, for instance the planning of a vacation or the running of errands [Hayes-Roth & Hayes-Roth 1979]. In planning a vacation we come across a number of practical


92

decisions. For instance, when we choose a destination we need to consider the following

� The type of vacation, for instance do we want an active vacation, do we want to see many things and educate ourselves at the same time or do we just want to lie on the beach. This decision relates to the goal of our vacation, which is seldom precisely defined or clear-cut.

� What facilities are available at the place of destination? This decision relates to the availability of information, which seldom is certain and complete

� Valuing the available information about our destination. This decision relates to the reliability of the sources of information.

� How to get to the place of destination

� Considering alternative destinations

These are some aspects to consider, but there are probably more. And then there is a difference in planning for yourself and planning for others; just think about picking a travel destination for a group of friends. So, planning involves many aspects and actions such as in the vacation-example. However, we surprisingly often seem to plan without thinking, combining the different aspects of planning in a natural way.

Probably because of its complexity, within the academic world different aspects of planning have been studied by a variety of disciplines ranging from economics, management and organization, operations research to mathematics, artificial intelligence and cognitive psychology [Kiewiet, Jorna, & van Wezel 2005]. Taking on different perspectives, these disciplines all deal with attuning and coordinating objects, activities and actions in time and space. To illustrate the impact that planning can have we will discuss the example of the seven bridges and the bridge masters taken from Jorna, Gazendam, Heesen, and van Wezel [1996].

5.2.2 The bridge masters example

The following example illustrates the difference a planning can make. Before the planning was introduced the situation was as follows. A busy part of an important waterway had seven bridges across. Each bridge was operated by a single [different] bridge master. The seven bridge masters all lived in their own bridge

Chapter 5 Planning

93

master house as part of their own bridge. During the night all the seven bridge masters were asleep and consequently each of the seven bridges was closed. During vacation periods all of the seven bridge masters had to find their own replacement. All seven bridges were operated independently without a planning schedule; the seven bridge masters all planned for themselve.

The introduction of modern technology made it possible for one bridge master to operate all seven bridges, which resulted in radical changes. First of all, every day was divided into different parts; each day now contained three shifts [early shift, late shift, and night shift]. And instead of seven bridge masters all operating individually and at the same time, each shift now only required one bridge master to operate all seven bridges [plus one bridge master as a sub in case of emergency]. Then, the bridges were now also operated during the night. As a result of this new approach only six bridge masters are needed instead of the former seven. The six bridge masters all work 38 hours a week and they all have a four-week period of vacation each year.

The bridge masters example shows that planning can have a great impact on required capacity and services. The example shows, furthermore, that planning calls for the uniting of different independent organizational units [the seven bridge masters] into one bigger unit [six bridge masters working together]. So the consequences of planning are often more far reaching than one might expect at first and organizational planning is no exception. For example, limited capacity of personnel calls for a highly qualitative duty roster [Jorna et al. 1996].

5.2.3 Comparing everyday planning to organization related planning

As the above examples of planning a vacation and the bridge masters show, there are some differences between everyday planning and professional planning in organizations. Whereas everyday life planning often goes by tacitly, organizational planning is much more explicit. Typical for everyday planning is that many different goals arise simultaneously and planning opportunistically seems to be a strategy to cope with these simultaneous goals and keep focus [Hayes-Roth & Hayes Roth 1979; Patalano & Seifert 1997]. Planning within organizations is a coordinating activity aiming to improve the performance of the organization, as with most organizational activities [Mietus 1994: 18].

In general we can distinguish three types of organizational planning, namely strategic planning, tactical planning and operational planning [Mietus 1994]. Strategic planning focuses on organizational policy and organizational goals.


94

Tactical planning focuses on acquisition and use of [organizational] resources. Anthony [1965] uses the term of management control rather than tactical planning and emphasizes the information handling aspect of organizational planning. And thirdly, operation planning assigns operational tasks that should be carried out immediately [Mietus 1994: 18]. It is very difficult to make clear-cut distinctions here, and therefore our focus of attention is on tactical planning and especially operational planning.

5.3 Planning as a problem solving task, the knowledge to make a duty roster

In defining the concept of task, literature on this subject emphasizes aspects such as 'a complex situation capable of eliciting goal directed behavior' [Farina and Heaton 1973 in Lamain 2000]. Other emphasized aspects are [human] actions [Drury, Pramore, Van Cott, Grey, & Corlett 1987 in Lamain 2000; Hacker 1986], [associated] goals [Hacker 1986 in Lamain 2000; [Filkes 1982 in Wærn 1989] and achieving these goals [Filkes 1982 in Wærn 1989]. Generally speaking, people perform tasks in order to achieve goals [Lamain 2000]. Thus, a task involves setting goals and accomplishing them through behavior that is required to perform the task.

Several types of planning tasks can be distinguished, such as production planning, transport planning and planning of personnel, but there are more. These tasks essentially attune different types of entities, for example persons, vehicles, locations, machines et cetera, while taking into account a number of constraints. The aim is either minimizing or maximizing different goal functions [Jorna et al. 1996]. In this way, planning as a task can be viewed as a form of problem solving [Hayes-Roth & Hayes-Roth 1979]. In the following we further elaborate on the task planning of personnel, the focus of the present study.

Manpower planning to make a duty roster is very complex. Figure 5.1 shows a conceptual model of planning as a task for the domain of nurse scheduling [Mietus 1994]. The figure shows that the planning task involves many entities of different caliber; this is precisely what makes planning such a highly cognitive task, putting a great strain on the planner. We will discuss the five knowledge clusters for making the duty roster as put forward in figure 5.1.

Chapter 5 Planning

95

5.3.1 Knowledge of the schedule itself

The planner needs to have knowledge about the schedule itself. This knowledge implies that the planner can place the planning periods, the shifts, and the personnel in the empty framework of a schedule.

5.3.2 Knowledge of shift

A second knowledge domain concerns the three shifts, which are day, evening, and night. Each of these three shifts has its own demands in terms of quality and quantity of staffing. For instance, a planner has to consider the irregular working hours, that result in obligatory resting hours for recuperation from work. But also the mandatory days off and the compensation days around the vacation periods, such as Christmas, are important. And then there are all kinds of variations on this three-split in shifts caused by the contractual hours and labor agreements, for instance resulting in 7.2 hour-shifts.

5.3.3 Knowledge of personnel

The third knowledge domain is concerned with the personnel to be scheduled, a complicated domain. This knowledge entails the personal data related to individual staff members, for example function, the degree of experience, contractual hours and labor contracts. Knowledge about the labor contracts is important as it puts different restraints on the assignment of shifts. For example, personnel with particular part time contracts can work only in the evenings or at nights.

Knowledge of personnel also concerns established data, such as courses and vacations. Courses are considered official working time and should therefore be taken into account in the duty roster. Historical data concern the historical progression of scheduled shifts for each colleague that is scheduled. This data needs to be considered over an entire year and needs continuous updating.

And finally there are the personnel’s wishes. These preferences for specific shifts or dislikes to a specific combination of shifts are often considered to be an acquired right for the employee and should therefore be handled delicately.

5.3.4 Knowledge of constraints

The fourth knowledge domain is concerned with the constraints that are put on the duty roster, which contain labor agreements, quantity of staffing and quality of


96

staffing. Labor agreements include many precise rules about hours of rest. For instance, the completion of a night shift sequence should be followed by a period of two days off, which implies that an employee cannot be assigned to a shift in the schedule for the two following days. Quantity and quality of staffing involves the number of staffing required in combination with a balanced combination of qualified staffing.

Figure 5.1: Conceptual model of the task domain in nurse scheduling [Mietus 2004: 68]

Clusters of domain entities [Mietus 1994: 67]

Chapter 5 Planning

97

5.3.5 Knowledge of goals

And fifthly, in order to make the duty roster the planner needs to be aware of the goals involved for making the duty roster. This knowledge domain concerns honoring the wishes of personnel, assuring the quality and quantity of staffing, continuity over days and the balance between work days and days of. But there are also more complex aspects, even to experienced planners. These concern continuity between days and the distribution of shifts, per employee as well as among employees.

The above elaboration on the knowledge domains of planning shows that making the duty roster is a complex task with many different aspects. To study the dynamics of making the duty roster it is more suitable to take on a task performance view, as this allows linking the knowledge to specific actions. The next section elaborates on the subtasks involved in planning.

5.4 Subtasks of planning

5.4.1 Planning as a task

Planning is considered to be a synthetic task [see Clancey 1985 or Schreiber, Wielinga, & Breuker 1993 in Jorna et al.1996: 66 for more details]. In contrast to an analytic task, such as diagnosis, where decomposition is important, in the planning task various activities or entities have to be merged or integrated. Furthermore, planning is considered to be a generic task, meaning that despite strong domain differences similar subtasks and similar reasoning patterns between planners in different domains can be discerned. Mietus [1994: 93] distinguishes three clusters of sub tasks, namely administration, problem solving and evaluation. The first cluster of administration within planning involves processing of personal data, determining fixed data, determining historical data, and determining wishes for the planning period. The second cluster of problem solving involves scheduling a shift, scheduling personnel, scheduling part of a planning period, counting the quantity of staffing and counting the number of assigned shifts. And all this to see whether the goals are met. And finally the third cluster involves weighing the schedule goals, checking the constraints, counting the quantity of staffing and counting the number of assigned shifts. For practical reasons we renamed the task clusters and fine-tuned their content to fit the empirical setting of our study. We renamed the three subtasks of planning into gathering information [part of the administration cluster], scheduling [part of the problem solving cluster], and


98

negotiating [relevant for problem solving]. They cover the most important aspects of making the duty roster. This reshuffling implies that we have left out both the subtasks of counting and evaluating; the weight of the subtask of counting decreases enormously with the introduction of DSS and the subtask of evaluating has more or less been put together with the subtask of negotiating.

Redefining the planning task in terms of the three subtasks of gathering information, scheduling and negotiating allows us to bridge the knowledge types and the knowledge dynamics to actual activities. Furthermore, it allows us to differentiate between the knowledge dynamics of the different activities involved in the innovation process.

5.4.2 Gathering information

The process of gathering information involves the selection, integration, and retrieval of data from diverse information sources, distributed as well as heterogeneous [e.g. Knoblock 1995]. Information gathering can be ascribed a key role in the planning process. Le Breton and Henning [1961] even state that a successful plan can be no better than the information made available and the judgment exercised with this information. The gathered information concerns all information required to plan, varying from constraints to practical information such as availability of resources. Information can be gathered in a number of ways, but most important is to do so in the light of its expected use [Le Breton & Henning 1961].

Le Breton and Henning [1961] remark that the significance of certain data does not lie in their total quantity; making up only a relatively small part of all necessary information, it can nevertheless have a key role in the final duty roster.

We distinguish five types of data to be gathered, namely personal data, wishes, established data, historical data and constraints [Mietus 1994]. In this way gathering information can be considered to rely heavily on coded knowledge. However, the information used for making a planning is not always gathered explicitly. Situations that provide a great deal of information for some planners may not provide any information for other planners.

5.4.3 Scheduling

Once all the data has been gathered the actual puzzling to fit all the required data into the roster can begin, the subtask of scheduling. Generally speaking scheduling consists of three major parts, attuning, adjusting and valuing. Attuning refers to

Chapter 5 Planning

99

fitting the instances of object types. This is the real puzzling, it deals with reasoning, weighing of alternatives and choosing what entities to combine, to attune or put together. Justifying the final duty roster is an important goal, so theoretical knowledge is expected to play an important role in this subtask.

Adjusting is different from attuning in that it normally is done after the duty roster is finished. Adjustment is required in case of changing circumstances, for instance due to illness or in case the duty roster is unacceptable to the personnel who have been scheduled, but also when the scheduled people are dissatisfied with the roster. Valuing is mostly done when the duty roster is more or less finished; the planner evaluates to see whether adjustments can be made for improvement. In other words, this part of scheduling weighs alternatives. Most of the time this is done implicitly [Mietus 1994].

5.4.4 Negotiating

Negotiating can be considered part of the adjustment process, when the draft version of the duty roster is finished. Although negotiating is often associated with great conflicts such as peace negotiations between countries at war, negotiating is used in many other situations as well, such as everyday life situations in traffic or standing in line at the grocery store [Lewicki, Saunders, & Minton 1998]. The aim of negotiating is agreement and this implies compromise [Morley & Stephenson 1977]. According to Lewicki, Saunders, and Minton [1997] negotiating situations have the following in common

� Two or more parties are involved

� There is a conflict of interest between the two or more parties

� It is a voluntary process; the parties involved think that they can influence each other

� The parties prefer to search for agreement, rather than fight openly

� It is expected that the involved parties both give and take

� Successful negotiation involves the management of the intangibles as well as the resolution of the tangibles

Furthermore, negotiating is characterized by joint decision making and talking about the relationship. This last characteristic distinguishes negotiating from bargaining – in research literature – as bargaining does not necessarily involve


100

verbal expression; bargaining could be denoted as the manner in which agreement is negotiated [Morley & Stephenson 1977]. Diplomacy can also be considered a form of negotiating as making compromises in international political matters [Neal 1964 in Morley & Stephenson 1977].

In the case of making the duty roster, the preliminary duty roster often creates a situation for negotiation, as most of the time at least some people are dissatisfied with the preliminary duty roster. So, negotiating as part of planning implies that the roster is not finished yet and that certain aspects of the roster need to be changed, which calls for fine tuning and calibrating, negotiating. For instance, it might be impossible to grant all wishes concerning weekend shifts or night shifts. Then, decisions need to be made as to who gets his or her wish granted and who does not. Much depends on the constraints concerning these special shifts in combination with the grounds of the wish. We therefore expect the subtask of negotiating to involve a substantial part of sensory knowledge.

5.5 Planning with decision support software

5.5.1 Computer supported tasks

The use of computer support to perform a task such as staff planning or manpower planning becomes more and more popular. In the first place, this increase in popularity can be explained by the possibilities that this computer support offers to improve the quality of the duty roster, in process as well as in outcome. As the above showed, planning puts a strain on cognitive capacity while decision support can remove some of this strain. A second explanation relates to administrative aspects, such as manipulating the data. For instance, as the data can be processed uniformly, using these data for other purposes becomes incredibly easy: that is, analyzing the data or linking them to other, for example financial, data.

The use of computer support does have a threshold; working with a computer program is something you need to learn. Especially, people who are unfamiliar with computers are not particularly fond of computer support. Wærn [1989] notes that learning a computer task will not be facilitated when performed under stress, such as time pressure or noisy environments. Other dismays include getting headaches from looking at the computer screen, not having the overview, which planners do have when using pen and paper, or not being able to scratch a few notes on the paper copy. But also restrictions in terms of having to work indoors and having to work with access codes can be considered a disadvantage

Chapter 5 Planning

101

to computer support. Furthermore, people who have much experience in planning [experts] and who will now have to learn to work with computer support do so only to continue their expertise at their main task of planning; experts may not particularly be interested in understanding the computer support itself, compared to novices.

So, the use of computer support affects the execution of a task in different ways. New skills need to be learned, such as using a new software tool; some people might even need to get used to working with a computer in general. Side effects can be that the new way of performing the task becomes time consuming. Thus, the introduction of computer support is considered an organizational innovation, from outside coming into the organization [Leonard-Barton 1995].

To get an idea of the impact that starting to work with DSS has, we will discuss the decision support system ZKR, a computer software program to support planners in making the duty roster [Jorna et al. 1996]. This software program was also introduced in Bartiméus [see chapter 7].

5.5.2 ZKR: Planning in health care

Introduction

ZKR is a software program for planning support. ZKR stands for ZieKenhuis-Rooster-systeem, meaning duty roster for hospitals. It was designed and developed at the university of Groningen [van Wezel, Jorna, & Mietus 1996] and was commercially explored by IKS and it was a widely used product between 1998 and 2003. In 2004 it was integrated with ‘Harmony’ [ORTEC], another DSS. No commercial or organizational connection between IKS, ORTEC and the university of Groningen has existed since 1994. ZKR was specifically designed to function as a personal support for the planner; it helps in structuring data, it ‘makes various schedule proposals, in conjunction with the user, for the different shifts which have to be scheduled’ [Numan 1998: 74]. Starting point in the design process was to stick to scheduling by hand as closely as possible. So, the aim of ZKR was to make the duty roster look as similar as possible on the computer to the manual duty roster. Figure 5.2a [on bookmark] shows a ZKR duty roster. The personnel are listed in rows at the left hand side. The days to be scheduled are represented by the columns; the different colors indicate different types of shifts. At the right hand side as well as at the bottom are the [automatic] calculations of hours are represented per person and per day respectively. Figure 5.2b [on bookmark] shows how a type of shift can be selected through a pop-up menu.


102

Thus, the computer software functions as a means to support the planner. It does not function as a tool to replace the planner. That is, the planner will always be the vital link in making the duty roster. ZKR neither functions as a tool that corrects on-line, for instance in the way that ‘Word’ provides the ‘AutoCorrect’ tool. New software programs such as Harmony do provide this functionality, but it is up to the planner to decide to use it or not. Programs such as ZKR and later Harmony provide functionalities in that they generate all kinds of overviews. These types of programs are therefore also especially interesting for management purposes as they provide the functionality to combine information.

The primary focus of ZKR is that it supports the subtasks of gathering information and scheduling, which are subject to the present study; it does not support the subtask of negotiating. Although ZKR does not support the subtask of negotiating in the same way that it supports both gathering information and scheduling, we do expect the subtask of negotiating to be influenced by computer support. Hence, planners will have a great amount of information at his/their disposal. So it becomes easier to take different points of view, as ZKR generates proposals. For instance, ZKR clearly points out the restrictions and constraints of changing the duty roster. In this sense planners can use ZKR to back up their argument.

The SEC-model

ZKR takes the SEC-model – Scheduling Expertise Concept – as a starting point, distinguishing two dimensions in computer support in planning [Jorna et al. 1996]. The first dimension consists of two modes, planning-by-hand [computer supported] and automatic planning by the computer. The planning-by-hand mode involves everything that a planner normally does when planning without a computer, using pen and paper et cetera. In other words, the computer provides an electric pen and paper. On top of that the computer provides easy access to information, it does the counting and everything is available at the same time. This mode also enables checking and evaluating preliminary plans. In contrast, using the mode of automatic planning the planner commands the computer to generate a possible duty roster, although the planner still provides the starting point in supplying the data. The automatic planning mode still entails a planning-by-hand aspect as the planner can make some adjustments [or of course reject the roster completely].

The distinction between planning-by-hand and automatic planning can be made more fine-grained by distinguishing four cumulating forms of support: 1] offering and requesting information [editor], 2] checking criteria [inspector], 3] evaluating

Chapter 5 Planning

103

goal functions [evaluator]', and 4] generating planning solutions [generator]. The editor provides the simplest form of support; in addition to the same options as when planning by hand the planner has the possibility of automatic counting and easy access to information. The second form of support, provided by the inspector, offers a much more advanced type of support. Additional to the possibilities of the editor mode, the inspector offers the option of having the computer automatically check the constraints that were set up front by the planner in creating the duty roster. When the computer signals a violation of the constraints the planner can choose whether or not to adjust the duty roster. An example often encountered is the constraint of forward rotation. This constraint, based on physiological research, dictates that a person may not take an early shift following a late shift. In practice this constraint will often be violated as it implies ‘sacrificing’ days off

At the third level of support, provided by the evaluator, goal functions are used and evaluated in terms of minimizing and maximizing these goal functions [see figure 5.2c on bookmark]. This level of support can be viewed as an extension of the inspector level; when a constraint has been violated the planner will try to minimize this violation in changing this constraint into a goal function. Whereas the inspector formulates the constraints very strictly, such as either a maximum of two nightshifts or not [a dichotomous choice: yes or no], the evaluator formulates the goal functions in terms of percentages [weighing of the goal functions, see the green beams in the pop-up menu of figure 5.2d on bookmark]. For example, using percentages a planner can see how to value a certain goal function, for example a planner is satisfied if 60 percent continuity is met during day shifts.

The generator offers the fourth form of support; the computer generates a few partial alternative duty rosters. The program will show a number of solutions among which the optimal solution, in meeting all constraints and goal functions as best as possible. Through a pop-up menu the planner can select the type of shift for which a solution is requested and the week in which this shift should take place. After the information is put in the computer ZKR will generate a solution [see figure 5.2e on bookmark]. The planners can now decide either to take the generated solution or to try another option.

5.5.3 Differences between manual planning and supported planning

The change from planning manually to planning with decision support implies that the execution of the planning task will change [Roth & Woods 1989 in Mietus 1994], for instance improvement of performance and increase of possible


104

solutions for a specific problem. Mietus [1994] more specifically studied the difference between planning manually and planning with the decision support of ZKR. She concludes the following. First of all, decision support to make a duty roster takes over some of the time consuming activities, such as counting. But it also guards the constraints. As a consequence, the planner will have less cognitive limitations, which allows a different allocation of cognitive resources. In particular, she found that the domain knowledge was used to weigh alternative possible rosters generated by ZKR; an activity seldom undertaken in the manual situation, as this required too much additional counting. Secondly, the ZKR structures the planning task, which makes it easier to focus on one particular aspect of the duty roster. Mietus also studied the differences between novices and expert planners and found these to be bigger in the manual situation than in the supported situation

5.5.4 In conclusion

We expect the coded knowledge to increase in the new and more standardized situation. As the new situation with ZKR opened up the possibility to weigh alternative rosters, we expect this to influence the theoretical knowledge of planners. Furthermore, we expect more knowledge dynamics for novices than for experienced planners.

Now that we have studied the theoretical issues underlying our research question the following chapter will reformulate the research question formulated in chapter 1. We will present our conceptual model and draw hypotheses from it.

105

Chapter 6

The Conceptual Model

6.1 Introduction

The previous four chapters explored the theoretical background to the research questions posed in the introductory chapter. They will now function to provide a theoretical framework for a more precise research focus. We [re]formulate our research questions, construct a conceptual model and deduct hypotheses from this model.

6.2 Research questions

1. What happens to the knowledge [types] of planners during an organizational

innovation such as the implementation of planning support software?

Furthermore, we formulate the following related questions

2. What kinds of knowledge will change in the introduction of a new way of

working, such as the implementation of planning support software? And what

differences can we discern in the presence and dominance of knowledge in

terms of content and type?

3. With respect to the innovation and the expected change, in what way do the

subtasks of planning differ in terms of knowledge types?

4. What is the effect of personal characteristics of the planners on knowledge

change in general and specifically on the knowledge dynamics of the planners

during the implementation of planning support software?

6.3 Integrated theoretical framework

6.3.1 Introduction

The conceptual model in the most basic form thus captures the effect of innovation on the knowledge dynamics of the individual end-users [see figure 6.1a].


106

The innovation process will cause a shift in the configuration of these three knowledge types at the level of the end-user and indirectly at the organizational level. Hence, the newly created situation, after the innovation has taken place asks for a new configuration of knowledge of the end-user in order to function within this new situation; this implies that the knowledge of the end-user will change. The innovation triggers the process of knowledge [dynamics] in terms of knowledge types. Essential is perception, and in particular the perception of difference as this triggers the process of knowledge dynamics. And in order to perceive a discrepancy between our perception of what was and our perception of what is, our cognitive-semiotic model is leading as a theoretical model.

Figure 6.1a: Conceptual model in the most basic form

Figure 6.1b: The innovation process influences the three individual knowledge types and leads to conversions and changes

Extending our basic conceptual model we assume that the innovation process causes the three individual knowledge types to change [see figure 6.1b]. Furthermore, two specific combinations in knowledge type change can be characterized as knowledge conversion [see above].

Thus, in extending and specifying the influence of innovation on knowledge types we see five different influences [see figure 6.1b].

Chapter 6 The Conceptual Model

107

1. changes in sensory knowledge

2. changes in coded knowledge

3. changes in theoretical knowledge

4. conversion from sensory knowledge to coded knowledge

5. conversion from coded knowledge to theoretical knowledge

6.3.2 Personal characteristics

We expect that knowledge of individual planners is related to some of their personal characteristics. We focus on education, job experience, contractual hours per week and age [see figure 6.1c below].

Figure 6.1c: The moderating influence of specific personal factors on the relationship between knowledge and innovation

Education

Education offers the use of a framework to perceive the world. In other words, it helps one to structure. This is the essence of theoretical knowledge. Thus, we expect education to be of influence on the quality of theoretical knowledge of the planners. The framework, in turn, is also essential for perceiving and processing new information. That is, to acquire knowledge. Thus, we expect education to influence the relationship between innovation and knowledge types, as education is one of the main tools to perceive. That is, education hands us models to process and interpret information. And these models determine what we perceive and thus what influences our thinking and knowledge.


108

Job experience

We also expect job experience to influence the relationship between innovation and knowledge. A planner with greater job experience will have encountered more different situations and will thus have more to compare to. In a sense, the knowledge of planners with greater job experience as well as older planners, is deeper imbedded in other knowledge.

Age

Then, we are also interested in the effects of age on the relation between knowledge dynamics and innovation, as younger planners often have a more restricted knowledge. Therefore, younger planners are expected to experience more change than older planners.

Contractual hours per week

Finally we expect the number of contractual hours per week to influence the relationship between knowledge and innovation. We argue that more contractual hours per week leads to easier build up of routines. On the other hand, fewer contractual hours necessitates communication.

6.3.3 Subtasks

Our conceptual model is almost complete. The last refinement forms the distinction between three subtasks based on their different characteristics as chapter 5 showed. Although we have no specific expectations on the differences in knowledge dynamics between the subtasks, we do expect that these differences in characteristics will differentiate the knowledge types and the knowledge type dynamics caused by innovation [see figure 6.1d]. So gathering information will show a different change pattern than either negotiating or scheduling. But also, knowledge type patterns before the implementation will differ between the subtasks.

Gathering information

Important to the execution of the gathering information task is the coded type of knowledge. We therefore expect this subtask to more heavily rely on coded knowledge than the other two types of knowledge, sensory and theoretical. Furthermore, we expect this subtask to rely more heavily on coded knowledge in comparison two the other two subtasks, negotiating and scheduling.


109

Figure 6.1d: The 15 dependent variables of our conceptual model in groupings of the subtasks

Negotiating

Important to the execution of the negotiating task is the sensory type of knowledge. We therefore expect this subtask to more heavily rely on sensory knowledge than on the other two types of knowledge, coded and theoretical.

Scheduling

Important to the execution of the scheduling task is the theoretical type of knowledge. We therefore expect this subtask to more heavily rely on theoretical knowledge than on the other two types of knowledge, sensory and coded. Furthermore, we expect this subtask to also rely more heavily on theoretical knowledge in comparison to the other two subtasks, gathering information and negotiating.

6.4 Conceptual model

Now that we have reviewed all the variables of which our conceptual model consists we can construct our total model [see figure 6.1e]. In the following we will formulate the hypotheses drawn from this conceptual model.


110

Figure 6.1e: Conceptual model

6.5 Hypotheses

We expect that the implementation of ZKR will lead to an increased need for communication as the making of the duty roster is no longer restricted to individual planners. Rather, the implementation of ZKR implies a need for consensus on making the duty roster. We therefore expect the coded knowledge to increase. Note that, as we explicitly expect the greatest changes to occur for coded knowledge most of our hypotheses focus on this particular knowledge type.

The emergence of codes and consensus implies a great potential time gain and more effective communication. This effective way of communication will have an additional effect in that the sensory knowledge will be exercised less and less.

Therefore, we formulate hypothesis 1a and hypothesis 1b as follows

Hypothesis 1a

The implementation process of ZKR will lead to a decrease of sensory

knowledge for all subtasks

Hypothesis 1b

The implementation process of ZKR will lead to an increase of coded


1a

1b

1c

2

3a 3b 4 56


111

The use of ZKR will also lead to a better understanding of the planning task and its individual subtasks. We therefore formulate hypothesis 1c as follows

Hypothesis 1c

The implementation process of ZKR will lead to an increase of theoretical


Then, our second expectation concerns the conversion of knowledge. Combining hypothesis 1a and hypothesis 1b we expect a conversion of sensory knowledge into coded knowledge; we expect sensory knowledge to decrease during the innovation and we expect coded knowledge to increase. We therefore formulate hypothesis 2 as follows

Hypothesis 2

The implementation process of ZKR will lead to a conversion of knowledge

from sensory knowledge into coded knowledge

Personal characteristics are expected to moderate the influence of the implementation of ZKR on knowledge. The first personal characteristic that we want to focus on is education, as education is a learning process; one learns to consider alternative perspectives and approaches to certain situations. Education can be described as the learning that one has already experienced; and ones learning experiences influences the learning experiences that will take place in the future. That is, ways to go about situations, in a practical sense as well as in a more abstracted sense. It might be that the higher the education, the crystallized the way to perceive has become. Therefore, we expect that planners with ahigher education will have more theoretical knowledge than planners with a lower level of education. And we argue, the more theoretical knowledge the more susceptible to change and learning. Furthermore, the more theoretical knowledge a planner has the easier they can perceive differences. And perceiving differences is the critical factor for developing sensory knowledge. Now, our expectation regarding the moderating effect of education can be formulated as follows.

Hypothesis 3a

During the implementation process of ZKR planners with higher professional

vocational education [in the Netherlands: HBO] will show a stronger increase


112

of their sensory knowledge in comparison to planners with senior vocational

education [in the Netherlands: MBO].

Hypothesis 3b

During the implementation process of ZKR senior vocationally educated

planners [MBO] will show a stronger increase of coded knowledge in

comparison to higher professional vocationally educated planners [HBO].

A second personal characteristic, which we expect to moderate the influence of the implementation of ZKR on knowledge types of the individual users, is job experience. We argue that experienced planners have strong coded knowledge embedded in detailed sensory knowledge. Furthermore, we argue that the stronger the coded knowledge is, the more difficult it will be to change these codes. In this line of arguing we reason that less experienced planners will show a greater increase of coded knowledge than will more experienced planners who start to work with the DSS. In other words, novice planners will code their knowledge more quickly based on less sensory knowledge [rougher sensory knowledge] in comparison to more experienced planners.

Hypothesis 4

During the implementation of ZKR novice planners will show a bigger increase

of coded knowledge than experienced planners.

Using a similar line of reasoning for the personal characteristic of age as we have done for job experience we formulate

Hypothesis 5

During the implementation of ZKR younger planners will show more increase

of coded knowledge than older planners

A fourth personal characteristic that we expect to moderate the influence of the implementation of ZKR on knowledge types of individual users is the amount of hours per week that an individual planner works. We argue that the amount of contractual hours per week is a strong indicator for the use of sensory knowledge and coded knowledge [in opposite directions]. That is, the fewer hours per week an individual works the stronger this individual relies on the use of codes, as the amount of personal contact is restricted [this personal contact is a foundation for the development of detailed sensory knowledge]. However, these used codes are


113

not strongly embedded in firm, detailed sensory knowledge. Therefore, these codes are strongly related to particular situations. We argue that the use of these codes in different and changing context will be more difficult for individuals who work fewer hours per week. Thus, whereas they rely more heavily on codes, the codes they rely on are weaker. Consequently, we expect that during the implementation of ZKR the fewer hours per week an individual works the easier it will be to adopt the new way of working, but only in the narrow sense. That is, a vulnerable use of new codes. We formulate our last hypothesis as follows

Hypothesis 6

During the implementation of ZKR part time planners show less dynamics in

their coded knowledge than do full time planners

The hypotheses formulated above will be tested in chapter 9, the data analysis chapter. The following chapter [7] will stage the empirical setting of the present research in the Bartiméus organization after which, in chapter 8, we discuss the methodology to justify our choices made on the operational issues.


114

115

Chapter 7

Bartiméus

7.1 The organization

Bartiméus is a big health care institution in Doorn and Zeist in the centre of the Netherlands with over 7000 clients/patients in its care. Bartiméus receives about 70 million euro’s from the government and they receive additional revenues of about ten million euro’s. Almost three quarters of this total amount of 80 million euro’s goes to the 2000 employees.

Inspired by a Christian philosophy Bartiméus was founded in 1915. Its main aim is to support visually impaired children, young people and adults. Bartiméus realizes this aim in providing education, care and services to blind people and visually impaired people; it is not uncommon that impaired sight is combined with learning difficulties. An important part of Bartiméus’ philosophy is that it encourages independence and self-reliance on the part of visually impaired people in their own, familiar environment. Bartiméus has expertise in many different aspects of visually impairment; therefore, it also functions as a research center and an expertise center providing advice to third parties. The organization chart [see figure 7.1] shows the four focus points of Bartiméus, namely to live life, to live, to learn, and to work.

Bartiméus supports visually impaired people to live their life to the fullest. For example, Davey is a 10-year-old boy who lives in a hostel for multiply disabled children who get 24-hour care. Davey is blind and has difficulty with many things, such as eating. He cannot walk or talk and it is difficult for him to keep his head up straight. Both Davey and the employees of the hostel get support from Bartiméus. Bartiméus does research and gives advice about the supervision of the children. Bartiméus also supports people to live on their own.

Then, Bartiméus helps children and young people with school matters. For example, Willemijn is 19 years old. She is the only one at her school who has a walking stick for the blind. She has bad sight since her birth, but the last few years her disability has become worse. One eye is totally blind and the other eye can see for about ten percent. At school she gets help from the ‘Ambulante


116

Onderwijskundige Begeleiding’ [ambulant pedagogical supervision] of Bartiméus. Together with her mentor she reviews the adjustments that need to be made, for example for when she has to sit for exams. Bartiméus also specializes in supporting people in their jobs.

In 2006 Bartiméus merged with Sonneheerdt, which is now mainly covered by the Proson-division [see figure 7.1]. Proson offers a working environment ranging from a printing office for offset or Braille to making furniture. The main focus is to create an environment in which visually disabled people can function as normal as possible.

Figure 7.1: The organization chart of Bartiméus

Chapter 7 Bartiméus

117

7.2 Planning at Bartiméus: A situation for innovation

7.2.1 General vision on planning

Bartiméus considers planning a strategic aspect of its organization. The quality of planning influences the organization on multiple levels; this includes the use of capacity and in line of this employing new people, but also the practical efficiency of duty rosters. Essentially, the quality of the planning is reflected in the quality of the provided care; this implies attuning clients and professionals, with care and recourses. One of the ways to secure this is through a constant quality in the service provided by the professionals. The duty roster is the key element to ensure this.

7.2.2 Situation at Bartiméus before ZKR

Before the implementation of ZKR there was no clear planning policy. That is, all the separate units had one or two planners who all had their own way of planning. Implicitly, making the duty roster was considered to be an autonomous task per unit. Little was known among different units about their planning and the management had no insight into the [absence of] uniformity of planning. As a consequence, there was little insight in the use of the personnel capacity. That is, planning at the unit level was based on the vision of the planner and the unit, but there was no integrated vision or policy at the level of the organization. Thus, the organization neither imposed a general policy on planning nor did it formulate a general guideline or methodology for how to deal in general with the planning in practice. The planning greatly relied on the expertise of the planner.

A great advantage to this approach is the freedom that planners have to use their expertise with great nuance. However, the flip side to this exercising of sophistications is a - potential - great difference between the planning of the different units. Furthermore, in streaming the data flow in terms of budgeting and allocating resources the freedom of the planners to exercise their expertise was greatly impaired. The different ways of planning did not enable a general overview on the planning at Bartiméus, nor did they enable efficiency in data processing.

7.2.3 Planners at Bartiméus

Bartiméus underlines the central position of the client in the services that they offer. Therefore, they emphasize patience as an important cultural characteristic and working at Bartiméus one should be able to prioritize the interests of the


118

client/patient. This includes understanding ‘that undesired behavior is a way of communicating’ [site]. The support plan forms the central starting point to provide the care. The care is explicitly considered to be the result of teamwork. The planners who participated in our study all worked in the above described environment.

Let us look at an example of a planner at Bartiméus making the duty roster. Planner Laan works at a unit of Bartiméus, which provides care for visually disabled people. At this unit the clients live during the day and at night. The unit has on average twelve caretakers, which is average. These twelve caretakers help clients in their daily routine of living.

Laan has been a planner for about ten years now, so she gained much experience over the years. Making a duty roster now takes about three hours for her, although when there are not enough people available it takes much longer. Especially to find the right people in the recruitment pool then takes up much of her time.

The duty roster ideally has a three-month range. However, Laan never seems to manage this as too many changes occur over this period of time. She therefore plans for a period of five weeks. Laan starts with making the new duty roster about three weeks in advance, gathering information and checking the available capacity. She then starts with the actual planning. First come the weekends, these are the hardest to plan as it is important to her to grant as many wishes as possible. Sometimes it is not possible to grant any wishes. Now, with ten years of experience saying ‘no’ to her colleagues is much easier then in the beginning when she tried not to upset her colleagues with the duty roster choices that she made. Checking all the balances, horizontally for the employees as well as vertically for the days, takes up much of her time. There are many different codes, rules and exceptions, and mistakes in adding up the numbers are easily made. Laan hardly takes notice of the rules and regulations on planning. This is not doable, she says. These regulations were not made for us. People who are used to a certain pattern of working are not easily put in a different schedule just because someone says it is better. They do not see it that way. So, an important part of making the duty roster is also to know your colleagues, to know what they feel is important. That way you can make a qualitatively good duty roster while keeping your colleagues happy at the same time.

The introduction of ZKR would be a great help in the time consuming adding up of the numbers, which easily elicits mistakes. Laan is anxious to try the new

Chapter 7 Bartiméus

119

software, and curious what it can do. She is especially curious about the constraints in the rules and regulations. She wonders how many constraints she normally ‘breaks’ in making the duty roster.

7.2.4 Why innovate

As the above shows, before the implementation of ZKR Bartiméus did not have a strategic aim in planning. The management of Bartiméus sought for an integrated system, which enabled both simplification for the planners as well as enhancement of data processing, the planning from the units to the administration both personnel and financial. Introducing planning support software would create an overview in the use and allocation of resources and would secure a stable quality of provided care and good use of resources.

7.2.5 Other people involved in planning at Bartiméus

The first group of people to be involved in the planning are of course the planners themselves, but indirectly more people are involved as we have showed in the chapter on organizational innovation. These groups involve firstly the employees who are to be scheduled. As the above job description shows, the care provided by Bartiméus is considered to be teamwork. This implies that the duty roster is very important in achieving this goal and the employees directly experience the changes in the making of the duty roster.


120

121

Chapter 8

Methodology

8.1 Introduction

Discussing the methodology of a research involves discussing the empirical issues; in other words, describing and explaining the ‘actions that we undertook for the empirical research’ [de Groot 1961]. The present chapter therefore respectively discusses the design of the present research [8.1], the group that participated [8.2], the instruments used to operationalize our theoretical concepts [8.3] and how we went about conducting our research, discussing the procedure [8.4].

8.2 Design

Our research is based on the one-group pre-test – post-test design, a quasi-expe-rimental design [see figure 8.1a]. This design is based on within–individual treat-ment comparison in which one group of individuals undergoes treatment as op-posed to between-individuals, in which two or more different groups each get a different treatment. So in our study, 18 planners participated in all three measure-ments. In a within-individual design, one group of individuals will be subject of a pre-test [T0], then undergo the treatment [E] after which a post-test [T1] follows [see figure 8a]. The effect of the treatment will simply be established by exami-ning the [average] difference between T1 and T0 [Judd, Smith, & Kidder 1991].

T0 E T1

Figure 8.1a: One group pre-test – post-test design T0: pre-test, T1: post-test, E: event

The design of the present study is a variation on the one-group pre-test – post-test design; it involves two consecutive events. We therefore introduce a second post-test, one for each event. The first ‘event’ was the training on the new planning software ZKR [T1] and the second ‘event’ was the actual working with


122

the planning software ZKR, so the experience that the planners gained [T2, see figure 8.1b]. We did not include a control group for this study. A control group functions as a comparison, to evaluate the effect of the treatment. The main aim of our study is to understand the knowledge dynamics within a field setting. Therefore, our first interest is one of longitudinal concern rather than comparing a ‘treatment’ situation to a ‘non-treatment’ situation.

T0 training T1 experience T2

Figure 8.1b: Design present study: A variation on one group pre-test – post-test design

T0: pre-test, T1: 1st post-test, T2: 2nd post-test

We expanded our variation on the one group pre-test – post-test design by using three sub tasks of planning, namely gathering information, negotiating and [the actual] scheduling [see also chapter 5 on planning]. This resulted in a 3 [activity: gathering information, negotiating, scheduling] x 3 [measurement: 1, 2 and 3] within-subject design, consisting of eight cells, as the first measurement did not involve the subtask of scheduling. [see figure 8.1c].

Figure 8.1c: Research design

We chose this design [figure 8.1c] for the following reasons. First of all, the three different measure points in time provide the opportunity to establish possible knowledge changes during the innovation process. Furthermore, the first measurement functions as a base, which enables comparison of a non-innovational situation to an innovational situation. Secondly, the second

Chapter 8 Methodology

123

measurement was chosen to measure the effect of the training of ZKR. And the third measurement in time was chosen to measure the effect of actually working with ZKR. Then, to understand the effect of task domain we differentiated between three subtasks. Finally, a consideration to choose this design was a result of the organizational constraints to conduct empirical research.

8.3 Participants

The participants all worked at the care institution of Bartiméus. As the change in personnel was substantial over the one and a half years that we gathered the data for this study, it was not possible to have one and the same group of planners. Figure 8.2 shows an overview of all the participating planners and their ‘distribution’ over the three measurements. Before the implementation of ZKR 35 planners participated, 31 planners after the training with ZKR and 24 planners in the third measurement. In total, 18 planners participated in all three measurements.

Figure 8.2: Overview of participating planners over three measurements

8.4 Operationalization

8.4.1 Innovation

The independent variable of this study is innovation in the sense that we measured the effect of the innovation process at different points in time; we did not oppose an innovative situation to a non-innovative situation.

4

3

1

10

1

6

18

2nd measurement

n = 31

1st measurement

n = 35

3rd measurement

n = 24


124

We studied this variable at three values: 1] no innovation of the new planning software, the first measurement, 2] at the beginning of the innovation, the second measurement, at this time the planners had had training in ZKR, and 3] the innovation was operational, the planners had gained experience with ZKR over a period of six months, the third measurement.

The first measurement

During the first measurement there was no innovation and application of the planning software ZKR. Although some planners were aware of the coming implementation of ZKR, the planners had no specific idea of what was going to happen. Some planners were a bit skeptic as they suspected a great time investment would be needed to work with ZKR, which would be a waste of time as they were satisfied with the way things were [see also chapter 5]. However, some were eager to start working with ZKR. Thus the planners at this time were not confronted with the innovation and they made the duty roster by hand as they had always done. Therefore, their knowledge of planning had not been influenced by the innovation of the new planning support software ZKR.

The second measurement

At the time of the second measurement the implementation of ZKR had started. The planners had had their training on the new planning support software ZKR. The effect of this training was that the planners became more aware of the way that they [used to] plan and the consequences that the implementation of ZKR

could have on the way that they made the duty roster. Therefore, their knowledge of [the subtasks of] planning started to be influenced by the implementation of ZKR.

The third measurement

During the third measurement the planning support software of ZKR was operational. The planners had worked with ZKR for half a year, which changed their way of planning from making the duty roster by hand to making the duty roster with planning support software. In effect this meant that the planners had made at least three duty rosters with the new planning software ZKR. In studying the relationship between knowledge and innovation this measurement shows the impact of the applied innovation.


125

8.4.2 The three knowledge types

Chapter 4 discussed the three knowledge types, emphasizing sensory knowledge to emerge when difference is perceived, coded knowledge when communication is needed and theoretical knowledge as the essence of a concept is revealed. These three starting points form the foundation of our operationalization.

Sensory knowledge

Sensory knowledge is essentially knowledge through the senses, without the intervention of a code. This entails applied knowledge. Therefore, a key characteristic of this type of knowledge is that it needs demonstration in a face-to-face situation to communicate. In order to make a construct of sensory knowledge we derive other characteristics from this key characteristic [see the Appendix for details].

� Personal transfer of knowledge

� Learned through experience

� Communication through use of metaphors and analogies

Coded knowledge

Coded knowledge is essentially knowledge that frames; it is represented in a sign. This entails verbalization or another kind of sign to capture this knowledge. Therefore, a key characteristic of this type of knowledge is that it can be captured in words. In order to make a construct of coded knowledge we derive other characteristics from this key characteristic [see the Appendix for details].

� Can be verbalized

� Learned from handbooks

� Use of impersonal communication media, such as e-mail or other information systems.

Theoretical knowledge

Theoretical knowledge is essentially knowledge about the meaning of a concept. This entails reason and structure, taking different sides and putting the concept into perspective. Therefore, a key characteristic of this type of knowledge is that it can be reasoned about or with. In order to make a construct of coded knowledge we derive other characteristics from this key characteristic [see the Appendix for details].


126

� Answer the why-question

� Obtained during education

� Easy communication with lay-men

� Having a vision of how things should supposed to be

8.4.3 Moderating variables: Personal characteristics

Education was operationalized as the highest level of education that the planners had completed. Job experience was operationalized as number of years that the planners had had experience in their present job. Age was operationalized as number of years and contractual hours were operationalized as the amount of hours that the planners worked per week.

8.4.4 Questionnaire

The Castor-research group of the University of Groningen developed the questionnaire that was used in this study. This research group included specialists on semiotics, cognitive psychology and communication studies. Furthermore, the questionnaire was also co-developed by experts in the field of nursing as well as planning. The questionnaire was piloted on planners in the field [hospital planners - Leeuwarden].

The questionnaire for the first measurement contained three parts 1] a general part, about the planner [e.g. about education, experience, age etc.], the organization and questions about the implementation of ZKR, 2] a specific part about gathering information, 3] a specific part about negotiating. For part 2 and 3 the same cluster of questions was used in our questionnaire concerning the subtasks. These questions included questions on knowledge types, learning, information sources, expectation and communication. The Appendix shows the questionnaire used in the first measurement.

For all three measurements the same questionnaire was used. However, during the first measurement it turned out that filling out the questionnaire was very time consuming. As this did not improve the motivation of the planner for filling out the questionnaire the questionnaire for the second measurement was carefully narrowed down – we omitted the open questions – maintaining the possibility for comparison; so the questions in the questionnaire for the second measurement were the same as for the first measurement.


127

The second and third questionnaire included an additional set of [closed] questions on a third subtask, scheduling. Thus, the sub task of scheduling was not included in the first questionnaire.

8.4.5 Validity and reliability

Carmines and Zeller [1994: 4] distinguish validity and reliability as follows:

… while reliability focuses on a particular property of empirical indicators – the extent to which they provide consistent results across repeated measurements – validity concerns the crucial relationship between concept and indicator

The previous section described the operationalization of our three theoretical concepts. In terms of validity we state that this operationalization was based on content validity and face validity. That is, the operationalization was based directly on the theoretical framework, after which we consulted experts for a face validity check on this operationalization. We will elaborate on the reliability in the following chapter.

8.5 Procedure

Three questionnaires were sent out to the planners during the implementation process, 1] before the implementation of the planning software ZKR [January /February 2001], 2] after the training for the planning software [July / August 2001] and 3] approximately one and a half years after the first measurement, when the software program ZKR had been operational for about half a year – this comes down to making about three duty rosters using ZKR [July / August 2002].

Before the questionnaire was sent out the participants received a letter in which the aim of the research was explained [before the first questionnaire the planners also received a letter from the organization]. After that the questionnaire was sent out. When the participants received the questionnaire, the ones that did not return it were approached and kindly requested to fill in the questionnaire. When there were any difficulties with the filling in of the questionnaire the participants could contact us. The third questionnaire was personally handed to the planners, which had the advantage of being able to directly communicate with the planners.


128

129

Chapter 9

Results and Data Analysis

9.1 Introduction

In chapter 6 nine hypotheses were formulated derived from the cognitive-semiotic model presented in chapter 4. We discuss the data in the following order. First we discuss the data at a descriptive level; that is, we provide a profile of the planners that participated in our study [9.2] and we present the general means and standard deviations to the three knowledge types [9.3]. Secondly, we subject the nine formulated hypotheses to the data obtained for the three knowledge types of sensory knowledge, coded knowledge and theoretical knowledge [9.4]. We then subject the data to more, exploratory, analyses [9.5]. The final section of the present chapter summarizes the main findings [9.6]. The last chapter of this thesis, subsequent to this chapter, elaborates on the implications of the results put forward in the present chapter.

9.2 Planners’ profile

In total 43 planners participated in our study. However, not all these planners participated in each of the three measurements; only 18 planners did. As the primary aim of our study is to understand the knowledge dynamics of individual planners in an innovative organizational setting we want to compare withinsubject. Thus, we focus on the 18 planners who participated in all three measurements. The first measurement was before the implementation process of ZKR had started, the second measurement was just after the planners had had their training on ZKR and the third measurement was after the planners had had half a year of experience with ZKR.

The 18 planners had an average age of 34.6 [SD: 9.0 years] and an average job experience in this position of 9.5 years [SD: 6.7 years]. Figure 9.1a shows the numbers of planners of 30 years of age and younger [N = 7] and planners who are older than 30 years of age [N = 11]; figure 9.1b shows the amount of novice planners [N = 7] and experienced planners [N = 11]. The planners were relatively highly educated: 13 planners [72 %] were educated at higher level,


130

which corresponds to a bachelor’s degree or higher professional vocational education [in the Netherlands: HBO]; the other five planners [28 %] were educated at senior level, which corresponds to senior vocational education [in the Netherlands: MBO; see figure 9.1c]. Regarding their position within the organization of Bartiméus 16 planners were unit members [88 %] and twoplanners were head of the team for which they worked [12 %]. More than half of the planners worked 32 hours up to full time [10 planners equals 56 % - see also figure 9.1d] and eight planners worked part time [44%]. All but one planner had prior experience with computers. However, this experience was limited to word processing programs such as Word and Word Perfect.

Figure 9.1a: Age of the planners: 30 and younger versus older than 30

Figure 9.1c: Education: senior versus higher

Figure 9.1b: Job experience: novice versus experienced

Figure 9.1d: Contractual hours per week: part time versus full time

In short, the planners started working in their current positions at the average age of 25 years. At the time of our study they were around 35 years of age with an average job experience of almost ten years. More than half of this group of

Chapter 9 Results and Data Analysis

131

relatively highly educated planners – in comparison to other groups of planners studied in the Netherlands Railways [Kiewiet et al. 2005] – worked at least 32 hours [= four days] per week.

9.3 Some preliminary data

In the previous chapter we presented the operationalization of the three con-cepts of the knowledge types based on content validity and face validity. In order to determine the reliability of the suggested operationalization into the constructs, we conducted a factor analysis after which we calculated Cronbach’s alpha for the three knowledge-type constructs. We will discuss the calculation in this section, followed by a descriptive analysis of the data of the first measurement.

9.3.1 Reliability of the constructs

Focusing on reliability our aim was to explore the options of building three stable theoretical constructs from the questions of the questionnaire. Hereto, we took all questions – related to the three different knowledge types – together [43 in total] and subjected this group of questions to a factor analysis. Rather than suggesting three factors that each represented one of the three knowledge types of sensory, coded, and theoretical knowledge, the factor analysis showed a more ambiguous result. The analysis showed that the first three factors only explained 54 percent of the variance [six factors would explain 85 percent]. Also, the three suggested factors contained cross loaded variables.

Therefore, based on this first factor analysis, we selected a group of questions and we conducted three additional factor analyses, one separate factor analysis for each of the three knowledge type constructs. These three separate factor analyses did not suggest three unambiguous knowledge type constructs either.

In determining the three knowledge type constructs our next step was to calculate Cronbach’s alpha for the three separate knowledge types. We therefore made two assumptions:

1. We expect a similar configuration for the three subtasks, which would imply that the data over the subtasks are stable

2. We expect a similar configuration for the three measurements, which would imply that the data over the measurements are stable


132

We used the factor analyses as a starting point to revalue the selection for our three knowledge type constructs. The three groups of questions for which the calculated Cronbach’s alpha met our two assumptions [see table 9.1a].

The factor analyses showed

� no clear cut distinction between the three knowledge types

� sensory knowledge was the most clear of all three factors

� reevaluating the highly inter correlated questions, some could not be founded properly enough in theory

� some question that were expected to correlate with one knowledge type in fact showed a higher correlation with another knowledge type

� coded knowledge was the least clear of all three clusters

The calculation of Cronbach’s alpha showed

� acceptable correlations that met our two formulated assumptions

So, although the factor analysis was not satisfactory, the calculation of Cronbach’s alpha did seem to meet our standards. However, reconsidering the three groups of questions on which Cronbach’s alpha had been calculated the outliers [based on the factor analyses] that we excluded, in fact included the questions that we on forehand determined to be the key-questions of the three knowledge types.

Table 9.1a: Cronbach’s alpha scores for the three knowledge types groups of questions [key questions not included]

Sensory Coded Theoretical

1ST 3RD 1ST 3RD 1ST 3RD

Gathering information .629 .724 .657 .570 .380 .653

Negotiating .799 .630 .746 .270 .610 .877

We considered three explanations for these results. Firstly, the data could contain random errors, which we can partly explain by the relatively small N. Secondly, the data could contain systematical errors, which could be explained by possible


133

misinterpretation of the questions by the subjects. Thirdly, the validity of the construct is too low. Although these three explanations are all plausible and a mixture of these three explanations could also be in order we cannot – at this point – be sure, which explanation is leading. This left us with a dilemma, choosing between validity, as a theoretical fundament, and reliability, coherence between the set of questions. Therefore, to be safe we chose to prioritize validity above reliability. Although, a group of questions sounds more attractive as an operationalization than does just one question, we did do just that. So we continued the data analysis and the hypotheses testing operationalizing the three knowledge types with the one key question.

Table 9.1b: Mean scores and SD’s for the three knowledge types for the 1ST

measurement split on the four personal characteristics

Sensory Coded Theoretical

M SD M SD M SD

Overall 3,28 0,69 2,67 0,75 4,26 0,46

Age Younger 3,00 0,70 2,79 0,81 4,14 0,48

Older 3,45 0,65 2,59 0,74 4,34 0,45

Difference +0,45 -0,20 +0,20

Education Senior 3,30 0,57 2,20 0,76 4,40 0,42

Higher 3,27 0,75 2,85 0,69 4,21 0,48

Difference +0,03 +0,66 -0,19

Job experience

Novice 3,07 0,73 2,79 0,81 4,29 0,57

Experienced 3,41 0,66 2,59 0,74 4,25 0,40

Difference +0,34 -0,20 -0,04

Contractual Hours

Part time 3,50 0,65 2,56 0,86 4,35 0,52

Full time 3,10 0,70 2,75 0,68 4,20 0,42

Difference -0,40 +0,19 -0,15


134

9.3.2 Means and standard deviations split on personal characteristics

Before we test our hypotheses we present the mean values and standard deviations for the three knowledge types, split on the four personal characteristics of age, education, job experience and contractual hours [see also table 9.1b]. We added an extra row per personal characteristic, which shows the difference in mean score between the two [sub] groups. The data resemble the first measurement.

As the table shows the greatest difference is found between the coded knowledge of higher educated planners [HBO] versus senior educated planners [MBO]. Overall, theoretical knowledge shows the smallest difference in score. An interesting contrast forms the group of senior educated planners. Whereas they show the lowest score on coded knowledge compared to all the other grouping of characteristics they show the highest score of theoretical knowledge compared to all the other groups; that translates to a theoretical knowledge that is twice as high as their coded knowledge.

Table 9.1c: Terminology and abbreviations used for in the present chapter

1 / 2 The 1ST measurement compared to the 2ND measurement:The effect of the training with ZKR

2 / 3 The 2ND measurement compared to the 3RD measurement:The effect of the experience with ZKR

1 / 3 The 1ST measurement compared to the 3rd measurement:The effect of the implementation of ZKR

All tasks All the subtasks together / over all subtasks

GI Gathering information

N Negotiating

S Scheduling

Sens Sensory knowledge

Cod Coded knowledge

Theor Theoretical knowledge


135

9.4 Hypotheses testing

To test the hypotheses we focused on the effects of both the training [the first measurement versus the second measurement] and the experience with ZKR [the second measurement versus the third measurement] as well as the effect of thewhole implementation process [the first measurement versus the third measurement]. Table 9.1c shows the three phrases that we use in the followingsection for these effects. Furthermore, the table shows the abbreviations, for the knowledge types and for the subtasks, that we will use in this chapter to accompany tables and figures. To test the hypotheses we used a paired sampled t-test. The first set of hypotheses is concerned with the main effects of the three knowledge types. Our expectations on sensory knowledge are formulated as follows.

1ST 2ND 3RD

M SD M SD M SD

All tasks 3,28 0,69 3,39 0,86 3,31 0,53

Gathering information 3,50 0,86 3,33 0,91 3,39 0,92

Negotiating 3,00 0,69 3,28 1,02 3,18 0,86

Scheduling 3,58 0,98 3,29 0,67

Figure 9.2 and table 9.2a: Mean scores for sensory knowledge for three subtasks over three measurements


136

9.4.1 Hypotheses 1

Hypothesis 1a

The implementation process of ZKR will lead to a decrease of sensory


Figure 9.2 and table 9.2a below show the mean scores and standard deviations for sensory knowledge over all three measurements for all three subtasks together [first row] and for the three subtasks separately. Figure 9.2 shows an [non-significant] increase of sensory knowledge during the implementation of ZKR

[d = 0.04, t = 0.195, df = 17, p = 0.848].

Table 9.2b: Results of the t-test for sensory knowledge

Sensory �� t df p

All tasks Training 0,11 0,456 17 0,654

Experience -0,74 -0,345 17 0,734

Implementation 0,04 0,195 17 0,848

Gathering information Training -0,17 -0,644 17 0,528

Experience 0,06 0,181 17 0,859

Implementation -0,11 -0,524 17 0,607

Negotiating Training 0,28 1,045 17 0,311

Experience -0,10 -0,338 17 0,740


Scheduling Experience -0,26 -1,042 17 0,312

At the level of the subtasks we see an inconsistent pattern. Gathering informationshows a decrease for sensory knowledge [d = -0.11, t = -0.524, df = 17, p = 0.607] and negotiating shows an increase for sensory knowledge [d = 0.18, t = 0.66, df = 17, p = 0.520]. See table 9.2b for the other t-test results. In other


137

words, the sensory knowledge of planners does not seem to decrease when they start to make their duty rosters with planning support software. Hypothesis 1acould, therefore, not be corroborated. Our expectations on coded knowledge are formulated as follows.

Hypothesis 1b

The implementation process of ZKR will lead to an increase of coded


Figure 9.3 and table 9.3a below show the mean scores and standard deviations for coded knowledge over all three measurements for all three subtasks together [first row of table 9.3a] as well as for the three subtasks separately. The graph shows an increase of coded knowledge for the implementation for all three subtasks together. The t-test confirms this effect [d = 0.34, t = 1.891, df = 17, p = 0.076]. The effect of just the training on ZKR also shows a significant increase of coded knowledge [d = 0.43, t = 1.811, df = 17, p = 0.088]; the experience with ZKR shows a [non-significant] decrease [d = -0.09, t = -0.401, df = 17, p = 0.694]. Thus, during the whole implementation of ZKR, coded knowledge increases, in line with hypothesis 1b.

At task level the graph shows – over the whole implementation stage – an increase for the subtasks gathering information and negotiating; negotiating shows a significant increase [d = 0.56, t = 1.968, df = 17, p = 0.066*]. Interestingly, gathering information also shows a significant increase on coded knowledge, for the training on ZKR [d = 0.78, t = 2.961, df = 17, p = 0.066*]. However, the coded knowledge of gathering information also shows a decreasing effect for the experience with ZKR [d = -0.56, t = -1.966, df = 17, p = 0.009***]. These two contrasting, significant, effects balance each other in that the increase of coded knowledge disappears over the whole implementation of ZKR [d = 0.22, t = 1.074, df = 17, p = 0.298]. The coded knowledge for the subtask of scheduling showed a [non-significant] decrease of coded knowledge for experience with ZKR

[d = -0.06, t = -0.215, df = 17, p = 0.832]. See table 9.3b below for the other t-test results.


138

Coded 1ST 2ND 3RD

M SD M SD M SD

All tasks 2,67 0,75 3,10 0,93 3,01 0,77


Negotiating 2,57 0,74 2,71 1,13 3,13 1,08

Scheduling 3,00 1,14 2,94 0,87

Figure 9.3 and table 9.3a: Mean scores and SD’s for coded knowledge for three subtasks for three measurements

In other words, the implementation of ZKR has the effect of an increase in the coded knowledge of the planners. This effect manifests itself most prominently in the subtask of negotiating; this task shows a linear increase of coded knowledge and a significant increase over the whole implementation process. Then, for gathering information, the planners seem to be able to better verbalize their knowledge after the training on ZKR. But this amelioration on verbalization almost disappears after the planners have had half a year of experience with ZKR. Only a shred of an increasing trend seems to be left. The coded knowledge on scheduling does not seem to be affected by the experience with ZKR.

Our expectations on theoretical knowledge are formulated as follows.


139

Table 9.3b: Results of the t-test for coded knowledge

Coded �� t df p

All tasks Training 0,43 1,811 17 0,088*

Experience -0,09 -0,401 17 0,694

Implementation 0,34 1,891 17 0,076*

Gathering information Training 0,78 2,961 17 0,009***

Experience -0,56 -1,966 17 0,066*



Experience 0,43 1,312 17 0,207

Implementation 0,56 1,968 17 0,066*

Scheduling Experience -0,06 -0,215 17 0,832

Hypothesis 1c

The implementation process of ZKR will lead to an increase of theoretical


Figure 9.4 and table 9.4a show the mean scores and standard deviations for theoretical knowledge over all three measurements for all three subtasks together [first row] and for the three subtasks separately. The graph appears to show that the theoretical knowledge of the planners decreases during the implementation of ZKR. The mean scores for the three subtasks together show a very small increase of theoretical knowledge for the training on ZKR [d = 0.0, t = 0.511, df = 17, p = 0.616] and a small decrease for the experience with ZKR [d = -0.17, t = -1.319, df = 17, p = 0.205]. All together, the theoretical knowledge of the planners [non-significantly] decreases during the implementation of ZKR [d = -0.10, t = -0.703, df = 17, p = 0.491].

At the level of the subtasks we see a [non-significant] decrease for both gathering information [d = -0.18, t = -0.634, df = 16, p = 0.534] and negotiating [d = -0.21, t = -0.929, df = 16, p= 0.366] on theoretical knowledge during the


140

implementation of ZKR. Scheduling shows an increase for theoretical knowledge for the experience with ZKR [d = 0.05, t = 0.369, df = 16, p= 0.716]. In other words, the theoretical knowledge of planners did not change during the implementation of ZKR. We did, however, see a consistent decrease of theoretical knowledge [see table 9.4b for the other t-test results].

Theoretical 1ST 2ND 3RD

M SD M SD M SD

All tasks 4,27 0,46 4,33 0,55 4,17 0,55


Negotiating 4,41 0,60 4,44 0,62 4,20 0,98

Scheduling 4,35 0,68 4,40 0,57

Figure 9.4 and table and figure 9.4a: Mean scores for theoretical knowledge for three subtasks over three measurements

Summarizing hypothesis 1, overall, during the implementation of ZKR the coded knowledge of the planners increases. The separate subtask of negotiating shows this effect as well. Gathering information shows an increase of coded knowledge, however only for the training, while the experience with ZKR for gathering information even shows a decrease of coded knowledge. The implementation neither shows a significant increase nor a significant decrease for both theoretical


141

knowledge and sensory knowledge. Theoretical knowledge does show a decreasing trend overall, for gathering information and for negotiating on experience as well as over the whole implementation process. Sensory knowledge shows the least consistent change patterns of all three knowledge types.

Table 9.4b: Results of the t-test for theoretical knowledge

Theoretical �� t df p


Experience -0,17 -1,319 17 0,205

Implementation -0,10 -0,703 17 0,491

Gathering information Training 0,10 0,433 17 0,670

Experience -0,28 -0,813 17 0,428

Implementation -0,18 -0,634 17 0,534


Experience -0,24 -0,887 17 0,388

Implementation -0,21 -0,929 17 0,366

Scheduling Experience 0,05 0,369 17 0,716

Hypothesis 2 combines hypothesis 1a and hypothesis 1b; that is, we hypothesized an increase for coded knowledge in combination with a decrease for sensory knowledge. We formulate hypothesis 2 as follows.

9.4.2 Hypothesis 2

Hypothesis 2

The implementation process of ZKR will lead to a conversion of knowledge

from sensory knowledge into coded knowledge


142

Hypothesis 1b showed that the implementation of ZKR resulted in an increase for the coded knowledge of the planners; however, hypothesis 1a showed that this was not combined with a decreasing effect for their sensory knowledge.

Conversion 1ST 2ND 3RD

All tasks coded 2,67 3,10 3,01

All tasks sensory 3,28 3,39 3,31

GI coded 2,78 3,56 3,00

GI sensory 3,50 3,33 3,39

N coded 2,57 2,71 3,13

N sensory 3,00 3,28 3,18

S coded 3,00 2,94

S sensory 3,56 3,29

Figure 9.5 and table 9.5a: The mean values for the three knowledge types for all subtasks, as well as for all three measurements

[To optimize the visibility of the direction of the trends we changed the range of the scale of figure 9.5 from 1 - 5 to 2,5 - 3,7]

We now want to investigate whether increases of coded knowledge are accompanied by decreases of sensory knowledge, apart from the size of the change. In other words, we focus on the combination of directions of the


143

knowledge change rather than on the size of the knowledge change [in p-values]. Figure 9.5 and table 9.5a show the mean scores for all subtasks together and the three subtasks separately. When we take a closer look at the coded knowledge and the sensory knowledge for all subtasks [the two column groups on the left of figure 9.5] we see for both sensory knowledge and coded knowledge an increase for the training [black and grey columns] Then for the experience with ZKR [greyand white columns] we see a decrease for both sensory knowledge and coded knowledge. Thus, over all subtasks do neither the training on ZKR nor the experience with ZKR show the expected knowledge conversion pattern as formulated in hypothesis 2, an increase of coded knowledge in combination with a decrease of sensory knowledge.

To test hypothesis 2 there are ten possible combination patterns showed in figure9.5. Table 9.5b specifies these ten combinations.

Table 9.5b: Overview of the possible combination patterns to test the conversion of sensory knowledge into coded knowledge

Conversion Training effect[black-gray]

Experience effect[gray-white]

Implementation effect

[black-white]

All tasks together X X X

Gathering information X X X

Negotiating X X X

Scheduling X

Each of these ten combinations can be placed in one of four combination-categories [see table 9.5c], that is, a 2 [increasing coded knowledge or decreasing coded knowledge] x 2 [increasing sensory knowledge or decreasing sensory knowledge] table:

� An increase of coded knowledge combined with an increase of sensory knowledge

� An increase of coded knowledge combined with a decrease of sensory knowledge


144

� A decrease of coded knowledge combined with an increase of sensory knowledge

� A decrease of coded knowledge combined with a decrease of sensory knowledge

For the purpose of testing hypothesis 2 we are interested in the second group. Table 9.5c [below] places all ten combinations in one the four quadrants; each quadrant is divided into twelve cells, three columns for effect [training, experience, and implementation] x four rows for task [all tasks together, gathering information, negotiating, scheduling]. Table 9.5c also shows the effect size [in p-values] for the differences in knowledge. For example, quadrant 3 shows onecombination; decreasing coded knowledge with a p-value of .07 together with increasing sensory knowledge with a p-value of .86. This combination applies to the effect of experience with ZKR for the subtask of gathering information. Quadrant II supports hypothesis 2 and quadrant III contradicts hypothesis 2. Both quadrant I and quadrant IV combine a knowledge type change in line with hypothesis together with a knowledge type change contradictory to hypothesis 2.

Table 9.5c shows three combinations in line with hypothesis 2 [quadrant II], one combination which contradicts hypothesis 2 [quadrant III], and 6 combinations that have both a change direction in line with hypothesis 2 as well as a change direction that contradicts hypothesis 2. Thus, table 9.5c shows that three of the ten possible combinations are in line with hypothesis 2. These three combinations concern gathering information [two times] and negotiating [once] and they refer to the training, the experience as well as the overall implementation. We conclude that our expectation on the conversion of coded knowledge into sensory knowledge is not supported.

The hypotheses 1 and 2 concerned the main effects of the three knowledge types. In sum, the coded knowledge of the planners increased when they started to make their duty rosters with ZKR, the training that the planners had on ZKR

[not the experience] mainly caused this effect. In fact, the coded knowledge during the experience that these planners had with ZKR reduced the overall effect, over the whole implementation period. Neither sensory knowledge nor theoretical knowledge either increased or decreased significantly, although we did see that the theoretical knowledge patterned a decrease. Finally, we did not find support for the assumption that planners who started to work with ZKR

converted their sensory knowledge into coded knowledge.


145

Hypotheses 3 to 6 formulate expectations on the moderating effect of four planner-related-characteristics on the relationship between the knowledge types and the implementation process of ZKR. One of these four characteristics is job experience. So, for example, we expect the knowledge change pattern of novice planners to differ from the knowledge change pattern of experienced planners.

Table 9.5c: P-values the effects of training [1/2], experience [2/3] and implementation [1/3] for sensory knowledge and coded knowledge

� = increase, � = decrease

� sensory � sensory

1 / 2 1 / 3 2 / 3 1 / 2 1 / 3 2 / 3

� coded All C: .09*S: .66

C: .08*S: .85

GI C: .01*S: .53

C: .30S: .61

N C: .64S: .31

C: .07S: .52

C: .21S: .74

S I II

� coded All C: .07*S: .86

III IV C: .69S: .73

GI

N

S C: .83S: .31

The four characteristics on which we formulated expectations of their moderating effect are respectively education [senior vocational versus higher vocational education: hypothesis 3], job experience [novice versus experienced: hypothesis 4], age [younger versus older: hypothesis 5] and contractual hours per week [part time versus full time: hypothesis 6]. To illustrate the testing procedure that we followed we will first formulate hypothesis 3a and use this hypothesis as an example.


146

9.4.3 Hypotheses 3

Hypothesis 3a

During the implementation process of ZKR planners with higher professional

vocational education will show a stronger increase of their sensory knowledge

in comparison to planners with senior vocational education.

Hypothesis 3a formulates the expectation that the education of planners influences their coded knowledge when they start to work with ZKR. More specifically, we expect that starting to work with ZKR will cause a greater increase in the sensory knowledge of higher educated planners than in the sensory knowledge of senior educated planners. Table 9.6a shows fictitious raw data – as illustration to explain the computational procedure – for the sensory knowledge of four planners. The rows present the data per planner, the first column [starting from the left] showing the scores for sensory knowledge for the first measurement, the second column showing the scores for the second measurement and the third column showing the scores for the third

1st 2nd 3rd Education 1/2 2/3 1/3

Planner 1 3 4 5 Higher 1 1 2

Planner 2 1 2 3 Higher 1 1 2

Planner 3 4 5 4 Senior 1 -1 0

Planner 4 4 5 5 Senior 1 0 1

Table 9.6a: Fictitious raw data for the sensory knowledge of four planners to

illustrate the data used for the different hypotheses

Table 9.6b: New variables: The difference scores derived from table 9.6a

measurement for sensory knowledge. Table 9.6b displays the difference scores differentiated from table 9.6a; the 1/2-column shows the score differences between the first and the second measurement – the effect of the training, the 2/3-column shows the score differences between the second and third

Hyp

othe

sis 3

a

Hypothesis 1a


147

measurement – the effect of the experience with ZKR – and the 1/3-column shows the score differences between the first and third measurement – the effect of the whole implementation.

Hypothesis 1 was tested on the raw data, as presented in table 9.6a. For example, to test the effect of the training on the sensory knowledge [hypothesis 1a] we compared the first column to the second column using a dependent t-test. Hypotheses 3 to 6 will test differentiated data, as in table 9.6b. For example, to test the difference in effect of the training on sensory knowledge between higher educated planners and senior educated planners [hypothesis 3a] we compared the first two rows of column 1/2 to the third and fourth row of column 1/2.

Education sensory 1ST 2ND 3RD

M SD M SD M SD

Senior 3,30 0,57 3,00 1,05 3,27 0,64

Higher 3,27 0,75 3,54 0,76 3,33 0,51

Figure 9.6a and table 9.6d: Mean values and SD’s for sensory knowledge over three measurements split on education

So in fact, to test the difference in sensory knowledge between the higher and senior educated planners we calculated ten new difference-score-variables [see also table 9.4b]. Note that table 9.6a only shows the three new variables for the overall sensory knowledge; one level deeper we also have the sensory knowledge


148

of the three subtasks – seven additionally. The testing of hypothesis 1 differs from hypotheses 3 to 6 on the following points

1. Hypothesis 1 is based on raw data, hypotheses 3 to 6 are based on differentiated data

2. Hypothesis 1 is t-tested as dependent and hypotheses 3 to 6 are t-tested as independent

3. Hypothesis 1 compares within subjects and hypotheses 3 to 6 compare between subjects

Figure 9.6a and table 9.6d show the mean scores and standard deviations for sensory knowledge split on education. The education of the planners has no

effect on the overall implementation of ZKR [�� = 0.003, t = -0.223, df = 16, p = 0.826]. The graph does show that the education of the planners affects their sensory knowledge after the training on ZKR; however, this effect does not hold in

the t-test [�� = 0.064, t = -1.050, df = 16, p = 0.310]. When the planners have had experience with ZKR the knowledge change caused by the training on ZKR

disappears [�� = 0.057, t = 0.983, df = 16, p = 0.340]. Table 9.6e shows an overview of all the t-test scores. Thus,

• Overall education does not affect the innovation – sensory knowledge relation

• The combination of change directions caused by the training confirms that higher vocational planners show a stronger increase of sensory knowledge than do senior vocational planners

When we go one level deeper [see figure 9.6b and table 9.6e] the sensory knowledge pattern of senior vocational planners for gathering information is most striking: a decrease of sensory knowledge caused by the training and then an increase after experience with ZKR. In contrast, the sensory knowledge of the higher vocational planners hardly seems to have been affected by the implementation of ZKR. Although the graph shows this effect for education, the t-

test only shows a trend for training [the training: �� = 0.135, t = -1.583, df = 16,

p = 0.133, and the experience [�� = 0.071, t = 1.105, df = 16, p = 0.286].

The sensory knowledge change patterns for negotiating differ from gathering information in that the higher vocational planners show an effect for the trainingand the senior vocational planners hardly seem to be affected by the

implementation of ZKR. This does not influence education however [training: �� = 0.025, t = -0.637, df = 16, p = 0.533 – see table 9.6f for the other t-test results].


149

Education sensory 1ST 2ND 3RD

M SD M SD M SD

GI Senior 3,60 0,89 2,80 0,84 3,40 0,89

GI Higher 3,46 0,88 3,54 0,88 3,38 0,96

N Senior 3,00 1,00 3,00 1,41 3,04 0,08

N Higher 3,00 0,58 3,39 0,87 3,23 1,01

S Senior 3,20 1,48 3,06 0,72

S Higher 3,69 0,75 3,39 0,65

Figure 9.6b and table 9.6e: Mean values for sensory knowledge split on education

The sensory knowledge change pattern for scheduling does not seem to be affected by the education of the planners; although the higher vocational planners score higher on sensory knowledge their change pattern does not differ from the

change pattern of senior vocational planners [�� = 0.005, t = 0.219, df = 16, p = 0.776]. Table 9.6f gives an overview of all the t-test scores.

Thus, testing hypothesis 3a for the effect of the education of the planners on the relation between their sensory knowledge and the implementation of ZKR we conclude

� Education does not have a significant effect on the sensory knowledge of planners


150

� Although not significant, all implementation knowledge change patterns and training knowledge change patterns are in line with hypothesis 3a – the whole implementationprocess of ZKR and the training on ZKR separately cause a stronger increase of sensory knowledge for higher vocational planners than for senior vocational planners

� All knowledge change patterns on experience with ZKR

contradict hypothesis 3a

Table 9.6f: T-test scores for sensory knowledge split on education for all three subtasks and all differences in measurements

�� t df p

All tasks Training 0,064 -1,050 16 0,310

Experience 0,057 0,983 16 0,340

Implementation 0,003 -0,223 16 0,826

Gathering information Training 0,135 -1,583 16 0,133

Experience 0,071 1,105 16 0,286

Implementation 0,004 -0,253 16 0,804

Negotiating Training 0,025 -0,637 16 0,533

Experience 0,005 0,275 16 0,787

Implementation 0,006 -0,317 16 0,755


In addition to the effect of education that we expected on sensory knowledge we also expected an effect of education on coded knowledge. In contrast to our expectations for sensory knowledge, we formulated the following hypothesis for the effect of education on coded knowledge.


151

Hypothesis 3b

During the implementation process of ZKR senior vocationally educated

planners will show a stronger increase of coded knowledge in comparison to

higher vocationally educated planners.

Figure 9.7a and table 9.7a below show the mean scores and standard deviations for coded knowledge split on education.

The graph suggests a greater effect on coded knowledge for the senior vocational planners compared to the higher vocational planners, but this effect does not

show in the t-test [training: �� = 0.052, t = 0.938, df = 16, p = 0.362;

implementation: �� = 0.073, t = 1.121, df = 16, p = 0.279]. Secondly, the trainingappears to have a greater effect than the experience with ZKR, for both the higher educated and the senior educated planners [see table 9.7c for an overview of all the t-test results]. Thus, we sum up

Education coded 1ST 2ND 3RD

M SD M SD M SD

Senior 2,20 0,76 3,00 0,97 2,87 0,69

Higher 2,85 0,69 3,14 0,95 3,06 0,82

Figure 9.7a and table 9.7a: Mean values and SD’s for coded knowledge split on education


152

� Overall education does not affect the innovation – coded knowledge relation

� The training on ZKR has a greater effect on coded knowledge than does the experience with ZKR

� The knowledge change direction of the implementation and the training are in line with hypothesis 3b – the implementation of ZKR causes a stronger increase of coded knowledge for the senior vocational educated planners than it does for the higher vocational educated planners

Education coded 1ST 2ND 3RD

M SD M SD M SD

GI Senior 2,20 1,10 3,40 0,89 3,00 1,22

GI Higher 3,00 0,91 3,62 0,87 3,00 0,91

N Senior 2,51 1,12 2,40 0,89 2,83 0,85

N Higher 2,59 0,61 2,82 1,21 3,25 1,16

S Senior 3,20 1,30 2,79 0,83

S Higher 2,92 1,12 3,00 0,91

Figure 9.7b and table 9.7b: Mean values for coded knowledge split on education


153

When we go one level deeper [see figure 9.7b and table 9.7b], we see that the effect for training over all tasks [figure 9.7a] is probably mainly caused by the subtask of gathering information, which shows similar patterns. The training on

ZKR does not seem to be affected by the education of the planners [�� = 0.058, t = 0.997, df = 16, p = 0.334], but the whole implementation process is affected

by the education of the planners [�� = 0.176, t = 1.850, df = 16, p = 0.083]; the increase of the coded knowledge of the senior vocational educated planners is bigger than that of the higher vocational educated planners [in line with hypothesis 3b].

In contrast, for negotiating the higher vocationally educated planners seem to show a stronger increase of coded knowledge than do the senior vocational

planners during the implementation of ZKR [�� = 0.017, t = -0.531, df = 16, p = 0.603]. Scheduling also shows a contradictory pattern to hypothesis 3b, but not

strong enough to show in the t-test [�� = 0.038, t = -0.793, df = 16, p = 0.440].

Thus, for hypothesis 3b, the effect of the education of the planners on the relation between their coded knowledge and the implementation of ZKR, we conclude

� The knowledge change direction set by the education on coded knowledge over all tasks, for the whole implementation process as well as the training on ZKR, is conform our expectation – senior educated planners have stronger increasing coded knowledge than do higher educated planners

� Education does not differentiate the planners’ coded knowledge of gathering information

� The change directions of negotiating and scheduling, although not significant, are contradictory to hypothesis 3b; they show a greater increase of coded knowledge for the higher educated planners compared to the senior educated planners

� Gathering information shows a knowledge change direction opposite to those of negotiating and scheduling

Interestingly, over all three subtasks we see a decrease of sensoryknowledge in combination with an increase of coded knowledge for the senior vocationally educated planners for the effect of the


154

training. This pattern manifests itself also for the subtask of gathering information.

Table 9.7c: T-test scores for coded knowledge split on education for all three subtasks and all difference in measurements

Education coded �� t df p

All tasks Training 0.052 0.938 16 0,362

Experience 0,001 -0,103 16 0,919



Experience 0,007 0,332 16 0,744

Implementation 0,176 1,850 16 0,083*

Negotiating Training 0,017 -0,529 16 0,604

Experience 0,000 -0,001 16 0,999

Implementation 0,017 -0,531 16 0,603

Scheduling Experience 0,038 -0,793 16 0,440

Hypothesis 4 formulates the expected effects of the job experience on the relationship between innovation and coded knowledge. We formulate this hypothesis as follows.

9.4.4 Hypothesis 4

Hypothesis 4

During the implementation of ZKR novice planners will show more increase of

coded knowledge than experienced planners.

Figure 9.8a and table 9.8a below show the mean scores and standard deviations for coded knowledge split on job experience.


155

Job experience coded 1ST 2ND 3RD

M SD M SD M SD

Novice 2,79 0,81 3,14 0,92 3,31 1,06

Experienced 2,59 0,74 3,08 0,98 2,82 0,48

Figure 9.8a and table 9.8a: Mean values for coded knowledge split on job experience

The graph shows an increase in coded knowledge for both the novice planners and the experienced planners, after the training on ZKR as well as during the whole implementation of ZKR. From the graph there appears to be an effect for job experience for the implementation as well as for experience, but the t-test

does not show this effect [training: �� = 0.004, t = -0.252, df = 16, p = 0.804;

implementation: �� = 0.037, t = -0.785, df = 16, p = 0.444].

When we go one level deeper [see figure 9.8b and table 9.8b], we see the greatest difference between novice planners and experienced planners for the subtask of negotiating. In fact, the t-test shows an effect for job experience of planners on their coded knowledge; during the implementation process of ZKR

novice planners show a stronger increase in their coded knowledge than do the

experienced planners [�� = 0.297, t = 2.600, df = 16, p = 0.019]. The experience

with ZKR showed a trend [�� = 0.123, t = 1.501, df = 16, p = 0.153].


156

Job experience coded 1ST 2ND 3RD

M SD M SD M SD

GI Novice 3,00 1,00 3,57 0,98 3,00 1,15

GI Experienced 2,64 1,03 3,55 0,82 3,00 0,89

N Novice 2,37 0,48 2,71 1,11 3,73 1,37

N Experienced 2,70 0,87 2,70 1,19 2,75 0,66

S Novice 3,14 1,21 3,29 0,95

S Experienced 2,91 1,14 2,72 0,78

Figure 9.8b and table 9.8b: Mean values for coded knowledge split on job experience

For gathering information the graph hardly shows an effect for job experience; only the training shows some difference, but the t-test confirms the none-effect

[�� = 0.023, t = -0.615, df = 16, p = 0.547]. Scheduling does not show an effect

for job experience either [�� = 0.020, t = 0.577, df = 16, p = 0.572]. Table 9.8cbelow gives an overview of all the t-test results.

Thus, for hypothesis 4, the effect of the job experience of the planners on the relation between their coded knowledge and the implementation of ZKR we conclude


157

Table 9.8c: T-test scores for coded knowledge split on job experience for all threesubtasks and all differences in measurements

Job experience coded �� t df p


Experience 0,048 0,893 16 0,385



Experience 0,000 -0,043 16 0,966

Implementation 0,043 -0,849 16 0,408


Experience 0,123 1,501 16 0,153

Implementation 0,297 2,600 16 0,019**


� Job experience does not [significantly] affect the coded knowledge over all subtasks

� The job experience of planners does [significantly] affect the coded knowledge of the individual subtask negotiating in line with hypothesis 4 – the implementation of ZKR

affects the coded knowledge of novice planners more than the coded knowledge of experience planners; novice planners show a greater increase

� Gathering information shows a knowledge change direction opposite to that of negotiating and scheduling

� Overall, novice planners have more coded knowledge than do experienced planners.

Hypothesis 5 formulates the expected effects of the moderator age on the relationship between innovation and coded knowledge. We formulate this hypothesis as follows.


158

9.4.5 Hypothesis 5

Hypothesis 5

During the implementation process of ZKR younger planners will show more

increase of coded knowledge than older planners.

Figure 9.9a and table 9.9a below show the mean scores and standard deviations for coded knowledge split on age. The graph shows little difference in the coded knowledge of the younger planners compared to the older planners for the

implementation of ZKR [�� = 0.004, t = -0.245, df = 16, p = 0.809]. The knowledge patterns do differ; the younger planners show a linear increasing

pattern, while older planners show a stronger increase for the training [�� = 0.032, t = -0.723, df = 16, p = 0.480] and then they show a decrease for the

experience with ZKR [�� = 0.019, t = 0.556, df = 16, p = 0.586].

Age coded 1ST 2ND 3RD

M SD M SD M SD

Younger 2,97 0,81 3,00 0,84 3,07 1,06

Older 2,59 0,74 3,17 1,02 2,97 0,57

Figure 9.9a and table 9.9a: Mean values and SD’s for coded knowledge over all three measurements split on age.


159

When we go one level deeper [see figure 9.9b and table 9.9b] we see that the patterns of the younger planners show similarities with the patterns of novice planners. And the patterns of the older planners show similarities with the patterns of the experienced planners. This is not surprising, as the job experience has a natural correlation with age.

Gathering information shows a trend for the implementation [�� = 0.116, t = -

1.452, df = 16, p = 0.166], but not for the training [�� = 0.066, t = -1.065, df = 16, p = 0.303].

Age coded 1ST 2ND 3RD

M SD M SD M SD

GI Younger 3,14 1,07 3,57 0,98 3,00 1,15

GI Older 2,55 0,93 3,55 0,82 3,00 0,89

N Younger 2,22 0,40 2,57 1,13 3,30 1,38

N Older 2,79 0,84 2,79 1,17 3,02 0,90

S Younger 2,86 0,90 3,00 1,00

S Older 3,09 1,30 2,90 0,83

Figure 9.9b and table 9.9b: Mean values for coded knowledge split on age


160

Negotiating does not have any effects on age either. However, the trend for the

implementation is in line with hypothesis 5 [�� = 0.123, t = 1.501, df = 16, p = 0.153], in contrast to gathering information. Scheduling shows no effect for age

[�� = 0.020, t = 0.577, df = 16, p = 0.572]. See table 9.9c for an overview of all the t-test results.

Table 9.9c: T-test scores for coded knowledge split on age for all three subtasks and all differences in measurements

�� t df p


Experience 0,019 0,556 16 0,586

Implementation 0,004 -0,245 16 0,809


Experience 0,000 -0,043 16 0,966

Implementation 0,116 -1,452 16 0,166


Experience 0,033 0,739 16 0,471



Thus, for hypothesis 5, the effect of the age of the planners on the relation between their coded knowledge and the implementation of ZKR we conclude

� No [significant] effects for the age of planners on the relationship between their coded knowledge and the implementation of ZKR

� Negotiating and scheduling show change patterns in line with hypothesis 5 – the implementation of ZKR affects the coded knowledge of the younger planners more than the coded knowledge of the older planners; the younger


161

planners show a greater increase. Gathering information shows a trend in contrast to hypothesis 5.

Hypothesis 6 formulates the expected effects of the hours per week that a planner works on the relationship between innovation and coded knowledge. We formulate this last hypothesis as follows.

9.4.6 Hypothesis 6

Hypothesis 6

During the implementation process of ZKR the part time planners will show

less increase in their coded knowledge in comparison to full time planners.

Figure 9.10a and table 9.10a below show the mean scores and standard deviations for coded knowledge split on contractual hours per week.

Working hours coded 1ST 2ND 3RD

M SD M SD M SD

Part time 2,56 0,86 3,40 1,06 2,96 0,65

Full time 2,75 0,68 2,87 0,79 3,05 0,88

Figure 9.10a and table 9.10a: Mean values and SD’s for coded knowledge over all three measurements split on age.


162

The graph shows a small linear increasing pattern for full time planners [= 32 hours or more per week]; the part-time planners show a bigger increase for the training on ZKR, but they show a decrease for the experience with ZKR. Neither

the training trend [�� = 0.129, t = 1.540, df = 16, p = 0.143, nor the knowledge

change for the total implementation of ZKR [�� = 0.004, t = 0.252, df = 16, p = 0.804] are in line with hypothesis 6. The experience trend is in line with hypothesis

6 [�� = 0.106, t = 1.381, df = 16, p = 0.186].

Working hours coded 1ST 2ND 3RD

M SD M SD M SD

GI Part time 2,63 1,06 3,75 0,89 3,00 0,93

GI Full time 2,90 0,99 3,40 0,84 3,00 1,05

N Part time 2,71 0,84 2,96 1,07 3,14 0,99

N Full time 2,46 0,69 2,50 1,18 3,17 1,20

S Part time 3,38 1,41 2,75 0,89

S Full time 2,70 0,82 3,09 0,88

Figure 9.10b and table 9.10b: Mean values for coded split on working hours

At the level of the subtasks the graph shows an increase for all groups for the training on ZKR and a mixing trend of increasing a decreasing coded knowledge


163

for the experience with ZKR; gathering information shows a decrease and negotiating shows an increase, while scheduling is mixed.

For the whole implementation of ZKR as well as for the training on ZKR gathering information shows a knowledge change pattern which contradicts hypothesis 6

[implementation: �� = 0.026, t = 0.649, df = 16, p = 0.526; training: �� = 0.082, t = 1.197, df = 16, p = 0.249], while negotiating shows a change pattern in line

with hypothesis 6 for the whole implementation [�� = 0.010, t = 0.411, df = 16, p

= 0.686] and for the experience on ZKR [�� = 0.010, t = -0.411, df = 16, p = 0.686].

Table 9.10c: T-test scores for coded knowledge split on contractual hours for all three subtasks and all differences in measurements

Contractual hours coded �� t df p


Experience 0,106 -1,381 16 0,186



Experience 0,022 -0,604 16 0,554



Experience 0,028 -0,673 16 0,511

Implementation 0,010 -0,411 16 0,686

Scheduling Experience 0,202 -2,014 16 0,061*

Interestingly, scheduling shows an effect for contractual hours on the experience

with ZKR [�� = 0.202, t = -2.014, df = 16, p = 0.061]; the part time planners show less increase than the full time planners. In fact, the part time planners show a decrease in their coded knowledge compared to an increase of coded knowledge for the full time planners. In addition, all change patterns for the


164

experience of ZKR are in line with hypothesis 6; over all tasks we see a trend, but only scheduling shows an effect of the contractual hours on the coded knowledge of the planners. See table 9.10c for an overview of the t-test results.

Thus, we conclude the testing of hypothesis 6 with the following summation

� The number of contractual hours of the planners did not have an effect for the whole implementation of ZKR over all tasks

� The number of contractual hours showed a [significant] effect on the experience of scheduling in line with hypothesis 6 – while part time planners showed a decrease in their coded knowledge as a result of their experience with ZKR, full time planners showed an increase

� The experience knowledge change patterns are all in line with hypothesis 6; scheduling shows a [significant] effect for the experience with ZKR

We will now conclude this section on the hypotheses testing with some preliminary conclusions.

9.4.7 Preliminary conclusions

� The testing of the hypotheses showed eight significant effects of which seven were in line with our expectations. Four of the significant effects applied to the whole implementation process. The one significant effect that was not in line with our expectations concerned the experience for coded knowledge on gathering information; however, this settled over the whole implementation process.

� The task of negotiating is considerably more in line with our expectations compared to the other two subtasks and the changes over all subtasks [14 changes in line with our expectations versus seven changes not in line]; overall, gathering information and scheduling respectively had the ratio’s 11/10, 10/11 and 4/3. As for the size of the effects the tasks do not differ.

� The whole implementation process shows more effect than either training or experience. Also, the whole implementation process is more in line with our expectations than either the training on ZKR


165

or the experience with ZKR. The implementation shows 13 changes in line with our expectations versus eight expectations not in line compared to 11/10 for training and 15/13 for experience. The whole implementation process also shows more significant effects compared to training and experience [training: 2, both in line with our expectation; experience: 2, one in line and the other not in line with our expectation; implementation: 4, all in line with our expectation].

We preliminary conclude the following. In line with our expectations, coded knowledge [over all subtasks] increased during the implementation of ZKR. Thus, the implementation of ZKR makes the used codes stronger with more consensus. Interestingly, this effect is most prominent for the subtask of negotiating. So, although ZKR does not directly support the activity of negotiating, its use does lead to a structural change in the knowledge of negotiation. That is, the introduction of ZKR leads to the use of stronger concepts for negotiating.

A second remark regarding the increase of coded knowledge is that the impact of ZKR was most prominent after the training [compared to after the experience with ZKR or the implementation process as a whole]. That is, the training on ZKR

showed a greater effect on the use of codes than did the actual use of ZKR

overall. So the increase in code strength might not have established enough to endure actually working with ZKR; the new knowledge needs to be confirmed [trained] repeatedly in order to stabilize the stronger codes.

Theoretical knowledge and sensory knowledge do not seem to be affected significantly by the use of planning support software. The expected conversion of sensory knowledge into coded knowledge was not supported by the data.

In addition to the main effects we also focused on four person-related-characteristics, which we expected to influence the relation between innovation and knowledge: 1] education, 2] job experience, 3] age, and 4] contractual hours. Of these four personal characteristics age was the only characteristic that did not show a single significant effect. That is, we did not find any differences between younger and older planners. We did find education to affect gathering information during the total implementation of ZKR [coded knowledge] and to have an effect after the training [sensory knowledge], job experience affected [the coded knowledge of] negotiating during the implementation of ZKR and the number of contractual hours affected scheduling on the experience with ZKR. Thus, age does not seem to influence the learning process of planners in an implementation


166

process; their education, job experience and contractual hours on the other hand do influence their knowledge types.

The hypothesis testing does not clearly conclude anything concerning the differences between the subtasks. We want to address this issue in the secondary analysis. In particular we want to explore in more detail possible differences between gathering information and negotiating [e.g. negotiating showed more results in line with our expectations than did gathering information]. In addition we also want the secondary analysis to address the differences between the stages of the implementation process, which seemed to result in contradictive effects.

Table 9.11: T-test scores for comparing the subtasks of GI, N, and S – for all three knowledge types for the effects of training, experience and implementation

GI vs. N GI vs. S N vs. S

Sensory Training 0,204

Experience 1,000 0,627 0,605

Implementation 0,336

Coded Training 0,205

Experience 0,082* 0,072* 0,277

Implementation 0,175

Theoretical Training 0,854

Experience 0,920 0,336 0,404

Implementation 1,00

9.5 Secondary analysis

9.5.1 Comparing the subtasks

A t-test comparing the subtasks per knowledge type again showed significant effects for coded knowledge; gathering information significantly differed from


167

negotiating as well as from scheduling for the effect of having experience with ZKR

[see table 9.11]. In other words, actually starting to work with ZKR had a different effect when comparing gathering information and negotiating, but also comparing gathering information and scheduling. Whereas the coded knowledge of negotiating steadily increased, the coded knowledge for gathering information strongly decreased after a strong increase for the training [see also tables and figures 9.12]. All in all coded knowledge did increase for gathering information, although not significantly. So we can conclude differences in the effect of ZKR on the different subtasks. Negotiating is effected more strongly.

9.5.2 Training versus experience

A t-test compared the differences in effect for the three knowledge types with a focus on: 1] the effect of the training with the effect of the experience with ZKR, 2] the effect of the training with the effect of the whole implementation, and 3] the effect of the experience with ZKR with the effect of the whole implementation.

Gathering Information Negotiating

1st 2nd 3rd 1st 2nd 3rd

Sensory 3,50 3,33 3,39 3,00 3,28 3,18

Coded 2,78 3,56 3,00 2,57 2,71 3,13

Theoretical 4,12 4,22 3,94 4,41 4,44 4,20

Figures 9.12 a [left: GI], 9.12b [right: N] and table 9.12: The mean scores for the three knowledge types for the subtasks of gathering information and negotiating


168

Again coded knowledge showed significant effects, but only for gathering information [see also table 9.13]. Firstly, the effect of the training was stronger than the effect of actually working with ZKR. Then, the effect of the training was bigger than that for the whole implementation and thirdly, actually working with ZKR had a different effect compared to the whole implementation process. The most interesting of these three effects is the difference between the effect of the training and the effect of the whole implementation. Hypothesis 1b [significant] effect for the implementation of ZKR for the coded knowledge of gathering information. However, comparing the increase of coded knowledge of the training to the increase of coded knowledge of the whole implementation does reveal a significant difference. Thus, although the whole implementation does not seem to show a significant effect on the coded knowledge of the planners for gathering information there is a difference in effect; this shows the importance of research over longer periods of time, as the whole implementation process of ZKR does seem to have a stronger effect on planners than just the training.

Table 9.13: T-test scores for comparing the effects of training, experience and implementation – for all three knowledge types for the subtasks of GI and N

Trainingvs.

Experience

Trainingvs.

Implementation

Experiencevs.

Implementation

Sensory Gathering information

0,679 0,859 0,528

Negotiating 0,807 0,836 0,512

Coded Gathering information

0,022** 0,066* 0,014**

Negotiating 0,615 0,339 0,795

Theoretical Gathering information

0,541 0,535 0,653

Negotiating 0,765 0,686 1,00


169

9.6 Findings

9.6.1 Increase of coded knowledge

The introduction of a new way of working, in our study the implementation ofthe planning support software of ZKR, structurally affected the knowledge types of the planners who worked with ZKR. More specifically, the coded knowledge significantly increased. That is, the planners of Bartiméus can more easily verbalize their actions after they have started to work with the planning support software ZKR. We understand this structural change in knowledge types in that the need for communication will emerge. This will elicit reflection on ones actions and sets the coding process in motion, which takes the knowledge of the planners to a ‘higher’ level of knowledge, a more coded knowledge.

Theoretical knowledge and sensory knowledge do not seem to be affected significantly by the use of planning support software. Both theoretical knowledgeand sensory knowledge neither have an [significant] effect, nor a trend. We do, however, notice that the theoretical knowledge of the planners shows a consistent change pattern in that it decreases after the planners have had the training on ZKR as well as over the whole implementation process of ZKR. Sensory knowledge does not show a consistent change pattern during the implementation process of the planning support software. The expected conversion of sensory knowledge into coded knowledge was not supported by the data. Thus, working with ZKR only increases the coded knowledge of the planners; sensory knowledge and theoretical knowledge are not affected. Thus, we observed a knowledge change in the sense that planners seem to be more aware of their actions and they can put these actions into words.

9.6.2 Differences between tasks

As for the difference in performed task we first note an interesting difference in dominance of the knowledge types. Gathering information scores higher on sensory knowledge and on coded knowledge, but negotiating scores higher on theoretical knowledge.

As for the differences in knowledge dynamics between the subtasks the most striking difference is the effect of the training for coded knowledge; whereas the verbalization for gathering information increased sharply the verbalization on negotiating was not in the least affected by the training on ZKR [.01 vs. .63]. We


170

do note that over the whole implementation process the verbalization skills on negotiating did increase [significantly], but the verbalization skills on gathering information did not [we might say that they moved towards a trend]. Gathering information seems to be more sensitive to training than is negotiating. But negotiating on the other hand does seem to show some sort of incubation period and overall is affected by the implementation of ZKR.

9.6.3 Personal characteristics

In addition to the main effects we also focused on four person-related-characteristics, which we expected to influence the relation between innovation and knowledge: Of personal education, job experience, age, and contractual hours age was the only characteristic that did not have a single significant effect. Job experience and the education level of the planners influenced their verbal expression. That is, senior educated planners profited more from the implementation of ZKR in that their verbal expression on gathering information showed a greater increase in comparison to the higher educated planners. We do note that the ‘level’ of the coded knowledge of senior educated planners was lower than that of the higher educated planners to begin with. After the implementation on ZKR the coded knowledge of these two groups of planners had reached the same level. As for the influence of job experience, novice planners profited more from the implementation process than experienced planners did.

The verbal skills on negotiating of the novice planners increased significantly more than the verbal skills of the experienced planners. In fact, the novice planners started out inferior to the experienced planners but surpassed them after having worked with ZKR; the experienced planners did not seem to be affected at all by the changes in their work. We do want to note at this point that it could well be that the more experienced planners might have used the planning support software of ZKR in a different way than the novice planners did. From the in-depth interviews we know that some planners did not change their way of working; they continued to make the duty roster by hand and then used ZKR only to register this duty roster. Unfortunately, we do not have this detailed information on all planners, so we cannot conclude that this distinguished the novice planners from the experienced planners.


171

9.6.4 Implementation stages

Then, in studying what actually happens to the knowledge of the planners during the implementation process we came across an interesting difference in effect between the training on ZKR and the experience with ZKR. That is, we implicitly assumed that actually working with ZKR would extend an effect on the knowledge types of the planners initiated by the training that the planners had on ZKR. But this was not what we found. In fact, in some cases [check, formulated differently] the effect established by the training was completely undone by the experience on ZKR. Or maybe we should say that one single training was not enough; the planners could not incorporate the newly obtained knowledge just by using the new planning support software.

9.6.5 In conclusion

In answering the research questions posed in chapter 6 we can say that an organizational innovation such as the implementation of planning support software causes structural changes in the knowledge types of planners. Particularly, the coded knowledge of the planners is affected. Interestingly, the coded knowledge of negotiating is affected the most, while this subtask of planning is not even directly supported by the ZKR panning support software. This is a strong indication that organizational innovation does not only affect the activities that are directly sup-ported by ZKR; it indicates that the impact of ZKR

reaches further, to indirectly related activities. Organizational innovation actually structurally changes the way of working and, more dramatically, the knowledge of people. The personal characteristics of planners were shown to be a factor of consideration and also the stages within the implementation process show differences in knowledge dynamics.

In the final chapter of this thesis we will reflect on these results.


172

173

Chapter 10

Discussion

10.1 Introduction

Knowledge plays an important role in the innovation process, but what actually happens to this knowledge [accounting for its importance] mostly remains unclear. The aim of the present study was to develop a model to study the knowledge dynamics during organizational innovation. A particular focus was to use this model to study knowledge dynamics of individuals in a field setting. We hereto proposed a cognitive-semiotic model, in which three – cumulatively related – types of knowledge were distinguished. We investigated and explored the tangibility of knowledge types and factors that influenced the innovation – knowledge relationship.

This final chapter will reflect on the model presented in chapter 4, the field study using this model and results of our field study. We start with discussing the methodological part of our empirical study on knowledge dynamics [10.2]. Then we discuss the results of our study [10.3] split into four parts, focusing on organizational aspects [10.3.1], planning aspects [10.3.2], knowledge aspects [10.3.3] and innovation aspects [10.3.4] respectively. We conclude this thesis with an evaluation of de cognitive-semiotic model that we presented [10.4].

10.2 Methodological aspects

Regarding the methodological part of this study we want to make three concluding remarks. First of all, our study was the first [as far as we know] to directly measure [types of] knowledge at the individual level [of the planners] instead of using indicators, such as level of education. Secondly, in addition to the first remark, we also made the dynamics of knowledge tangible, as we measured the knowledge types of planners at different points in time during the innovation process. And thirdly, our study provided insights in measuring knowledge and doing [semi] longitudinal research on knowledge dynamics during organizational innovation. These insights can be considered a starting point for new research. The insights gained can be summed up as follows.


174

First of all, most people hardly reflect on the knowledge that they use to perform their jobs. So, asking people about this knowledge stirs things up. They start to think about their tasks and they try to envision situations in which the knowledge is used. This effort is often perceived as difficult and therefore it takes time. Furthermore, the knowledge that people have may be very vivid, but very implicit [at first reflection] as well. Note that this does not necessarily mean that this knowledge is only sensory, as our study shows. So, studying knowledge types has an impact on the subjects participating in the study.

Secondly, our empirical study on knowledge types had no precedent reference. We therefore chose an operationalization strategy, which allowed some latitude; we chose to set up our study broadly and we tried to capture as many related aspects as possible. We formulated a divers set of questions. However, this strategy did not have the outcome we hoped for, namely a set of questions that together formed three constructs representing the three knowledge types of the cognitive semiotic model. As the previous chapter showed, the three groups of questions that were distilled did not include the key-questions that we formulated. Interestingly, though, these key-questions seemed to behave more as outliers than as centre questions within the groups of questions. This insight shows us that these key questions indeed have something unique within the overall set of questions. Further research on the operationalization, preferably a larger population to study, will help us to get a firmer grip on bridging this gap between theory and practice.

Thirdly, in our attempt to link our theoretical framework to an empirical study we operationalized this study as closely as possible to the theoretical concepts. This may not come as a surprise to academics. However, although, this strategy may be technically possible and academically correct many of the planners who participated in our study experienced the questionnaire as very academic and sometimes even farfetched and too analytical; they indicated that the questionnaire wandered in too much detail and that it seemed that we asked the same question more than once. So in a sense, for fear of not correctly operationalizing our theoretical point we dared not to simplify the formulation of our questions in our translation of the theoretical concepts to indicators that can be operationalized. In communication with non-academics the nuance of the selectively formulated questions is lost most of the time, as non-academics or even academics outside the field do not understand the difficultly formulated questions. This leaves us with the choice to either formulate the questions more loosely, a hard thing to do for academics, or to choose a more personal method

Chapter 10 Discussion

175

of data collection [and therefore requiring more labor]: gathering the data in person; being present when the questionnaire is filled out can support the planner through rephrasing questions and giving additional information. Furthermore, the planners provide useful extra information for the interpretation of the data. Of course a combination of the two will do extra.

Finally, in our attempt to capture the knowledge that planners have, we formulated various questions aiming at what planners know and how they apply their knowledge. It turned out that the reliability of the various groups of questions asked was rather low [Cronbach’s alpha < 0.50] and after manipulating the set of questions by putting some new ones in and taking others out it turned out that the most relevant questions in terms of validity should be omitted. We found this not acceptable. This left us with the choice between validity and reliability. We chose for validity. Having to make this choice was a disappointment. Future research should point whether we made the right choice, or not.

10.3 Theory

10.3.1 Organization

In interpreting our data and exploring its possible generalization we want to note the following regarding the data collection at Bartiméus, a large care institution. We particularly want to relate the size of Bartiméus to the increase of coded knowledge that our study showed. Coded knowledge is based on consensus, on agreeing on the interpretation of a code. The greater the group the longer it takes to reach consensus and to develop strong codes. So, this would imply that a bigger organization would take longer to develop strong codes, as the consensus should apply to a greater amount of its members. This could imply that smaller organizations will develop either stronger codes than we have observed at Bartiméus or they will develop their codes faster than we saw at Bartiméus. This is a discussion for further research. We demonstrated that the knowledge within organizations resides in individuals; this knowledge changes, has to be attuned, integrated and shared. Larger organizations make physical co-presence of individuals more difficult and therefore the role of coded knowledge becomes more relevant. However, it comes with a price.


176

10.3.2 Planning

As agreement on the interpretation of coded knowledge, consensus, is a crucial assumption in using coded knowledge, it is vital that coded knowledge is communicated and not just used by the one planner making the duty roster. Communicating coded knowledge will stabilize this knowledge and in this sense it will help to speed up the knowledge dynamics and thus the innovation process. We therefore suggest that planners be stimulated to communicate with each other about making the duty roster. Furthermore, communicating with other planners will ensure rehearsal of coded knowledge and it will improve the use of these codes. In order to stimulate communication among planners they should work together; pair the planners during the training and have them work together on making the duty roster additionally to the training period.

10.3.3 Knowledge

As a result of the introduction of the planning software ZKR we expected the theoretical knowledge to increase. However, the data showed a decreasing tendency, contrary to our expectations. In trying to understand and explain this possible structural trend we go back to the essence of theoretical knowledge, which is to understand concepts in relation to other concepts in its structure and in its coherent pattern. Furthermore, theoretical knowledge is built up of sensory knowledge and coded knowledge. Now, if we take this line of thought then a change in coded knowledge automatically affects theoretical knowledge, in the sense that the theoretical knowledge will be restructured and the concepts will be repositioned. So, an increase of coded knowledge in this sense can lead to a –probably temporal – decrease of theoretical knowledge, as a new structure needs to emerge. In other words, it takes time to adjust and make sense of the new situation. This is important to consider. The training of employees to work with the new planning software should pay considerable attention to the use of coded knowledge. Too great an emphasis on just the ‘logic’ of the software, without embedding it in the day to day routine of making the duty roster, leaves a gap in the knowledge development of the employees. Acquiring coded knowledge, knowing it and understanding it is important. This is particularly relevant to the experienced planners who have strong coded knowledge. Boisot refers to a lock-in effect that can be created once knowledge is coded. These planners will have greater difficulty to adapting their coded knowledge than will novice planners. Therefore, we suggest paying extra attention to this group of planners.


177

Then, we address the subtasks. Although no general [significant] patterns of differences between the subtasks emerged, we did come up with some interesting findings. For instance, the increase of coded knowledge was found over all tasks, but in particular we saw a steady increase for the subtask of negotiating compared to gathering information. We expected the coded knowledge of negotiating to be lower than that of gathering information, which in fact was the case. Then, although the coded knowledge for gathering information increased significantly after the training, over the whole implementation period negotiating showed more increase. It could be that gathering information is more directly a part of the planning process, whereas negotiating is more indirectly involved, in the sense that the ZKR software does not directly support it. In this way the effect of the new way of working will take longer to show for negotiating than it would for the subtask of gathering information. This implies that activities that are not directly supported by the implementation process should not be overlooked, as the innovation process will affect them as well. This is especially important to note, as it seems that these effects will take longer to show. The introduction of software, such as ZKR, to support planning will have effects that were not intended beforehand. The increase of coded knowledge on negotiating is an indicator that planners use more standardized procedures to negotiate about the duty roster. An often-heard problem is that rules and regulations used for the duty roster are not clear, neither for the planners nor for the people who are to be scheduled. The increase of coded knowledge that we measured on the planners might not just have been limited to them; the people who are to be scheduled might also show an increase of coded knowledge. Making the duty roster more transparent to all parties involved will create a greater level of acceptance. This is a crucial factor to ensure the success of the duty roster in the long term.

Regarding the training effect that we observed we want to make the following remark. Learning is not something that just happens [Leonard-Barton 1995]; it is a process that should be stimulated and facilitated. Boisot distinguishes N-learning from S-learning. An important difference between these two types of learning is in the approach to the value of knowledge. Whereas N-learning focuses on the ‘absolute’ accumulation of knowledge, S-learning sees knowledge as more relative. That is, the accumulation of knowledge is not building up to a higher sense of wisdom. Rather, the accumulation of knowledge should always be valued in a larger perspective. From our cognitive-semiotic standpoint this essentially implies, that theoretical knowledge is essential for knowledge dynamics; one needs to


178

have a model to perceive differences, the beginning of the learning process. So either the planners already have this model or such a model should be provided. We relate theoretical knowledge to experience en expertise. Reviewing the data of the novices and experts showed that the novice planners had a greater increase of coded knowledge than did the experienced planners. Interestingly, the theoretical knowledge between the novices and the experts after the implementation had been completed only differed fractionally in favor of the novices. So, although the novices had less coded knowledge before the implementation of ZKR they ended up with almost the same amount of theoretical knowledge as did the expert planners. This could imply that the novice planners are more aware of their own way of planning than are the experts. This idea is supported by the fact that the expert planners have more sensory knowledge before the implementation than the novice planners do. So interestingly, for the overall implementation of support software, [from the decision to start, to finally working with ZKR, almost 1,5 years later] it might well be that the expert planners need more guidance in working with the new software than the novice planners do; they have to learn and practice as well as ‘un-learn’ and un-practice’.

The contractual hours influenced the knowledge dynamics of planners in that planners who worked part-time showed a decrease in their coded knowledge and planners who worked full-time showed an increase. A possible explanation for this difference could be that full-time workers have more routine in scheduling with ZKR. For instance, because they have more time and more opportunities to explore the new planning software. Part-time workers might have a tighter schedule when they work, as they need more time to get into their working rhythm in comparison to full-time workers. So, for the implementation of decision support it should be kept in mind that part time employees can differ in their learning process from full time employees. Extra attention to this point could be translated into a decision to link part timers and full timers during the training to work together to get to know the program.

10.3.4 Innovation

All in all, we started this study to gain more insight into the knowledge dynamics of individuals to better understand the organizational innovation process. We can say we have. Now what are the practical implications for organizations that are on the verge of starting such an innovation process?


179

Innovation, especially introduction of software, is a process, which entails structural changes in the types of knowledge that people use to perform their jobs. These changes should be considered beforehand. An important key to facilitate the innovation process is to have an inventory of the knowledge that people have and use as this knowledge forms the starting point for the innovation process; it is the input. Making an inventory, as part of the innovation process, will have the positive side effect that people will reflect on their knowledge. This reflection will generate more awareness of making the duty roster, which can function as better input to set criteria for the implementation process.

The knowledge is at the same time the input, throughput as well as output of innovation. Therefore, knowledge changes and this will have unintended effects, positive ones as well as negative ones. One effect that we mentioned earlier is the increase of coded knowledge for negotiating. This allows more transparency in the process of making the duty roster. As perceived fairness is an important factor in making the duty roster, the increase of coded knowledge for negotiating might even be considered a must.

The implementation process has practical implications as well. For instance, one planner explained that the new way of working cost much more time. She used to make the duty roster during outside activities. She could then sit outside and make the duty roster while keeping an eye on the clients who were enjoying themselves on the lawn. The implementation of ZKR made this way of working no longer possible, as the making of the duty roster had been restricted to one particular place, behind the computer. An inventory beforehand on how people are used to make the duty roster could have helped in this situation as well. A solution could be to have notebooks available to make the duty roster. Other practical problems include not being able to make a full color print or not to have enough room behind the computer for more than one person to sit. These problems in the practical use of the new software tool could also be overcome by doing a thorough inventory at the beginning of the innovation process. Planners should be allowed to have enough time to gradually switch from the old way of working to the new way.

Innovations, and the implementation of new software that changes working practices, bring along mental load. The way of working changes and this interferes with routines. And of course knowledge of people changes. This will all take up time. So, perceiving an innovation process just in terms of efficiency and earning money will not help. Proper preparation and facilitating the end-users in both


180

content related and practical aspects of the innovation process should be considered more as a must rather than as a luxury that cannot be afforded.

Innovation processes revolve around people. If they do not accept the new way of working and everything that comes along with this then the innovation will most definitely fail. For management to take into account the ways of working and in particular the changes in the knowledge involved will create more involvement of all parties.

Training people on the new way of working is very good. However, just providing training will not do the job. The obtained knowledge will not stick if it is not imbedded in the whole innovation process. Therefore the training should be well timed. Preferably the training should not be the first encounter with the innovation process. Rather, the planners should have already reflected on their way of working before the training, for instance when the inventory is made in preparation of the implementation process. Then, the obtained knowledge should be used and communicated. Have planners practice together in making the duty roster with the new software.

We want to conclude this thesis elaborating on the use of the cognitive semiotic model that we presented.

10.4 Elaboration of thoughts

10.4.1 The I-Space model and the cognitive-semiotic model

A main difference between the I-Space model and the cognitive semiotic model is the different focus of aggregation level. Whereas the I-Space model focuses on the organizational level the cognitive semiotic model focuses on the individual level. We argue that our model therefore provides an extra. Determining the knowledge of an individual implies determining part of the organizational knowledge. So determining the knowledge of many individuals implies the determination of a [big] part of the organization. This can be considered organizational knowledge. So, the step from individual knowledge to organizational knowledge is not a big step. Furthermore, determining the knowledge of individuals gives insight into the diversity of the knowledge available within the organization. Hence, it is very unlikely that all persons within an organization have the same type of knowledge, for instance, very codified and concrete. Rather, these persons all have different knowledge type patterns. For


181

instance, person A has strongly coded and rough sensory knowledge while person B has abstract theoretical knowledge and detailed sensory knowledge. By differentiating between the different persons within an organization will do more justice on the knowledge available within an organization. And more importantly, it will increase the use of the knowledge repository.

10.4.2 Other types of innovation

In exploring the value of our cognitive semiotic model we want to consider a different kind of innovation, the innovation within the organization itself, for instance, product innovation. Organizational innovation, such as the implementation of decision support, involves the input of something new into an organization. Innovation within the organization itself, on the other hand, relies directly on change that is realized by the members of the organization, which is the case in product innovation or service innovation.

Using the cognitive-semiotic model to understand the process of product innovation poses some challenging issues. First of all, we cannot compare the knowledge of the product before it was there to the knowledge of the product after it has been developed. So we need to establish what knowledge we consider in terms of dynamics. Secondly, product innovation involves people from different disciplines. So when we consider their knowledge dynamics, which knowledge do we consider? And thirdly, what kinds of expectations do we have for this divers group of people working on product innovation in terms of knowledge dynamics? Can we understand the process and the activities of product innovation in terms of a general model on knowledge dynamics? So we are mainly looking for a reference point to be able to establish the knowledge dynamics.

Starting with the second issue, the different disciplines, this relates to the diversity of the group of people involved. The people that work on product innovation all have their expertise, which they use as input for the development of new products. In understanding the knowledge dynamics of the different experts we need to focus on the expert knowledge in relation to the new product, not their expert knowledge in general. Furthermore, we make the assumption that the process of product innovation does have a focus and direction. So when we want to study the knowledge dynamics of the experts involved we use the focus and direction of the new product as a reference point. For instance, the expert working on new ways to make coffee has the direction and focus of developing a


182

coffee machine or at least think about the ways to enjoy drinking coffee, either at home or in a public place. The knowledge [dynamics] that we consider in this process involves conceptual change of the new product. We could hypothesize that the knowledge type at the beginning of the process is primarily sensory, and gradually becomes coded. In communication with other experts the knowledge will become of a more theoretical nature. This line of arguing shows great parallels with the Nonaka and Takeuchi [1995] story on the development of new products, such as cars and bread making machines. The difference between their approach to knowledge change and our cognitive semiotic approach is that we explicitly add a theoretical dimension in the process of knowledge dynamics. This dimension is often implicitly meant in the more common term ‘tacit knowledge’ [see chapter 4].

Furthermore, we can expect that the role of theoretical knowledge is considerable. The experts are involved, as they understand their discipline in a broad sense. They are expected to be able to link their expertise to many different ideas and concepts and they are expected to be able to consider different points of view, which is an essential part of theoretical knowledge. We argue that theoretical knowledge is an essential part of product innovation and creativity in general as it provides powerful tools to perceive new things; and perception of change is what sets the process of knowledge dynamics in motion.

183

References

A

Ackerman, P. L., Sternberg, R. J., & Glaser, R. (Eds.). (1989). Learning and Individual Differences: Advances in theory and research. New York: W.H. Freeman and Company.

Anderson, J. R. (1990). Cognitive Psychology and Its Implications. New York: W.H. Freeman and Company.

Anthony, R. N. (1965). Planning and Control Systems: A Framework for Analysis. Boston: Harvard.

Antonacopoulou, E. P. (2001). The Paradoxical Nature of the Relationship Between Training and Learning. Journal of Management Studies, 38(3), May: 327-350.

Argyris, C., & Schön, D. A. (1978). Organizational Learning: A theory of action perspective. Reading: Addison-Wesley.

Argyris, C. (1999). On Organizational Learning (2nd ed.): Blackwell Publishing.

B

Blackler, F. (1995). Knowledge, Knowledge Work and Organizations: An overview and interpretation. Organization Studies, 16(6), 1021-1046.

Boisot, M. (1983). Convergence Revisited: The codification and diffusion of knowledge in a British and a Japanese firm. Journal of Management Studies, 20(2), 159–190.

Boisot, M. (1995). Information Space. London: Routledge.

Boisot, M. (1998). Knowledge Assets. Securing competitive advantage in the information economy. Oxford: Oxford University Press.

Boisot, M., & Canals, A. (2004). Data, Information and Knowledge: Have we got it right? Journal of Evolutionary Economics, 14(1), 43-67.

Borghini, S. (2005). Organizational Creativity: Breaking equilibrium and order to innovate. Journal of Knowledge Management, 9(4), 19-33.

Bosch-Sijtsema, P. (2001). Knowledge Development in a Virtual Organisation: An information processing perspective. Lund University, Lund.

Bostock, D. (1999). Plato. In T. Honderich (Ed.), The Philosophers. Introducing great Western thinkers (pp. 15-21). Oxford: Oxford University Press.


184

Bouissac, P. (1998). Converging Parallels: Semiotics and psychology in evolutionary perspective. Theory & Psychology, 8(6), 731-754.

Burnside, R. M. (1990). Improving corporate climates for creativity. In M. A. West & J. L. Farr (Eds.), Innovation and Creativity at Work (pp. 265-284). West Sussex: John Wiley & Sons Ltd.

C

Carmines, E. G., & Zeller, R. A. (1994). Reliability and Validity Assessment. In M. S. Lewis-Beck (Ed.), Basic Measurement (pp. 1-58). London: Sage Publications Toppan Publishing.

Carroll, J. D., & Henry, N. (1975). Introduction [to “a symposium, knowledge management”]. Public Administration Review, 35, 567.

Clement, J. (1983). A Conceptual Model Discussed by Galileo and Used Intuitively by Physics Students. In D. Gentner & A. L. Stevens (Eds.), Mental Models (pp. 325-340). Hillsdale, NJ: Lawrence Erlbaum Associates.

Cohen, W. M., & Levinthal, D. A. (1990). Absorptive Capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35, 128-152.

Coriat, B., & Dosi, G. (1998). Learning how to Govern and Learning how to Solve Problems: On the co-evolution of competences, conflicts and organizational routines. In A. Dupont, P. Chandler & Ö. S. Hagström (Eds.), The Dynamic Firm: The role of technology, strategy, organization, and regions (pp. 103-133). New York: Oxford University Press.

Cunningham, D. J. (1998). Cognition as Semiosis: The Role of Inference. Theory & Psychology, 8(Special Issue: Semiotics and Psychology - 6), 827.

D

Damanpour, F. (1987). The Adoption of Technological, Administrative, and Ancillary Innovations: Impact of organizational factors. Journal of Management, 13(4), 675-688.

Damanpour, F. (1990). Innovation Effectiveness, Adoption and Organizational Performance. In Innovation and Creativity at Work (pp. 125-141). West Sussex: John Wiley & Sons Ltd.

Damanpour, F., & Evan, W. M. (1984). Organizational Innovation and Performance: The problem of "organizational lag". Administrative Science Quarterly, 29(3), 392-409.

Davenport, T. H., & Prusak, L. (2000). Working Knowledge: How organizations manage what they know: Harvard Business School Press.

Dawson, S. (1995). Organisaties Analyseren. Schoonhoven: Academic Service.

References

185

de Groot, A. D. (1961). Methodologie: Grondslagen van onderzoek en denken in de gedragswetenschappen. 's-Gravenhage: Mouton.

Dennett, D. C. (1971). Intentional Systems. The Journal of Philosophy, 68(4 [Feb. 25]), 87-106.

Dennett, D. C. (1978). Brainstorms. Brighton, Sussex: Het Harvester Press.

du Plessis, M. (2007). The Role of Knowledge Management in Innovation. Journal of Knowledge Management, 11(4), 20-29.

Duh, R.-R., Chow, C. W., & Chen, H. (2006). Strategy, IT Applications For Planning and Control, and Firm Performance: The impact of impediments to IT implementation. Information & Management, 43, 939-949.

E

Eberhardt, B. J., & Shani, A. B. (1984). The Effects of Full-time Versus Part-time Employment Status on Attitudes Toward Specific Organizational Characteristics and Overall Job Satisfaction. Academy of Management Journal, 27(4), 893-900.

Edwards, P. (1972). The Encyclopedia of Philosophy. New York: Macmillan.

F

Farr, J. L., & Ford, C. M. (1990). Individual Innovation. In Innovation and Creativity at Work. West Sussex: John Wiley & Sons Ltd.

Feldman, D. C. (1990). Reconceptualizing the Nature and Consequences of Part-time Work. Academy of Management Review, 15(1), 103-112.

Fichman, R. G., & Moses, S. A. (1999). An Incremental Process for Software Implementation. Sloan Management Review, 40(2), 39-52.

G

Garcia, R., & Calantone, R. (2002). A Critical Look at Technological Innovation Typology and Innovativeness Terminology: A literature review. The Journal of Product Innovation Management, 19(2), 110-132.

Gazendam, H. W. M., & Jorna, R. J. (1998). Theories About Architecture and Performance of Multi-Agent Systems [Electronic Version]. University of Groningen: SOM Research Report, 98, 1-33. Retrieved July 28, 2008.

Geerts, H. (1999). Standaard Softwarepakketten Selecteren en Implementeren.Amersfoort: Twynstra Gudde.

George, J. M., & Jones, G. R. (2001). Towards a Process Model of Individual Change in Organizations. Human Relations, 54(4), 419-444.


186

Glynn, M. A. (1996). Innovative Genius: A framework for relating individual and organizational intelligences to innovation. The Academy of Management Review, 21(4), 1081-1111.

Goerl, G. F. (1975). Cybernetics, professionalizations, and knowledge management: an exercise in assumptive theory. Public Administration Review, 35(6), 581.

Golish, B. L., Besterfield-Sacre, M. E., & Shuman, L. J. (2008). Comparing Academic and Corporate Technology Development Processes*. Journal of Product Innovation Management, 25(1), 47-62.

Goodman, N. (1981). Languages of Art (2nd ed.). Brighton, Sussex: The Harvester Press.

Gopalakrishnan, S., & Damanpour, F. (1997). A Review of Innovation Research in Economics, Sociology and Technology Management. Omega, 25(1), 15-28.

Gourlay, S. (2006). Towards Conceptual Clarity for ‘Tacit Knowledge’: A review of empirical studies. Knowledge Management Research & Practice, 4(1), 60-69.

Grant, R. M. (1996). Toward a Knowledge-Based Theory of the Firm. Strategic Management Journal, 17(Winter Special Issue: Knowledge and the Firm), 109-122.

Greif, S., & Keller, H. (1990). Innovation and the Design of Work and Learning Environments: The concept of exploration in human-computer interaction. In Innovation and Creativity at Work (pp. 231-249). West Sussex: John Wiley & Sons Ltd.

H

Haberlandt, K. (1994). Cognitive Psychology. Needham Heights, Massachusetts 02194: Allyn and Bacon.

Hage, J. (1980). Theories of Organizations: Form, Process, and Transformation. New York: John Wiley & Sons.

Hage, J., & Aiken, M. (1970). Social Change in Complex Organizations. New York: Random House.

Hage, J., & Dewar, R. (1973). Elite Values Versus Organizational Structures in Predicting Innovation. Administrative Science Quarterly, 18(3), 279-290.

Hage, J. T. (1999). Organizational Innovation and Organizational Change. Annual Review of Sociology, 25(1), 597-622.

Hannan, M. T., & Freeman, J. (1984). Structural Inertia and Organizational Change. American Sociological Review, 49(2), April: 149-164.

References

187

Hayes-Roth, B., & Hayes-Roth, F. (1979). A Cognitive Model of Planning. Cognitive Science, 3, 275-310.

Hedlund, G. (1994). A Model of Knowledge Management and The N-Form Corporation. Strategic Management Journal, 15(Summer Special Issue: Strategy: Search for New Paradigms), 73-90.

Heijloo, R., & Eskens, E. (Eds.). (2005). Filosofen Lexicon. Leven, werk en citaten van tweehonderd grote denkers. Diemen: Veen Magazines.

Helmhout, M. (2006). The Social Cognitive Actor. A multi-actor simulation of organisations. University of Groningen, Groningen.

Henry, N. L. (1974). Knowledge management: a new concern for public administration. Public Administration Review, 34(3), 189-196.

Heywood, S. (1965). Toward a Sound Theory of Innovation. The Elementary School Journal, 66(3), 107-114.

Hilgard, E. R., & Bower, G. H. (1975). Theories of Learning. Englewood Cliffs, New Jersey: Prentice-Hall, Inc.

Horn, J., L. (1989). Cognitive Diversity: A framework of learning. In P. L. Ackerman, R. J. Sternberg & R. Glaser (Eds.), Learning and Individual Differences. Advances in theory and research (pp. 61-116). New York: W.H. Freeman and Company.

Hunt, E. (1999). Intelligence and Human Resources: Past, Present, and Future. In P. L. Ackerman, P. C. Kyllonen & R. D. Roberts (Eds.), Learning and Individual Differences (pp. 482). Washington, DC 20002: American Psychological Association.

Huysman, M. (1996). Dynamics of Organizational Learning. Vrije Universiteit, Amsterdam.

J

Jackofsky, E. F., & Peters, L. H. (1987). Part-time Versus Full-time Employment Status Differences: A replication and extension. Journal of Occupational Behaviour, 8(1), 1-9.

Johnson-Laird, P. N. (1993). Mental Models. In M. I. Posner (Ed.), Foundations of Cognitive Science (pp. 469-499). Cambridge, Massachusetts: MIT Press.

Johnson-Laird, P. N. (2006). How We Reason. Oxford: Oxford University Press.

Jorna, R. (Ed.). (2006). Sustainable Innovation. the organisational, human and knowledge dimension. Sheffield: Greenleaf Publishing Ltd.

Jorna, R. J. (1990). Kennisrepresentaties en Symbolen van de Geest. Amsterdam: Uitgeverij Thesis.


188

Jorna, R. J., Gazendam, H. W. M., Heesen, H. C., & Van Wezel, W. M. C. (1996). Plannen en Roosteren. Taakgericht analyseren, ontwerpen en ondersteunen. Leiderdorp: Lansa Publishing B.V.

Jorna, R. J., & Waalkens, J. (2006). Innovation: Many-headed and certainly important. In R. J. Jorna (Ed.), Sustainable Innovation. The organisational, human and knowledge dimension (pp. 15-27). Sheffield: Greenleaf Publishing Ltd.

Judd, C. M., Smith, E. R., & Kidder, L. H. (1991). Research Methods in Social Relations (6th ed.). Fort Worth, TX: Holt Rinehart and Winston.

K

Khilji, S. E., Mroczkowski, T., & Bernstein, B. (2006). From Invention to Innovation: Toward developing an integrated innovation model for biotech firms*. Journal of Product Innovation Management, 23(6), 528-540.

Kiewiet, D. J., Jorna, R. J., & van Wezel, W. (2005). Planners and their cognitive maps: An analysis of domain representations using multi dimensional scaling. Applied Ergonomics, 36(6), 695-708.

Kimberly, J. R., & Evanisko, M. J. (1981). Organizational Innovation: The influence of individual, organizational, and contextual factors on hospital adoption of technological and administrative innovations. Academy of Management Journal, 24(4), 689-713.

King, N. (1990). Innovation at Work: The research literature. In M. A. West & J. L. Farr (Eds.), Innovation and Creativity at work (pp. 15-59). West Sussex: John Wiley & Sons Ltd.

Klaus, H., Rosemann, M., & Gable, G. G. (2000). What is ERP? Information Systems Frontiers, 2(2), 141-162.

Klein, K. J., & Sorra, J. S. (1996). The Challenge of Innovation Implementation. The Academy of Management Review, 21(4), 1055-1080.

Knoblock, C. A. (1995). Planning, Executing, Sensing, and Replanning for Information Gathering. Paper presented at the Fourteenth International Joint Conference on Artificial Intelligence, Montreal, Canada.

Knudsen, M. P. (2007). The Relative Importance of Interfirm Relationships and Knowledge Transfer for New Product Development Success*. Journal of Product Innovation Management, 24(2), 117-138.

Kræmmerand, P., Møller, C., & Boer, H. (2003). ERP Implementation: An integrated process of radical change and continuous learning. Production Planning & Control, 14(4), 338-348.

References

189

Kratzer, J., Leenders, R. T. A. J., & Engelen, J. M. L. (2004). Stimulating the Potential: Creative Performance and Communication in Innovation Teams. Creativity and Innovation Management, 13(1), 63-71.

L

Lam, A. (1998). Tacit Knowledge, Organisational Learning and Innovation: A societal perspective. Paper presented at the Danish Research Unit for Industrial Dynamics [DRUID].

Lamain, W. (2000). Cognitive Task Analysis in Human-computer Interaction: Goals and information processing profiles. University of Groningen, Groningen.

Lammers, I. (2003). In Conflict, een Geschiedenis van Kennismanagement. Vrije Universiteit, Amsterdam.

Larkin, J. H. (1983). The Role of Problem Representation in Physics. In D. Gentner & A. L. Stevens (Eds.), Mental Models (pp. 75-98). Hillsdale, NJ: Lawrence Erlbaum Associates.

Laudon, & Laudon. (2002?). System Success and Failure: Implementation. In Management Information Systems: Organization and Technology in the Networked Enterprise (pp. 398-431).

Le Breton, P. P., & Henning, D. A. (1961). Planning Theory. Englewood Cliffs, NJ: Prentice-Hall.

Legge, J. (1992). The Competitive Edge: How Innovation Creates and Sustains the Competitive Advantage of Enterprises. North Sydney: Allen & Unwin.

Legge, K. (1984). Evaluating Planned Organizational Change. London: Academic Press.

Leiponen, A. (2006). Managing Knowledge for Innovation: The Case of Business-to-Business Services*. Journal of Product Innovation Management, 23(3), 238-258.

Leonard-Barton, D. (1985). Experts as Negative Opinion Leaders in the Diffusion of a Technical Innovation. Journal of Consumer Research, 11(4), 914-926.

Leonard-Barton, D. (1988). Implementation Characteristics of Organizational Innovations: Limits and opportunities for management strategies. Communication Research, 15(October), 603-631.

Leonard-Barton, D. (1995). Wellsprings of Knowledge. Building and sustaining the sources of innovation. Boston, Massachusetts: Harvard Business School Press.

Levanoni, E., & Sales, C. A. (1990). Differences in Job Attitudes Between Full-time and Part-time Canadian Employees. Journal of Social Psychology, 130(2), 231-237.


190

Lewicki, R. J., Saunders, D. M., & Minton, J. W. (1997). Essentials of Negotiating. Boston: The McGraw-Hill Companies, Inc.

Lewicki, R. J., Saunders, D. M., & Minton, J. W. (Eds.). (1998). Negotiating. Readings, exercises, and cases (3rd ed.). Boston: The McGraw-Hill Companies, Inc.

Linton, J. D. (2002). Implementation Research: State of the art and future directions. Technovation, 22(2), 65-79.

Löwstedt, J. (1993). Organizing Frameworks in Emerging Organizations: A Cognitive Approach to the Analysis of Change. Human Relations, 46(4), 501-526.

M

Maguire, C., Kazlauskas, E. J., & Weir, A. D. (1994). Information services for innovative organizations. San Diego: Academic Press.

Majchrzak, A. (1988). The human side of factory automation. San Francisco, CA: Jossey-Bass Inc., Publishers.

Majchrzak, A., Cooper, L. P., & Neece, O. E. (2004). Knowledge Reuse for Innovation. Management Science, 50(2), February : 174-188.

March, J. G., & Simon, H. A. (1958). Organizations. New York: John Wiley.

McCloskey, M. (1983). Naive Theories of Motion. In D. Gentner & A. L. Stevens (Eds.), Mental Models (pp. 299-324). Hillsdale, NJ: Lawrence Erlbaum Associates.

McCraw, T. K. (2007). Prophet of Innovation: Joseph Schumpeter and creative destruction. Cambridge, Massachusetts: The Belknap Press of Harvard University Press.

Michon, J. A., Jackson, J. L., & Jorna, R. J. (2003). Semiotic Aspects of Psychology: Psychosemiotics. In R. Posner & K. Robering (Eds.), Semiotik: Ein Handbuch Zu Den Zeichentheoretischen Grundlagen Von Natur und Kultur (A Handbook on the Sign-Theoretic Foundations of Nature and Culture). Berlin: Walter de Gruyter.

Mietus, D. M. (1994). Understanding Planning for Effective Decision Support. A cognitive task analysis of nurse scheduling. University of Groningen, Groningen.

Morley, I., & Stephenson, G. (1977). The Social Psychology of Bargaining. London: George Allen & Unwin Ltd.

References

191

N

Nakamura, A., & Nakamura, M. (1983). Part-time and Full-time Work Behaviour of Married Women: A Model with a Doubly Truncated Dependent Variable. Canadian Journal of Economics, 16(2), 229-257.

Nelson, R. R., & Winter, S. G. (1982). An Evolutionary Theory of Economic Change. Cambridge: The Belknap Press of Harvard University Press.

Nersessian, N. J. (1992). How Do Scientists Think? Capturing the dynamics of conceptual change in science. In R. N. Giere (Ed.), Cognitive Models of Science (pp. 3-45). Minneapolis, MN: University of Minnesota Press.

Newell, A., & Simon, H. A. (1972). Human Problem Solving. Englewood Cliffs: Prentice-Hall.

Nicholson, N. (1990). Organizational Innovation in Context: Culture, interpretation and application. In M. A. West & J. L. Farr (Eds.), Innovation and Creativity at Work (pp. 179-204). West Sussex: John Wiley & Sons Ltd.

Nonaka, I., & Takeuchi, H. (1995). The Knowledge-Creating Company. How Japanese companies create the dynamics of innovation. New York: Oxford University Press.

O

O'Connor, G. C. (1998). Market Learning and Radical Innovation: A cross case comparison of eight radical innovation projects. Journal of Product Innovation Management, 15, 151-166.

O'Mahony, M., & van Ark, B. (2003). EU Productivity and Competitiveness: An industry perspective: Can Europe resume the catching-up process? Retrieved April 19, 2008, from http://www.ggdc.net/pub/EU_productivity_and_competitiveness_exec.pdf.

P

Patalano, A. L., & Seifert, C. M. (1997). Opportunistic Planning: Being Reminded of Pending Goals. Cognitive Psychology, 34(1), 1-36.

Pavia, T. M. (1991). The Early Stages of New Product Development in Entrepreneurial High-Tech Firms. Journal of Product Innovation Management, 8(1), 18-31.

Polanyi, M. (1964). Personal Knowledge: Towards a post-critical philosophy. New York, NY: Harper & Row, Publishers, Incorporated.

Polanyi, M. (1966). The Tacit Dimension. London: Routledge & Kegan Paul.


192

Posner, M. I. (Ed.). (1993). Foundations of Cognitive Science. Cambridge, Massachusetts: The MIT Press.

Prusak, L. (Ed.). (1997). Knowledge in Organizations. Boston: Butterworth-Heinemann.

Pylyshyn, Z. W. (1993). Computing in Cognitive Science. In M. I. Posner (Ed.), Foundations of Cognitive Science (paperback ed., pp. 51-91). Cambridge, Massachusetts: The MIT Press.

R

Rogers, E. M. (1995). Diffusion of Innovations. New York: The Free Press.

Roth, E. M., & Woods, D. D. (1989). Cognitive Task Analysis: An approach to knowledge acquisition for intelligent system design. In G. Guida & C. Tasso (Eds.), Topics in Expert System Design. Amsterdam: Elsevier Science Publishers B.V.

S

Sapolsky, H. M. (1967). Organizational Structure and Innovation. The Journal of Business, 40(4 (October)), 497-510.

Satchell, P. M. (1998). Innovation and Automation. Aldershot, England: Ashgate Publishing Ltd.

Schuyt, C. J. M. (1997). Teken en Betekenis in het Sociale Leven. In J. C. A. Van der Lubbe & J. A. Van Zoest (Eds.), Teken en betekenis (pp. 156-172). Haarlem: Aramith Uitgevers.

Scott, S. G., & Bruce, R. A. (1994). Determinants of Innovative Behavior: A path model of individual innovation in the workplace. The Academy of Management Journal, 37(3), 580-607.

Senge, P. M. (1990). The Fifth Discipline: The art and practice of the learning organization. New York: Doubleday.

Senker, J. (1995). Tacit Knowledge and Models of Innovation. Industrial and Corporate Change, 4(2), 425-447.

Shank, G. (1998). The Extraordinary Ordinary Powers of Abductive Reasoning. Theory & Psychology, 8(6), 841-860.

Simon, H. A. (1947). Administrative Behavior. New York: Macmillan

Simon, H. A., & Kaplan, C. A. (1993). Foundations of Cognitive Science. In M. I. Posner (Ed.), Foundations of Cognitive Science (pp. 1-47). Cambridge, Massachusetts: The MIT Press.

References

193

Slappendel, C. (1996). Perspectives on Innovation in Organizations. Organization Studies, 17(1), 107-129.

Smith, H. A. (2000). Psychosemiotics and its Peircean Foundations [Electronic Version], 24. Retrieved January 5th, 2008 from http://www.digitalpeirce.fee.unicamp.br/PsychosemioticsandPeirce.pdf.

Smythe, W. E., & Chow, S. L. (1998). The Formal and Mental Structures of Semiotic Processes. Theory & Psychology, 8(6), 783-804.

Smythe, W. E., & Jorna, R. J. (1998). The Signs We Live by: The relationship between semiotics and psychology. Theory & Psychology, 8(6), 723-730.

Song, X. M., Thieme, R. J., & Xie, J. (1998). The Impact of Cross-Functional Joint Involvement Across Product Development Stages: An exploratory study. Journal of Product Innovation Management, 15(4), 289-303.

Sorge, A., & Warner, M. (Eds.). (1997). The IEBM Handbook of Organizational Behavior. London: International Thomson Business Press.

Sørensen, C., & Lundh-Snis, U. (2001). Innovation Through Knowledge Codification. Journal of Information Technology, 16(2), 83-97(15).

Steffy, B. D., & Jones, J. W. (1990). Differences Between Full-Time and Part-time Employees in Perceived Role Strain and Work Satisfaction. Journal of Organizational Behavior, 11(4), 321-329.

Strambach, S. (2002). Change in the Innovation Process: New knowledge production and competitive cities - the case of Stuttgart. European Planning Studies, 10(2), 215-231.

Subramanian, A., & Nilakanta, S. (1996). Organizational innovativeness: Exploring the relationship between organizational determinants of innovation, types of innovations, and measures of organizational performance. Omega, 24(6), 631-647.

Sundbo, J. (1998). The Theory of Innovation. Cheltenham, UK: Edward Elgar.

T

Tidd, J., Bessant, J. R., & Pavitt, K. (1997). Managing Innovation: Integrating technological, market and organizational change. Chichester, West Sussex: Wiley.

Tushman, M. L., & Anderson, P. (1986). Technological Discontinuities and Organizational Environments. Administrative Science Quarterly, 31, 439-465.


194

V

van Ark, B., & Pilat, D. (1993). Productivity Levels in Germany, Japan, and the United States: Differences and Causes. In M. N. R. Baily, Peter C.; Winston, Clifford (Ed.), Brookings Papers on Economic Activity: Macroeconomics (pp. 1-48). Washington, D.C.: Brookings Institution Press.

van Dam, R. E. A. (2008). Personal communication.

van de Ven, A. H. (1986). Central Problems in the Management of Innovation. Management Science, 32(5), May : 590-607.

van de Ven, A. H., & Polley, D. (1992). Learning While Innovating. Organization Science, 3(1), 92-116.

van de Ven, A. H., Polley, D. E., Garud, R., & Venkataraman, S. (1999). The Innovation Journey. Oxford: Oxford University Press.

van Heusden, B. (1994). Why Literature? An inquiry into the nature of literary semiosis. University of Groningen, Groningen.

van Heusden, B., & Jorna, R. J. (2001). Toward a Semiotic Theory of Cognitive Dynamics in Organizations. In K. Liu, R. J. Clarke, P. B. Andersen & R. K. Stamper (Eds.), Information, Organisation and Technology. Studies in Organisational Semiotics (pp. 83–113). Boston: Kluwer Academic Publishers.

van Stijn, E. J. (2006). Miracle or Mirage? : An exploration of the pervasive erp system phenomenon informed by the notion of conflicting memories. University of Twente, Enschede.

van Wezel, W., Jorna, R. J., & Mietus, D. (1996). Scheduling in a Generic Perspective. Knowledge -Based Decision Support by Domain Analysis and Cognitive Task Analysis. International Journal of Expert Systems, 9(3), 357-381.

Verjans, S. (1999?). Organisational and Psychological Analysis of the Implementation of Information Technology in Production Settings: Sense-Making and Stress at Two Levels of Analysis. Paper presented at the ICEIS.

Voss, C. A. (1988). Implementation: A Key Issue in Manufacturing Technology: The need for a field of study. Research Policy, 17(2), 55-63.

W

Waalkens, J. (2006). Building Capabilities in the Construction Sector. Absorptive capacity of architectural and engineering medium-sized enterprises.University of Groningen, Groningen.

Wærn, Y. (1989). Cognitive Aspects of Computer Supported Tasks. New York, NY, USA: John Wiley & Sons, Inc.

References

195

Weick, K. E., & Westley, F. (1996). Organizational Learning: Affirming an Oxymoron. In Handbook of Organization Studies (pp. 440-458).

West, M. A., & Farr, J. L. (1990a). Innovation at Work. In M. A. West & J. L. Farr (Eds.), Innovation and Creativity at Work (pp. 3-14). West Sussex: John Wiley & Sons Ltd.

West, M. A., & Farr, J. L. (Eds.). (1990b). Innovation and Creativity at Work. West Sussex: John Wiley & Sons Ltd.

Wijnen, G., & Kor, R. (1996). Het Managen van Unieke Opgaven. Samen werken aan projecten en programma's: Kluwer Bedrijfswetenschappen.

Wolfe, R. A. (1994). Organizational innovation: review, critique and suggested research directions. Journal of Management Studies, 31(3), 405-431.

Woll, S. (2002). Everyday Thinking. Memory, reasoning, and judgment in the real world. New Jersey: Lawrence Erlbaum Associates, Publishers.

Z

Zack, M. H. (1999). Managing Codified Knowledge. Sloan Management Review, 40(4), 45-58.

Zaltman, G., Duncan, R., & Holbek, J. (1973). Innovations and Organizations. New York: John Wiley & Sons Inc.


196

197

Appendix: Questionnaire

Part A: General questions

A1 Datum:

A2 Naam:

A3 Leeftijd: jaar

A4 Hoogst genoten (afgeronde) opleiding (indeling volgens CBS 1992):

O Elementair (vgl. Basisschool)

O Lager, LBO

O Middelbaar, MBO

O Hoger, HBO

O Wetenschappelijk, WO

O Anders, nl. ……………….

A5

Functie:

O Afdelingshoofd

O Hoofdverpleegkundige

O Verpleegkundige

O Leerling verpleegkundige

O Anders, nl. ……………….

A6 Op welke afdeling bent u werkzaam?

A7.1 Aanvang aanstelling van uw huidige functie:

Maand: ……………………….Jaar: ………………………….

A7.2 Omvang aanstelling (in uren per week):

Uren per week

A8.1

Ervaring in soortgelijke functies:

maand(en) en ja(a)r(en) of N.V.T.

A8.2 Andere werkervaring:

maand(en) en ja(a)r(en) of N.V.T.

A8.3 Eventuele toelichting op soort andere werkervaring:


198

A9.1 Hebt u ervaring met het werken met computers?

Ja / Nee

A9.2Zo ja, met wat voor pakketten hebt u gewerkt?• Tekstverwerkerprogramma’s, zoals WP en Word• Spreadsheetprogramma’s, zoals Excel• Databases, zoals Access• Analyseprogramma’s, zoals ….• Anders, namelijk …………

A10Hoe hebt u tot nog toe uw dienstroosters gemaakt?• Handmatig• Met behulp van een spreadsheet• Met behulp van een speciaal planprogramma• Anders, namellijk …………….

A11.1 E-mail: ………………………...@………………..……………….

A11.2 Telefoonnummer: ( )

A12.1 Is de kwaliteit van de personeelsroosters voor verbetering vatbaar wat betreft:

benutting capaciteit? Ja / Nee

garanderen continuïteit? Ja / Nee

tevreden stellen van verpleegkundigen? Ja / Nee

A12.2 Welke bijdrage verwacht u dat ZKR kan leveren aan het verbeteren van de kwaliteit?

A12.1 Welke problemen komt u nu tegen bij het roosteren?

A13.2 Hoe verwacht u dat ZKR u daarbij kan helpen bij deze problemen?

A14.1 Is de organisatie Bartiméushage het laatste jaar veranderd?

Niet veranderd Veel veranderd

1 2 3 4 5

A14.2 Is de organisatie van uw afdeling het laatste jaar veranderd*?

Niet veranderd Veel veranderd

1 2 3 4 5

A14.3 Op welke manier is deze organisatieverandering van invloed op het maken van een

rooster?

A15.1 Welk doel streeft Bartiméushage na?

A15.2 Hoe wordt dit doel gerealiseerd?

A16.1 Welk doel streeft de afdeling waar u werkt na?

A16.2 Hoe wordt dit doel gerealiseerd?

A14 Hieronder vindt u een vijftal omschrijvingen. Kunt u van elke omschrijving aangeven in

hoeverre deze past bij de organisatie van Bartiméushage (1) en in hoeverre deze past bij

de organisatie van de planning van Bartiméushage (2)?

Appendix

199

A14.1

1. BTH

2. plann

Omschrijving A:

Kennis is weinig vastgelegd (vuistregels, gewoonte, praktijkervaring, gevoel), concreet,

meest moeilijk te verwoorden en gedetailleerd, en alleen beschikbaar voor leden van de

(kleine) groep (ínsiders’). De kennis is vrij gelijk verdeeld, er is nauwelijks hiërarchie,

hoogstens iemand die een voorbeeldfunctie heeft. Samenwerking staat hoog in het vaandel

geschreven.

Weinig overeenkomst Veel overeenkomst

1 2 3 4 5

1 2 3 4 5

A14.2

1. BTH

2. plann

Omschrijving B:

Kennis is nauwelijks vastgelegd, vrij concreet en weinig gedetailleerd, meestal moeilijk te

verwoorden, en alleen beschikbaar voor een enkeling. Veel leden van de organisatie weten

niet wat ze doen en waarom. De kennis is geconcentreerd bij enkelen.


1 2 3 4 5

1 2 3 4 5

A14.3

1. BTH

2. plann

Omschrijving C:

Kennis is veelal vastgelegd, kan niet door iedereen ‘beredeneerd’ worden en is ook niet

voor iedereen toegankelijk. Kennis is macht, in dit type organisatie. De organisatie is sterk

hiërarchisch, kennis wordt niet gedeeld.


1 2 3 4 5

1 2 3 4 5

A14.4

1. BTH

2. plann

Omschrijving D:

Kennis is heel sterk vastgelegd (bijvoorbeeld in de vorm van regels, handboeken, jargon of

vaktaal), is theoretisch (veronderstelt een zekere specialisatie) maar is voor iedereen

beschikbaar en toegankelijk. Er is een taakverdeling op grond van competenties.


1 2 3 4 5

1 2 3 4 5

A14.5

1. BTH

2. plann

Omschrijving E:

Kennis is zowel vastgelegd als stilzwijgend, zowel abstract als concreet, globaal en

gedetailleerd, en in principe voor iedereen toegankelijk. Waar nodig ontstaan naar behoefte

tijdelijke samenwerkingsverbanden. De organisatie heeft geen duidelijk centrum, er zijn

verschillende centra, en de organisatie is open naar buiten.


1 2 3 4 5

1 2 3 4 5


200

The questions for the subtasks gathering information, negotiating and schedulingwere identical. We therefore chose to only include the questions for one subtasks, the subtask of gathering information.

Part B: Gathering Information

Met informatie verzamelen om te kunnen roosteren bedoelen we hier het volgende:

1) Informatie over bijvoorbeeld vakantie, studie of de individuele wensen,

2) Informatie over randvoorwaarden en kwaliteitskenmerken, bijvoorbeeld voorwaarden (CAO/

het bevorderen van de continuïteit etc.).

B1.1 Verwacht u dat uw manier van informatie verzamelen zal veranderen als gevolg van het

roosteren met ZKR?

Ja / Nee

B1.2 Waarom denkt u dat uw manier van informatie verzamelen wel/ niet zal gaan veranderen?

B2.1 In welke mate hebt u via de volgende kanalen geleerd informatie te verzamelen om te

kunnen roosteren?

Niet Zeer veel

Opleiding 1 2 3 4 5

Van collega’s 1 2 3 4 5

Ervaring 1 2 3 4 5

Anders, nl. 1 2 3 4 5

…………………….

B2.2 Hebt u bij de verschillende manieren van leren informatie verzamelen ook andere

onderdelen van informatie verzamelen geleerd?

Ja / Nee

B2.3 Zo ja, waarin zit dit verschil dan precies?

B3 Hoeveel tijd denkt u nodig te hebben om iemand te leren informatie te verzamelen als

diegene:

Een ervaren verpleegkundige is:

Een onervaren verpleegkundige is:

Een niet-verpleegkundige is:

B4.1 Op welke manier is het volgens u het beste om uw kennis van informatie verzamelen over

te brengen?

Appendix

201

B5 Wat voor zaken beïnvloeden de wijze waarop u informatie verzamelt met betrekking tot het

dienstrooster? De inhoud van de informatie die u met betrekking tot het dienstrooster

verzamelt?

Niet Zeer veel

Beleid ziekenhuis 1 2 3 4 5

Ervaring collega’s 1 2 3 4 5

Eigen ervaring 1 2 3 4 5

Visie afdeling 1 2 3 4 5

Informatiesysteem 1 2 3 4 5

Organisatiestructuur 1 2 3 4 5

Verpleegkundige 1 2 3 4 5

visie afdeling 1 2 3 4 5

Anders, nl 1 2 3 4 5

(…………………..)

B6 Hoe creatief kunt u zijn bij het informatie verzamelen?

Niet creatief Zeer creatief

1 2 3 4 5

B7 Waar/ bij wie verzamelt u informatie:

Nooit Voortdurend

Roosteraars 1 2 3 4 5

Leidinggevenden 1 2 3 4 5

Verpleegkundigen 1 2 3 4 5

Overige collega’s 1 2 3 4 5


(…………………..)

B8 Hoe gemakkelijk* kunt u overleggen/ communiceren over verzamelen van informatie met:

Zeer makkelijk Zeer moeilijk

Roosteraars 1 2 3 4 5

Te roosteren 1 2 3 4 5

medewerkers

Leidinggevenden 1 2 3 4 5

Leken(opl.niv.gelijk) 1 2 3 4 5

Overige collega’s 1 2 3 4 5


(…………………..)


202

B9 In welke situaties (wanneer) verzamelt u informatie?

B10 Waarover verzamelt u informatie?

B11.1 Hoe vaak maakt u bij het informatie verzamelen gebruik van:

Nooit Voortdurend

E-mail 1 2 3 4 5

Handboeken e.d. 1 2 3 4 5

Telefoon 1 2 3 4 5

‘face-to-face’ 1 2 3 4 5

Vergadering 1 2 3 4 5

Informatiesystemen 1 2 3 4 5

Anders, 1 2 3 4 5

(nl. … )

B11.2 Hoe vaak zou u bij het informatie verzamelen gebruik willen maken van:

Nooit Voortdurend

E-mail 1 2 3 4 5

Handboeken e.d. 1 2 3 4 5

Telefoon 1 2 3 4 5

‘face-to-face’ 1 2 3 4 5

Vergadering 1 2 3 4 5

Informatiesystemen 1 2 3 4 5

Anders, 1 2 3 4 5

(nl. … )

B12

Vindt u het belangrijk dat u uw manier van informatie verzamelen kunt overdragen door

middel van voordoen?

Ja / Nee

B13 Om informatie te kunnen verzamelen is een bepaalde mate van kennis vereist. Is het

gemakkelijk om deze kennis over te dragen door deze voor te doen?


1 2 3 4 5

B14.1 Ligt dit aan een specifiek onderdeel van informatie verzamelen? Zo ja, welk(e)

onderde(e)l(en) van informatie verzamelen zijn dan moeilijk te verwoorden en misschien

alleen over te dragen door het voor te doen (en dus niet uit te leggen)?

B14.2 Waarom is dit onderdeel van informatie verzamelen zo moeilijk te verwoorden?

Appendix

203

B14.3 Hebt u het idee dat u de moeilijk te verwoorden onderdelen van informatie verzamelen kunt

verbeteren?

Niet te verbeteren Goed te verbeteren

1 2 3 4 5

B15 Hoe gedetailleerd is de moeilijk te verwoorden kennis die nodig is om informatie te

verzamelen? Denk hierbij bijvoorbeeld aan het verschil tussen kennis die u hebt van de

verpleegkundigen die u inroostert (zeer gedetailleerd) en die u hebt van de

ziekteverzuimcijfers voor het komende jaar (zeer globaal).

Zeer globaal Zeer gedetailleerd

1 2 3 4 5

B16 Vindt u het belangrijk dat de (ongeschreven) regels van de manier waarop u informatie

verzamelt worden vastgelegd?

Ja / Nee

B17.1 Hoe gemakkelijk is het om uw kennis van informatie verzamelen uit te leggen (zonder het

voor te doen)?


1 2 3 4 5

B17.2 Hoe gemakkelijk is het om uw kennis van informatie verzamelen uit te leggen terwijl u het

voordoet?


1 2 3 4 5

B18.1 In hoeverre is de kennis die nodig is om informatie te verzamelen vastgelegd/

beschreven?

Niet vastgelegd Geheel vastgelegd

1 2 3 4 5

B18.2 Hoe gemakkelijk is het om de kennis omtrent informatie verzamelen vast te leggen?


1 2 3 4 5

B19.1 In hoeverre is de vast te leggen kennis van het informatie verzamelen maar voor één

interpretatie vatbaar? Het gaat er hier om in hoeverre de interpretatie van datgene wat is

vastgelegd de mogelijkheid biedt om erover in discussie te gaan. Denk hierbij bijvoorbeeld

aan het verschil tussen het Latijnse schrift (zeer eenduidig, een A is namelijk een A en niet

wat anders) en globale termen zoals liefde (volstrekt niet eenduidig, hier geeft iedereen een

eigen interpretatie aan).

Niet eenduidig Zeer eenduidig

1 2 3 4 5

B19.2 Welke vast te leggen onderdelen van informatie verzamelen zijn wel eenduidig?


204

B19.3 Welke vast te leggen onderdelen van informatie verzamelen zijn niet eenduidig?

B20.1 Is het mogelijk om de in theorie goed vast te leggen onderdelen van het informatie

verzamelen ook daadwerkelijk beter vast te leggen?

Niet mogelijk Goed mogelijk

1 2 3 4 5

B20.2 Op welke manier kan dit dan worden verbeterd?

B21 Vindt u het belangrijk dat u kunt aangeven waarom u bepaalde informatie verzamelt?

Ja / Nee

B22.1 Hoe vaak kunt u bij het informatie verzamelen aangeven waarom u iets doet zoals u het

doet?

Zelden Bijna altijd

1 2 3 4 5

B22.2 Zou het mogelijk zijn dat u in de toekomst van meer onderdelen van informatie verzamelen

(dan u op dit moment kunt) beter kunt aangeven waarom u die doet zoals u ze doet?

Goed mogelijk Niet mogelijk

1 2 3 4 5

B23 Als het mogelijk is, hoe zou u dit dan in de toekomst kunnen aangeven en als dit niet

mogelijk is, waarom dan niet?

B24.1 Welke onderdelen van informatie verzamelen die u kunt beredeneren zijn zeer concreet?

Bijvoorbeeld: u kunt precies uitleggen waarom u bepaalde informatie verzamelt en kunt dit

verduidelijken aan de hand van een concreet voorbeeld.

B24.2 Welke onderdelen van informatie verzamelen die u kunt beredeneren zijn juist abstract?

Bijvoorbeeld: u kunt alleen aan de hand van een abstract voorbeeld uitleggen waarom u

bepaalde informatie verzamelt.

205

Summary

Introduction

Organizational innovation is often accompanied by problems [e.g. van de Ven 1986; Leonard-Barton 1988/1995; Geerts 1999; Laudon & Laudon 2000/2002; van Stijn 2006]. We argue that knowledge is a crucial factor in understanding organizational innovation and, consequently, in understanding the typical problems that come along with this process. We introduce a cognitive-semiotic model that enables empirical study in the form of hypothesis testing. We focus on the level of the individual as we want to understand the knowledge dynamics in organizational innovation and these knowledge dynamics occur within the individual.

We pose the following question: What actually happens to the knowledge of individuals that undergo organizational innovation? Our aim is to capture knowledge dynamics and to understand the relation between an innovative situation on the one hand and the knowledge of people within that innovative situation on the other. An objective of this study is to bridge theory and practice. Literature at this point falls short.

Theoretical framework

For studying knowledge dynamics Boisot [1995/1998] offers a good starting point. He presents the Information Space model for analyzing information flows within organizations. He characterizes information in three ways, 1] its level of codification, 2] its level of abstraction, and 3] its level of diffusion. This characterization can capture information in terms of organizational dynamics in a purposeful way; it allows distinguishing between different types of information based on a characterization on three different levels and it can reveal information changes. However, for our purpose, the I-Space model leaves out an essential aspect to characterize information changes, or more specifically, the knowledge dynamics that we are interested in. We argue that the change of knowledge always follows a set pattern, moving from one dimension to the next. The three subsequent dimensions involved are not all represented in Boisot’s I-Space model. Boisot does incorporate the levels of codification and abstraction, which more or less equal our dimensions of coded knowledge and theoretical knowledge. However, we argue that these two dimensions are founded in the dimension of sensory knowledge, which is a fundamentally different type of knowledge than the


206

diffusion level of the I-Space model. Furthermore, we do not consider the diffusion level to be an essential aspect of the knowledge itself, but rather to be a consequence of the other two characteristics.

We therefore introduce a cognitive-semiotic model, which it combines the empirical focus of the cognitive sciences with the theoretical focus on the nature of change of semiotics. Our model discerns three types of knowledge, which are cumulatively related, namely sensory knowledge, coded knowledge, and theoretical knowledge.

Sensory knowledge forms the foundation; it refers to the knowledge that has not been coded [yet]. Sensory knowledge is a specific case of tacit knowledge, present in behavior, that is behavioral expression [expertise] and perception. Essential in forming sensory knowledge and using it is the perception of difference, as this starts the process of knowledge dynamics. Sensory knowledge is exemplified in the question to what extent knowledge can be mimicked.

The second type of knowledge is coded; it refers to the presence of various kinds of codes, varying form sketches, pictures, alphabet characters to mathematical symbols, etc. An assumption to coded knowledge is the presence of some sort of consensus concerning the codes. Coded knowledge is exemplified in the determination of codes that are used.

Finally, theoretical knowledge provides structures, coherence and interrelations. It allows different perspectives and it is exemplified in the possibility to explain, to answer ‘why-questions. From sensory through coded to theoretical, learning time increases and the need to be physically present decreases.

Our cognitive-semiotic model can capture the essential changes that occur to the knowledge of the individual during the innovation process; the knowledge dynamics of the individuals actually embody the innovation process.

Operationalization

The empirical setting that we studied entailed the implementation of a software tool to support the staff planning in health care. The software that was implemented concerned the commercial product ZKR [in the 1990s developed at the university of Groningen; Jorna, et al. 1996]. We performed three measurements: 1] before the actual implementation [the old way of working], 2] after the training with the support software, 3] when the planners worked with the new planning support software for half a year. 35 planners participated in the first measurement, 31 planners in the second one and 24 planners in the third

Summary

207

measurement. Due to organizational changes and personal preferences a group of 18 planners participated in all three measurements. The third measurement was 1,5 years later than the first one. Our study focused on three different subtasks of planning: information gathering, problem solving [actual scheduling] and negotiating. Questionnaires were used for data collection. We formulated hypotheses in line with the thought that the transition from working manually to working with decision support, and in particular on a highly cognitive task as planning, makes the knowledge transition parallel the accumulation principle of our proposed model. For instance, knowledge should become more of the coded type for all subtasks.

Conclusion

Our findings show that the use of planning support leads to stronger codes. So essentially, use of decision support leads to more standardization and confirmation.

From a methodological point of view this study has made an important step to making innovation tangible in terms of knowledge; we measured the knowledge types of individuals within an organizational innovation process. Moreover, in measuring the knowledge types we also captured the dynamics of knowledge. And then, this thesis provides insight into this process of measuring knowledge dynamics to better understand organizational innovation.

This thesis showed how the understanding of knowledge [dynamics] can help to gain more insight into many different aspects of organizational innovation. Such as, making the duty roster implies certain things as using codes. These codes should be well understood. Or, how day-to-day routine is interrupted. A subtask such as negotiating, which is not directly supported by the implementation of the software does also undergo changes. Or the consideration that knowledge change is not a linear development, which is easily triggered just by working differently. The model that we presented enabled these insights. We therefore consider this model to be a powerful tool to study organizational innovation, a tool, which has the capacity to generate more insights.

208

Samenvatting

Inleiding

Organisationele innovatie gaat vaak gepaard met problemen [e.g. van de Ven 1986; Leonard-Barton 1988/1995; Geerts 1999; Laudon & Laudon 2000/2002; van Stijn 2006]. Wij stellen dat kennis een cruciale factor is om organisationele innovatie te begrijpen en dus ook om de typische problemen van dit proces te kunnen begrijpen. Wij presenteren een cognitief-semiotisch model waarmee empirisch onderzoek en in het bijzonder hypothese toetsend onderzoek kan worden gedaan. Wij doen ons onderzoek op het niveau van het individu, omdat wij met name geïnteresseerd zijn in de kennisdynamiek van organisationele innovatie. En wij stellen dat dit zich voltrekt op het niveau van het individu.

Wij stellen de volgende vraag: Wat gebeurt er eigenlijk met de kennis van individuen die in het proces van organisationele innovatie zitten? Ons onderzoeksdoel is om de kennisdynamiek van deze individuen tastbaar te maken en om de relatie tussen innovatie en kennis te doorgronden. Hierbij willen wij nadrukken theorie en praktijk aan elkaar koppelen.

Theoretisch raamwerk

Boisot [1995/1998] biedt een mooi beginpunt voor onderzoek naar kennisdynamiek in organisaties. Zijn Information Space model analyseert informatiestromen in organisaties. Deze worden door drie factoren gekenmerkt, 1] het niveau van codificatie, 2] het niveau van abstractie en 3] de mate van spreiding. De verschillende combinaties van deze drie factoren kan veranderingen in kennis in kaart brengen en daarmee dus de organisationele kennisdynamiek. Wij zien codificatie en abstractie ook als belangrijke factoren voor het typeren van kennis; deze factoren lopen parallel met onze dimensies van gecodeerde en theoretische kennis. Toch mist dit I-Space model van Boisot een belangrijk element voor ons onderzoeksdoel, een element dat, in ons idee, de basis van kennisdynamiek vormt, namelijk sensorische kennis. Deze dimensie verschilt fundamenteel van de factor spreiding die Boisot noemt.

Het cognitief-semiotisch model dat wij in dit onderzoek presenteren combineert de empirische focus van cognitiewetenschappen met the theoretische focus van het proces van verandering van semiotiek. Ons model onderscheid drie typen kennis, die cumulatief gerelateerd zijn, te weten, sensorische, gecodeerde en theoretische kennis. Sensorische kennis vormt the basis; dit is kennis die [nog] niet

Samenvatting

209

is gecodeerd. Sensorische kennis kan worden gezien als een soort stilzwijgende kennis. Deze kennis komt tot uitdrukking in gedrag en perceptie is hiervoor essentieel. Het tweede type kennis is gecodeerde kennis en refereert naar de aanwezigheid van codes van allerlei aard, zoals plaatjes, mathematische symbolen of karakters. Als derde onderscheiden we theoretische kennis; deze kennis biedt structuur, coherentie en legt relaties. Gecodeerde kennis verondersteld sensorische kennis en theoretische kennis verondersteld gecodeerde kennis en dus ook sensorische kennis. Theoretische kennis is duurt dus het langst om tevergaren.

Ons cognitief-semiotisch model biedt de mogelijkheid om de kennisdynamiek te vatten die plaats vindt bij individuen die in een innovatieproces zitten.

Operationalisatie

De empirische situatie die wij onderzocht hebben was de implementatie van een software programma om the ondersteunen van het maken van dienstroosters in de gezondheidszorg. Wij richtten ons hierbij op drie kennisonderdelen van planning, te weten, informatie verzamelen, onderhandelen en het roosteren zelf. Wij hebben drie afnames gedaan bij de planners die met de nieuwe software gingen werken: 1] voor de implementatie [dus de oude manier van werken], 2] na de training met de nieuwe software, 3] toen de planners een half jaar hadden gewerkt met de nieuwe software. Aan deze eerste afname deden 35 planners mee, aan de tweede 31 planners en aan de derde deden 24 planners mee. Door organisationele veranderingen en persoonlijke voorkeuren waren er 18 planners die aan alle drie de afnames mee deden. Het onderzoek besloeg in totaal anderhalf jaar.

Voor de afnames gebruikten wij vragenlijsten. En wij formuleerden hypothesen waaronder: het overgaan van handmatig dienstroosters maken naar roosteren met computerondersteuning, in het bijzonder als het gaat om een zeer cognitieve taak als planning, gebeurt volgens het cumulatieprincipe van ons cognitief-semiotisch model. Dit houdt in dat de gecodeerde kennis toeneemt.

Conclusie

Onze bevindingen laten zien dat gebruik van planningsondersteuning leidt tot sterkere codering. Dus in essentie betekent dit ook meer standaardisatie en confirmatie.


210

Vanuit een methodologisch oogpunt hebben wij met ons onderzoek een belangrijke stap gedaan om innovatie tastbaar te maken in termen van kennis; wij hebben de kennistypen van mensen gemeten die in een organisationele innovatie zaten en dat op meerdere tijdstippen in het proces. Zodoende konden wij ook de kennisdynamiek van deze mensen tastbaar maken.

Dit onderzoek heeft laten zien dat een beter begrip van kennis en kennisdynamiek ook tot een beter inzicht kan leiden van verschillende aspecten van organisationele innovatie. De dynamiek van stilzwijgende, gecodeerde en theoretische kennis haken in op het daadwerkelijke roosteren. Maar ook op de subtaak onderhandelen die niet direct wordt ondersteund door de planningssoftware. Of de dagelijkse routine die wordt onderbroken en doorbroken. Het zijn zaken die tastbaar kunnen worden gemaakt met het cognitief-semiotische model dat wij in dit proefschrift hebben gepresenteerd.

211

Sammendrag

Innledning

Organisatorisk innovasjon følges ofte av problemer [e.g. van de Ven 1986; Leonard-Barton 1988/1995; Geerts 1999; Laudon & Laudon 2000/2002; van Stijn 2006]. Vi mener at kunnskap er en avgjørende faktor for å kunne forstå organisatorisk innovasjon og dermed også de typiske problemene som følger med en slik prosess. Vi presenterer en kognitiv-semiotisk modell som kan brukes i forbindelse med utprøvning av hypoteser i empirisk forskning. Vi fokuserer på individet fordi vi først og fremst er interessert i kunnskapsdynamikken ved organisatorisk innovasjon og mener at denne finner sted i individet.

Vi stiller følgende spørsmål: Hva skjer egentlig med kunnskapen til individer som gjennomgår en organisatorisk innovasjonsprosess? Vårt mål er å kartlegge dynamikken og forstå forholdet mellom en innovativ situasjon på den ene siden og individers kunnskap i den samme sitasjonen på den andre. Vi vil legge vekt på å sammenligne teori og praksis.

Teoretisk rammeverk

Boisot [1995/1998] gir et godt utgangspunkt for forskning på kunnskapsdynamikk i organisasjoner. Hans modell Information Space analyserer informasjonsstrømmen i organisasjoner. Denne kjennetegnes av tre faktorer, 1) nivået for kodifisering, 2) nivået for abstraksjon og 3) nivået for diffusjon. De forskjellige kombinasjonene av disse tre faktorene kan kartlegge forandringer i kunnskap og dermed også kunnskapsdynamikk. Vi ser på kodifisering og abstraksjon som viktige faktorer for selve defineringen av kunnskap; disse faktorene er parallelle med våre dimensjoner av kodet og teoretisk kunnskap.

Allikevel mangler denne Information Space modellen et element som er viktig for vårt forskningsmål, et element som vi mener former grunnlaget for kunnskapsdynamikk, nemlig sensorisk kunnskap. Denne dimensjonen skiller seg sterkt fra faktorfordelingen til Boisot.

Den kognitiv-semiotiske modellen som vi presenterer i denne undersøkelsen kombinerer det empiriske fokuset fra den kognitive vitenskapen med det teoretiske fokuset fra semiotikkens forandringsprosess. Vår modell skiller mellom tre typer kunnskap som er kumulativt relatert; sensorisk, kodet og teoretisk


212

kunnskap. Sensorisk kunnskap danner grunnlaget; dette er kunnskap som (enda) ikke er kodet.

Sensorisk kunnskap kan sees som stilltiende kunnskap. Denne kunnskapen gjør seg gjeldende i oppførsel og persepsjon er her essensielt. Den andre typen kunnskap er kodet kunnskap og referer til alle type koder; alt fra matematiske tabeller, alfabetiske symboler til tegn og bilder. Den tredje typen er teoretisk kunnskap; denne gir struktur, sammenheng og legger bånd i interkommunikative forhold. Kodet kunnskap forutsetter sensorisk kunnskap. Teoretisk kunnskap igjen forutsetter kodet kunnskap og dermed også sensorisk kunnskap. Det tar altså lengst tid å bygge opp teoretisk kunnskap.

Vår kognitiv-semiotiske modell gir muligheten til å forstå kunnskapsdynamikken som finner sted i individet og i en innovasjonsprosess.

Utførelse

Den empiriske situasjonen vi undersøkte var implenteringen av et software for oppsett av arbeidskalender innenfor helsevesenet. Vi fokuserte på følgende tre kunnskapsfaktorer; 1) samling av informasjon, 2) forhandling og 3) selve oppsettsprosessen. Vi foretok tre målinger; 1) før implenteringen (altså den gamle måten for oppsett), 2) etter opplæring til det nye softwaren og 3) etter at planleggerne hadde jobbet med softwaren i et halvt år. I den første målingen deltok 35 personer, i den andre deltok 31 personer og i den siste deltok 24 personer. På grunn av organisatoriske forandringer og personlige preferanser var det 18 personer som deltok i alle målingene. Undersøkelsen tok i alt ett og et halvt år. Ved målingene brukte vi spørreskjemaer. Vi formulerte følgende hypotese:

Overgangen fra å jobbe manuelt til å jobbe digitalt, spesielt ved en så kognitiv oppgave som planlegging, gjør at kunnskapsovergangen skjer parallelt med vår kognitiv-semiotiske modell.

Dette betyr altså at mengden kodet kunnskap vil øke.

Konklusjon

Våre resultater viser at digital støtte til planlegging fører til en sterkere koding. I bunn og grunn betyr dette økt standardisering og bekreftelse.

Ut ifra et metodologisk synspunkt har vi med vår undersøkelse gjort et viktig steg mot å gjøre innovasjon håndgripelig i forhold til kunnskap: Vi har målt kunnskapsformene til mennesker innenfor en organisatorisk innovasjonsprosess og

Sammendrag

213

på flere tidspunkter innenfor denne prosessen. På denne måten kunne vi også gjøre kunnskapsdynamikken håndgripelig for disse menneskene.

Ut ifra et teoretisk perspektiv kan vi si følgende: Vi har samlet data fra helseinstitusjonen Bartiméus. Denne institusjonen kjennetegnes som en stor institusjon som tilbyr små boenheter og omsorg av handikappede mennesker.

For å forstå verdien av våre resultater i forhold til generalisering vil vi vektlegge kunnskapsformen koding. Ettersom kodet kunnskap baserer seg på konsensus kan man argumentere for at å oppnå enighet innenfor en liten gruppe er forskjellig fra å oppnå enighet i en større gruppe. Siden Bartiméus er en stor institusjon er det mulig at dette påvirker økningen av kodet kunnskap på en annen måte enn den ville gjort hvis institusjonen var mindre. I tillegg er det mulig at mindre enheter er mer avhengig av sensorisk kunnskap, noe som igjen kan føre til en tilbakegang av kodet kunnskap i en liknende prosess, slik som den ved helseinstitusjonen Bartiméus. Når det gjelder generalisering forventer vi liknende resultater ved andre helseinstitusjoner.

Videre retter vi oss mot kunnskapsområdet planlegging. Vi har valgt dette kunnskapsområdet fordi det involverer en høy grad av kognitiv aktivitet og fordi vårt fakultet har utviklet ekspertise på dette området. Disse valgene har altså ikke ført til en begrensning av forståelsen av våre resultater.

Videre var planleggingen satt opp som en generell oppgave. Det vil si at den besto av flere uavhengige oppgaver. Planlegging som kunnskapsområde ble altså valgt vilkårlig og vi forventer derfor at andre kognitive oppgaver vil gi liknende resultater. På grunnlag av de overstående argumenter forventer vi at våre resultater også vil gjelde for andre typer planlegging, eksempelvis transportplanlegging og produksjonsplanlegging.

Videre har vi fokusert på følgene implikasjonsprosessen kan ha på kunnskapsaspektene. Når en organisasjon vil innovere vil vi understreke at særegenheter i kunnskap har enkelte konsekvenser. Ettersom kunnskap alltid er et kumulativt resultat bør det være klart at et ønsket kunnskapsresultat kan kun oppnås når den nødvendige kunnskapen er tilgjengelig.

Med andre ord, du kan ikke starte en implikasjonsprosess uten at de involverte parter har kunnskap om deres egen kunnskap, fordi det da ikke er noe fundament å bygge på. Kunnskap kan ikke forandre seg selv. Bare målrettet forandring kan føre til ønsket effekt, og dette må fasiliteres av strukturerte retningslinjer. Dette


214

krever en regelmessig oppfølging og ikke bare sporadisk. Det må skapes et miljø hvor de involverte kan utfolde seg.

For å konkludere, den kognitiv-semiotiske modellen som vi har satt frem ble brukt empirisk for å undersøke kunnskapsdynamikken til individer som gjennomgikk en organisatorisk innovasjonsprosess.

Denne innovasjonsprosessen inneholdt en ytre påvirkning. Andre former for innovasjonsprosesser kan også ha fordel av vår modell, for eksempel produktinnovasjon.

university of groningen the knowledge dynamics of ... · ter verkrijging van het doctoraat in de...

Documents