a transparent agile change -...
TRANSCRIPT
Faculty of Health and life sciences
Department of Psychology
Master Thesis 5PS22E, 30 ECTS
Spring 2020
A Transparent Agile Change
- Predicting a Transparent Organizational Change from
Change Recipients’ Beliefs and Trust in Management
Author: Towe Nilsson
Supervisor: Andrejs Ozolins
Examiner: Rickard Carlsson
Abstract
The popularity of agile methodologies is steadily increasing. This study is an intent to
balance the agile change literature with a psychological perspective and quantitative measures
of an agile change made within a Swedish organization. Organizational change recipients’
beliefs (discrepancy, appropriateness, valence, efficacy, & principal support) and trust in
management were measured in an online survey to see how well these variables could predict
a successful agile change towards transparency. The results indicate a lack of support for
several previously cited success factors in the agile literature and a need for more quantitative
and research-driven literature. No support could be found for a relationship between
discrepancy, appropriateness, valence, principal support, trust in management, and the
outcome of a successful implementation of transparency. Efficacy was found to be a
significant and robust predictor of the outcome. More research is needed to ensure the
generalizability of the results.
Keywords: organizational change, change recipients’ beliefs, trust in management,
agile adoption, Scaled Agile Framework (SAFe), Scrum, transparency
Acknowledgments
This thesis could never have been made without the support and interest from the right
persons. First, I want to thank all key persons within the studied organization who believed in
and contributed to this research. I want to especially highlight my two main points of contact
in the organization, who contributed with many great ideas and critical discussions. For seeing
the value of this study for their change initiative as well as for an increased understanding of
agile change initiatives in general. Thank you to all who participated and showed interest in
the study, for lending me your time, effort, and enthusiasm. This thesis is the result of your
support, interest, and well-aimed questions. Your reflections have not only produced the
results but also continuously challenged me to think critically about my research. Although
this thesis can only reflect a small part of your contribution, the value continues within your
organization as well as in my improved understanding of what it means to be a part of a large
agile change initiative. It has been a true adventure to be a part of your change journey and I
have learned more than I ever could have hoped for.
Thank you to my ever so supportive supervisor and program coordinator at Linnaeus
University, Andrejs Ozolins. Your encouraging words, timely and constructive feedback gave
me the energy I needed, when I needed it the most. Throughout the process, you have
encouraged me to feel proud of my work and saw the value I could not always see myself.
Your kind words made me blush more than once.
Finally, I want to thank my family and friends for hearing my thoughts and
proofreading my text, clearing out some confusion and mistakes (missing the “r” in Scrum is
not a mistake you want to publish…), and nicely telling me when my thoughts were not really
in print. And not least, thank you for supporting me throughout my studies, knowing when to
distract me, when to give me space and when to lend me your energy. I am lucky to have you.
With all these contributions this work is much better than I could have made on my
own, but of course, any potential confusion and mistakes that remain are on me. Some of
them I have already had the opportunity to learn from, others I hope to reflect upon in future
discussions.
Authors Note
This study was preregistered at Open Science Framework, including planned
hypothesis, measured variables and demographics, planned sample and stopping rule for data
collection, planned data analyses, alpha-level and rules of data exclusion (e.g.,, treatment of
outliers & influential cases). See registration at https://osf.io/8ntw7
4
A Transparent Agile Change: Predicting a Transparent Organizational Change from
Change Recipients’ Beliefs and Trust in Management
The continuous pressure to produce more, better, and faster than your competitors is
vital for the survival of an organization. Especially in the software business where the pace is
accelerating. Agile methodologies grew as a reaction to be able to faster deliver value to
customers with less risk of being outrun by your competitors without having had the
opportunity to release the product you are working on. The popularity of agile methodologies
is steadily increasing (CollabNet VersionOne, 2019). From traditionally being used in
software development, the methods are today considered useful in other areas of product
development as well (Long & Starr, 2008). They are adopted in a variety of sectors such as
education, communications, marketing, sales (Oprins et al., 2019), manufacturing, training,
research (Totten, 2017), hardware development, and business (Scaled Agile, 2019a). With this
increase in usage, there is also an increased need for empirical research to better understand
these changes and thus better guide practitioners. Even if claimed to have many benefits (e.g.,,
Scaled Agile, 2019a; Schwaber & Sutherland, 2017), agile methods are not widely
researched, with the increased popularity mostly driven by consultants and practitioners
(Tripp, 2012). The success factors and challenges of adopting agile methodologies have been
widely referenced, often in case studies describing success stories or more qualitative studies.
Although important, a balance is needed to gain more profound knowledge and a need for
more academic research can be noted in the field (e.g., Dikert et al., 2016). The trend can be
further problematized in large-scale agile frameworks, intending to implement agile on a
larger scale in organizations, where the research is considered scarce (Dingsøyr et al., 2019;
Turetken et al., 2017). According to a literature review by Dikert et al. (2016), 90 % of the
literature on large-scale agile changes is based on experience reports. Due to this, some
caution is appropriate when drawing conclusions from the agile literature. An interest in a
more psychological perspective can also be noted in the field (e.g., So & Scholl, 2009).
Researchers or practitioners are often from an engineering background, several of which
reference traditional psychological theories (e.g.,, Pedrycz et al., 2011; Tripp et al., 2016).
Findings by Gandomani and Nafchi (2016) indicate several of the most important challenges
are of a social, people-oriented nature, rather than a technological. Barroca et al. (2019) have
also found that the success of larger agile transitions is rather dependent on more
psychological factors such as commitment to change, communication, engaging people,
management support, team autonomy, and training.
5
One aspect of agility that is often referenced (e.g., Kalenda et al., 2018; Scaled Agile,
2019b; Schwaber & Sutherland, 2017) is transparency. This is both considered a prominent
part of commonly used agile methodologies and frameworks (Scaled Agile, 2019b; Schwaber
& Sutherland, 2017), as well as a found challenge in agile transitions (e.g., Kalenda et al.,
2018; Barroca et al. 2019). Transparency was also one of the biggest reported benefits among
respondents who had been working according to SAFe, a framework for scaling agile
methodologies in an organization, for several years (Laanti & Kettunen; 2019). This research
is an intent to follow an agile change within an organization, mainly focusing on
understanding the potential predictors of the intended outcome of increased transparency, and
by this enable a deepened understanding of changes specifically targeted at transparency.
To increase understanding of how to successfully implement agile methodologies and
transparency, this study is an intent to bridge the gap of the practitioners-driven and
qualitative focus in agile literature and shortage of quantitative studies. This will be done by
applying well-established psychological literature and theory of organizational change to
explore how well this theoretical foundation can be applied in agile change initiatives.
Specifically, the study intends to measure how change beliefs and trust in management can be
related to a successful implementation of transparency in a company going through an agile
transition.
The thesis is arranged in the following sections: First, a theoretical background,
including a short introduction to agile methodologies and an overview of research from the
field on transparency, change recipients’ beliefs, and trust in management. This will be
followed by the proposed research question and hypothesis, a methodological part, results,
and lastly a discussion including implications for future research.
Theoretical Background
Scrum, SAFe, and Agile
Agile methodologies started as a reaction in the software development business to
more traditional project methods (such as Waterfall). Compared to traditional linear methods
often focused on releasing a finished end-product, agile methods are created to be more
flexible to better adapt to changed circumstances (Beck et al., 2001). It is an iterative and
incremental way of developing, continuously adding value to a product, combined with a
close relationship with the customer, or end-user, to allow for continuous feedback and
improvements. There are a variety of agile methodologies (e.g., Scrum, Kanban, eXtreme
Programming, Crystal Clear, & DSDM, Björkholm & Brattberg, 2016). Although agile
methodologies are often discussed as a coherent set of working methods in research, they
6
include a wide range of practices, not necessarily compatible (Tripp, 2012). The
methodologies are often fragmentized with only parts being used (Pries-Heje & Baskerville,
2017). This research focuses on Scrum, which is the most commonly applied agile method
(e.g., CollabNet VersionOne, 2019) and SAFe, which is a framework for scaling agile within
the organization (Scaled Agile, n.d.). The term “scaling agile” refers to the introduction of
agile methodologies on a bigger scale within an organization. SAFe is the agile scaling
approach reported mostly used in practice (CollabNet VersionOne, 2019).
Transparency
Although there might be differences in conceptualizations of transparency regarding
what is disclosed and level of analysis, the core of defining transparency is similar in
conceptualizations (Palanski et al., 2011), involving availability, openness, and disclosure of
information. This study applies a definition close to that used by Palanski et al. (2011), who
defined transparency in a team as “the sharing of relevant information and explanations within
a team to enable its members to carry out their responsibilities within the team” (p. 203). A
similar definition is used by Laanti et al. (2011), defining visibility and transparency in an
agile context as information sharing and tracking of program/projects.
According to Iivari and Iivari (2011), openness, communication, and honesty are some
of the essential cultural attributes in the agile literature. Transparency is described as part of a
shift to a collaborative culture (Stettina & Hörz, 2015). It is one of four core values in SAFe, a
part of an alignment to work more efficiently in a fast-paced environment where “all work is
visible, debated, resolved and transparent” (Scaled Agile, 2019b, para. 5). It is also described
as a pillar in Scrum (Schwaber & Sutherland, 2017).
Transparency is described as a critical factor during agile transitions (Dikert et al.,
2016; Barroca et al., 2019). It supports the alignment of activities at different organizational
levels (Suomalainen et al., 2015). Transparency has been reported as one of the three biggest
benefits of adopting SAFe (Laanti & Kettunen, 2019). Agile practices have been found to
increase the transparency of projects on a team level as well as the organizational level
(McHugh, Conboy, & Lang 2012). Another benefit of transparency is the teams’ ability to see
the progress of their work (Asnawi et al., 2012). Transparency might be especially beneficial
in a larger agile context to visualize interdependence between multiple teams and the flow of
work items from a general portfolio down to team-level (Horlach et al., 2018).
Despite the found benefits, transparency is also described as one of the challenges in
an agile transformation (e.g., Figalist et al., 2019; Kalenda et al., 2018; Suomalainen et al.,
2015). This paradox was also found in research by Laanti et al. (2011), describing how
7
transparency, along with requirements management, was “perceived as among the most
beneficial, yet also among the most challenging, areas of agile development throughout the
survey population” (p.277). Although not academically approved, organizational
representatives from Healthwise (Long & Starr, 2008), describe how transparency was
considered one of the main reasons for implementing agile methods in the organization, yet
several developers were lost due to their perceived misfit with the transparent environment.
Schnitter and Mackert (2011) found that several of the team members’ arguments proposed
against Scrum were related to transparency, such as “Scrum forces me to report every failure
and impediment. This puts me under pressure” (p.212). The resistance was found to be
reduced by explaining that transparency was needed for the teams to be able to work more
independently. The importance of having the right tool to facilitate transparency and
visualization has been found in several studies in previous research (e.g., Conboy, 2009;
Kalenda et al., 2018).
Change Recipients’ Beliefs
Although it is important to measure performance and financial aspects, theorists (e.g.,
Oreg et al., 2013) argue psychological factors and employees need further consideration for
the success of organizational changes. One theory with an employee-centric approach is
organizational change recipients’ beliefs (OCRB, Armenakis & Harris, 2009) intending to
understand what factors determine if a change recipient will be motivated to support a change.
OCRB focuses on five different beliefs of interest for change motivation and buy-in among
employees (Armenakis & Harris, 2009). Mainly, discrepancy, appropriateness, efficacy,
principal support, and valence. Discrepancy is related to whether there is a perceived need for
a change, while the change is considered appropriate if it is thought to address the
discrepancy. Efficacy concerns employees’ perception of themselves to have the capability of
implementing the change. Principal support concerns the perceived change support from
managers, change agents, and peers. Lastly, valence is the perceived personal value of a
change.
Change recipients’ beliefs have been related to essential variables during
organizational change, such as change resistance and support for change (Abdel-Ghany, 2014;
Armenakis et al., 2007). Structural equational modeling has suggested change recipients’
beliefs mediate the relationship between employee readiness factors and behavior of either
supporting or resisting the change (Abdel-Ghany, 2014). However, indications of a reversed
causality have been found by Hameed et al. (2019) with results indicating that change
recipients’ beliefs positively effect employees’ readiness for change. OCRB have also been
8
used to describe change readiness prior to organizational change (Doyle et al., 2009; Hameed
et al., 2019).
People buy-in has also been found to be an essential prerequisite in organizations
going through large-scale agile transitions and is in a qualitative study by Gandomani and
Nafchi (2016) described as a prerequisite for agile transformation. Change beliefs have been
found to be higher among those going through a lean change compared to control groups who
did not (Mazur et al., 2011). Principal support and efficacy have previously been found to
predict technology acceptance (Brown, 2009), which might be closely related to the outcome
measured in this study as transparent tool usage. Change efficacy also predicted personal
initiative of adopting the new technological system.
Although not scientifically approved, to the knowledge of the author, awareness of the
importance of the beliefs in OCRB can be found in the Agile practitioner community. The
organization Mountain goat, a prominent actor in the Agile practitioners’ community, has
developed a model with similar variables as OCRB. The ADAPT model refers to the
importance to create Awareness and Desire when implementing agile methods. To develop
the right Ability, Promote successes of the implementation and to Transfer the implications of
agile to other parts of the organization (Cohn, 2010). Where awareness can be related to
perceived discrepancy, desire to valence, ability to efficacy, and promoting success to
principal support and appropriateness.
Discrepancy and Appropriateness
The importance of a perceived need for a change, discrepancy, can be seen in several
classic theories of organizational change, such as the eight-step model by Kotter (2012) and
the unfreeze-change-refreeze model by Kurt Lewin (see Burnes, 2020 for an analysis and
review of the model and its origin). Support can also be found in several studies (e.g.,
Armenakis & Harris, 2002; Abdel-Ghany, 2014; Hameed et al., 2019). Gandomani and
Nafchi (2016) found pre-startup assessment to be one of seven prerequisites before an
organization starts a large-scale agile transition. One challenge noted in the literature of agile
change initiatives is top-down steered changes which might make it harder to understand
reasons for initiating a change (Dikert et al., 2016).
The importance of a proper diagnosis prior to change has been raised in research
dating back to 1965 (Armenakis et al., 2007) and can be considered a way to assess needed
change interventions. Careful planning and deliberation were found to be related to less
uncertainty among change recipients (Rafferty & Griffin, 2006). Perceived weaknesses in a
proposed change have been found to be a reason for resisting change (Lunenburg, 2010). In a
9
report of readiness for implementation of a national framework for early childhood care in
Ireland, a relationship was found between knowledge of the change and appropriateness
(Doyle et al., 2009).
Although not directly measured, findings from the agile research literature indicate the
importance of appropriateness in this context as well. No clearly defined goal from the
management of an agile transition might risk a feeling among employees that agile
methodologies can be replaced (Dikert et al., 2016). Too strict, “by-the-book”
implementations, without adjusting for specific needs, have been problematized as a challenge
in agile transitions. Teams who adjusted practices to their environment did better than those
who did not, especially when implementing agile at a larger scale (Dikert et al., 2016). Dikert
et al. (2016) found that practices might need to be adjusted to different contexts. Starting with
a pilot before implementing the bigger change was found to demonstrate the suitableness of
the agile change and increase acceptance. The importance of appropriateness can potentially
also be related to the found positive influence of success stories (Dikert et al., 2016).
Valence
The importance of valence for buy-in among change recipients is based upon Vroom’s
VIE theory of motivation (Armenakis et al., 2007), lifting the importance of the perceived
attractiveness of an outcome. This can be of both an intrinsic and extrinsic nature (Armenakis
et al., 2007). Autonomy is proposed as intrinsic motivation (Van den Broeck et al., 2016) and
has also been found to be related to increases in job satisfaction in an agile context (Tripp et
al., 2016). Perceived and expected costs and benefits of a change have been related to attitude
toward job changes (van Dam, 2005; Giangreco & Peccei, 2005), resisting change (Giangreco
& Peccei, 2005; Lunenburg, 2010), and cognitive resistance (Oreg, 2006). A lack of a clear
understanding of the benefits of Agile transitions has been found to be related to skepticism
towards agile (Dikert et al., 2016). As discussed above, benefits of transparency have been
reported in previous research, as well as its challenges and drawbacks. Further, since
transparency could have the potential to negatively affect employees (e.g., information about
performance), it is not necessarily perceived as valuable. It can thus be proposed that valance
of transparency could be perceived as either high or low in an agile change.
Efficacy and Principal Support
Change efficacy concerns employees’ perception of themselves to have the capability
of implementing change (Armenakis et al., 2007). The theoretical foundation of this variable
dates back to Bandura’s (1997) theory of self-efficacy, where one’s perceived capability is
proposed to be context-specific, in this case measured in a change context. This variable
10
might also be related to change support since support from managers and role modeling are
considered important to improve self-efficacy (Stajkovic & Luthans, 1998). A relationship
between self-efficacy and perceived support has been found in agile contexts as well (Seger,
Hazzan, & Bar-Nahor, 2008). However, the relationship was only found on the managerial
level and not on a junior level. Chreim (2006) found that employees consider existing
capabilities before embracing a change. This is aligned with change-related self-efficacy as a
found predictor for a positive and proactive organizational change response (Ahmad, 2019).
Self-efficacy has also been found to mediate the relationship between transformational
leadership and reactions to change (Bayraktar & Jímenez, 2020).
One study explicitly measuring self-efficacy in an agile context was made by Seger et
al. (2008) which found self-efficacy to be related to an agile orientation among managers but
not among juniors. However, self-efficacy was found to be related to change motivation
among both juniors and managers. Although change efficacy has not been extensively
addressed in agile research, the importance of related factors such as prior knowledge and
management support have been found to be key success factors in large-scale agile transitions
in a study by Kalenda et al. (2018). Further, a study by Sundborg (2019), found self-efficacy
and principal support to be mediators in the relationship between knowledge and affective
commitment. The found importance of knowledge in agile transitions might thus rather have
an indirect effect on the outcome of a change, through the perceived efficacy. The importance
of knowledge, training, and coaching has been found by several researchers in agile
transitions (Asnawi et al., 2012; Dikert et al., 2016; Mohan, 2018). The importance of training
has also been found in a study by Gandomani and Nafchi (2016), where the absence of proper
training was perceived to cause several problems in the transition, low confidence being one
of them.
Supportive management is described as an essential part of the Lean leadership
proposed by SAFe, described as role models, and actively driving the change towards large-
scale agile adoption (Scaled Agile, 2020). Change support has also been found to be
important in research on agile transitions. Korhonen (2013) propose that winning the active
participation and support from all members of the organization is needed to reach the benefits
of an agile change. Although general support (Korhonen, 2013), as well as teamwork support
(Kalenda et al., 2018), has been referred to, the importance of management support has been
put forward as a key success factor in agile changes (e.g., Barroca et al., 2019; Kalenda et al.,
2018). A literature review of large-scale agile transformations by Dikert et al. (2016) found
management support to be considered a success factor in 38 % of the reviewed cases. Further,
11
resistance on management levels was reported as a challenge in 10 % of the literature
reviewed (Dikert et al., 2016). Gandomani and Nafchi (2016) found that participants
perceived that low commitment among management could create challenges in an agile
transformation, in both their own management role and responsibilities, as well as the roles of
others. Perceived level of support was found to be related to agile orientation (Seger et al.,
2008).
Trust
This research is built on the theoretical foundation and definition of trust proposed by
David, Mayor, and Schoolman (1995), defining trust as:
the willingness of a party to be vulnerable to the actions of another party based on the
expectation that the other will perform a particular action important to the trustor,
irrespective of the ability to monitor or control that other party (p.712).
A similar definition has been proposed by Rousseau et al. (1998), defining trust as a
‘‘psychological state comprising the intention to accept vulnerability based upon positive
expectations of the intentions or behavior of another.’’ (p. 395).
Several findings indicate trust-building is related to less resistance to change
(Bruckman, 2008; Oreg, 2006). Oreg (2006) found a relation to cognitive, affective, as well as
behavioral resistance and action against change initiatives. Trust in management has also been
positively related to commitment to change (Meyer & Hamilton, 2013). Apart from this, trust
in supervisors has been found to mediate the relationship between openness to change and
predictors such as managerial communication, and participations (Ertük, 2008).
According to Iivari and Iivari (2011), the agile literature is highly people-oriented,
where trust is one of the essential cultural attributes. Lack of trust between management and
engineers has been found to be a reason for reimplementing a waterfall model (Murphy &
Donnellan, 2009). In the agile literature, trust is often described from the perspective of teams
needing trust from managers to operate in an agile way (e.g., Sidky, 2007; van Manen & van
Vliet, 2014; Tessem & Maurer, 2007). It is also described of importance for teams to develop
trust from customers (e.g., Bang, 2007). The importance of trust within agile teams has also
been addressed (McHugh et al., 2012; van Manen & van Vliet, 2014), where agile practices
such as daily stand-ups and sprint planning are proposed to enhance trust in the team
(McHugh et al., 2012).
In a study by Gregory et al. (2016), trust was identified as one of the challenges in
agile methods. Transparency is described in the SAFe framework as “an enabler of trust”
(Scaled Agile, 2019b, para. 13) and has been found to improve trust and collaboration in
12
qualitative research (Stettina & Hörz, 2015). In a qualitative study by McHugh et al. (2012),
agile practices and transparency are described as related to increases in trust. However, to the
extent of the author’s knowledge, the importance of team members' trust in management
during an agile transition has not been addressed in the literature. McHugh et al. (2012)
propose future research to investigate other aspects of trust. Further, when increasing
transparency and availability, of for example performance-related information, trust might be
even more important. This since trust in management was found to be potentially critical for
employee’s change commitment during conditions that might be perceived as a risk (Meyer &
Hamilton, 2013). Although perhaps not the intention, performance-related information can be
argued to have the potential to be used against the benefits of an employee. The need for trust
in management acting in the best interest of the employee should thus potentially be even
more important due to the potentially higher vulnerability. One indication of the proposed
relationship between vulnerability and transparency can be found in a qualitative interview
study by Schnitter and Mackert (2011), where interviewees raised perceptions of having to
demonstrate their flaws due to the increased transparency. Examples of this could be noted in
the following two statements against Scrum: “My manager might get a wrong impression of
my work in case someone reports how slow my work progresses” and “Scrum forces me to
report to the team every failure and impediment. This puts me under pressure” (p. 212). Most
of the resistance was found to be removed when the necessity of transparency was pointed out
to the team members as an enabler for the team to work more independently (Schnitter &
Mackert, 2011). A similar finding was demonstrated by Mohan (2018) who found
transparency to be related to perceptions of having to show weaknesses and the importance of
assurance from managers that the information would be used for improvements rather than
demonstrating weaknesses. This challenge was more apparent in the initial phase of the
transformation.
Research Question and Hypothesis
As has been discussed above, this study made it possible to study transparency more
closely, one of the found challenging aspects of agile methodologies (e.g., Kalenda et al.,
2018). This study thus has the potential to enable valuable insights and a deepened
understanding of changes specifically targeted at transparency. The above described
theoretical foundation describes how change beliefs and trust in management have been
related to important variables for change success. And although the agile change literature
relies to a large extent on case studies and more qualitative research (e.g., Gandomani and
Nafchi, 2016; McHugh et al., 2012; Schnitter and Macker, 2011; Stettina & Hörz, 2015),
13
indications of the importance of the described variables can be found in agile change literature
as well. This study aims to contribute to a quantitative foundation and research-driven
approach for potential relationships between the proposed predictors and the outcome of an
agile change initiative. In this study, the measured outcome consists of a successful
implementation of transparency, through a transparent tool usage. In this case a transparent
tool usage refers to the use of the tool for transparency purposes, as this was the proposed
goal of the studied change initiative. The following research question and hypothesis are
proposed:
RQ1: Can beliefs about an organizational change and trust in management predict
successful implementation of transparency within an IT-department of a Swedish
organization?
H1: Change recipients’ beliefs and trust in management can predict successful
implementation of transparency.
Methodology
Organization
The study was conducted primarily in the IT-department of a larger heavy industry
organization in the southern part of Sweden. The organization was currently going through a
change of working methods towards agile methodologies when the study was conducted. The
change was driven by the implementation of a new tool for development (Azure DevOps
Services), aimed to be used for increased transparency of work within and between teams.
The tool also aimed to make the work of the teams more transparent on managerial levels
(e.g., project & program level), as well as for stakeholders outside the department.
Participants
A survey was sent out to all registered users of the tool. 220 of these were estimated to
be potential participants. In total, 149 responses were received. 118 of the respondents
indicated they were using the tool. However, a final sample of 110 participants remained
since 8 of the respondents indicated they had been in the company for less than half a year, or
less than a year without specifying how much less. All respondents who indicated they had
been in the organization for more than half a year remained in the sample. Since a tenure
below half a year would indicate respondents might not have been in the company for enough
time to have experienced the entire change, the responses were considered less reliable.
According to power guidelines by Cohen (1992) and Field (2009), the sample size was
enough for the study to have an acceptable power, 80 % power, to find relationships of
medium effect size with the proposed analysis and variables.
14
The respondents consisted of 77 males and 28 females, whereas five preferred not to
answer. The mean age was 40.8 years (Confidence Intervals, CI: 38.8, 42.8; SD: 10.0) and
had a range from 23 to 62 years. The respondents had on average been within the company
for 5.2 years (CI: 4.0, 6.5; SD 6.2). Years within the company ranged from 8 months to 25
years.
Measurements
All variables were measured with questionnaires in an online survey. Where possible,
previously developed, psychometrically valid, and reliable scales were used.
Transparency - Successful Implementation of Agile Methodologies
The study aimed to understand the predictors of a successful implementation of agile
methodologies. Success is operationalized to consist of an increase in transparency, measured
by a transparent use of the tool being implemented to drive the agile processes in the
organization. Since no scale could be found to measure the proposed outcome variable, a
scale was developed to measure a transparent tool usage. The scale was developed to be
aligned with operationalizations made by representatives of the organization of the goals of
the initiated change towards transparency, as well as the agile framework used (SAFe) and the
tool being implemented (Azure DevOps Services). Since transparency according to the SAFe
framework consists of visibility of information between several work roles and organizational
levels (Scaled Agile, 2019b), items were framed on a more general level to include different
perspectives and make it possible to include different roles in the sample. Inspiration could
also be found in other researchers’ definitions of transparency as visualizing (Laanti et al.
2011), giving alignment, keeping people updated and seeing the progress (Suomalainen et al.,
2015) as well as the amount of information shared as well as its thoroughness (Palanski et al.,
2011).
Items were measured on a 5-point Likert scale ranging from strongly disagree to
strongly agree. The items concerned whether the agile tool was used in a way that increased
transparency, e.g., if the tool was used to easily identify dependencies between teams (see
items in Appendix A). The scale used a design and phrasing of items similar to what has been
applied in a scale by Hameed et al. (2019) to measure employees’ readiness for change. The
items were tested in a pilot survey with representatives from the organization to ensure they
were properly understood, and minor adjustments were made in the descriptions to adjust for
feedback from the pilot.
The scale initially had six items. However, one of the scale items was excluded due to
low correlations with three other items of the scale (r = .11, r = .19, r = .24). An initial factor
15
analysis had been planned but since the sample excluding low correlating items was deemed
most appropriate size was too small to conduct a reliable factor analysis (see
recommendations for at least 300 cases, Tabachnick, Fidell, & Ullman, 2007), excluding low
correlating items was deemed most appropriate considering low correlations might indicate
the item does not measure the same underlying factor as the other items.
Change Recipients’ Beliefs
Change recipients’ beliefs were measured with a psychometrically sound, 24-item
scale developed by Armenakis et al. (2007). The items are measured on a 7-point Likert scale
ranging from strongly disagree to strongly agree. Change beliefs were categorized into the
following five subscales (with example items referenced within parenthesis): Discrepancy
(e.g., “We need to change the way we do some things in this organization”), appropriateness
(e.g., “This organizational change will prove to be best for our situation”), efficacy (e.g., “I
have the capability to implement the change that is initiated”), principal support (e.g., “The
top leaders support this change”), and valence (e.g., “This change will benefit me”). All sub-
scales had an acceptable coefficient alpha above .80 in the validation studies of the scale
(Armenakis et al., 2007).
Trust in Management
To measure trust in management, a scale developed by Mayer and Gavin (2005) was
used. The full 10-item scale was not used since it was argued by the researchers to be
multidimensional. Instead, the more generic, five-item solution proposed by the researchers
was used. The five-item solution is an updated version of the original scale by Mayer and
Davis (1999), with one additional item not included in the original scale. An example item
from the scale is “I really wish I had a good way to keep an eye on _____”. The updated
version had acceptable reliability with coefficient alphas above .70 (Mayer & Gavin, 2005).
The items were measured on a 7-point Likert scale, ranging from strongly disagree to strongly
agree. Alterations were made in the scale introduction to target the manager with whom you
had the closest relationship. This, since it is likely the closest manager which is more
practically involved and taking part in the increase of transparency of an individual
employees’ work. Higher management is less likely to be invested in the work of individuals
but rather focusing on a more strategic level and overall results. Further, Dirks and Ferrin
(2002) have found immediate leaders, for example supervisors, to be particularly important
for trust. Similar alterations to target specific levels of management have been made in other
cases where trust in management has been studied (e.g., Mayer & Davis, 1999). The scale
included two reversed items which were transformed before conducting any analyses.
16
Procedure
No random sampling could be made since all participants in the organization were
needed for maximum power of the study. After an introductory email by one of the change
agents, all registered users of the tool received an email invitation to participate in the study.
The invitation contained information regarding the purpose of the study, who would take part
in the results, and a link to a web-based survey. This information was also included in an
introductory letter in the web-based survey to ensure that participants received the right
information to make an informed decision on whether to participate in the study or not. Where
possible, teams were approached with an additional physical presentation. Although this
might introduce bias in the responses it was deemed necessary due to the high risk of low
power if not enough responses could be collected. The teams were introduced the same
information as in the initial email invite to reduce potential bias. All respondents filled in the
same web-based survey. Two additional email reminders were sent out before closing the data
collection.
Data Analysis
Descriptive analyses were made, presenting the mean scores from the different scales
along with confidence intervals. The reliability of all scales was tested using Cronbach’s
alpha. Since the sample was too small to conduct a reliable factor analysis, reliability of the
transparency scale was conducted including an assessment of correlations between the items
of the scale. To analyze the proposed research question, a multiple regression analysis was
performed, using a forced entry method. Tests and analyses of assumptions of multiple
regression were made according to criteria presented by Field (2009), complete results of
these tests can be found in the Appendix B. As proposed by Field (2009), a final regression
model was presented, only containing predictors that contribute substantially to the prediction
of the outcome. As the last step, the final model was controlled for demographic variables
(age, gender, & tenure), using a hierarchical regression analysis. Demographics were included
in the first block and found predictors in the second, controlling the significance of R2 change
in the second model. Comparisons of parameters and significance of individual predictors
were made.
Ethical Considerations
Conducting a study within an organization deserves some specific considerations.
This, since it is possible results will affect decisions made and if anonymity cannot be ensured
might also affect individuals in the organization negatively. It was therefore especially
important to ensure the anonymity of respondents in the organization, as well as present how
17
the results would be used in the organization. It was expressed in both email invitations,
during physical presentations and in the introductory page of the online survey, that no
individual answers would be shared with the organization and that the results would be used
to guide the change in the organization. The study was in all letters and physical presentations
presented as voluntary since it otherwise could have been a risk that the participants perceived
themselves as obliged to participate due to the close relation to the change work in the
organization.
Results
Descriptives
Averages for the predictors ranged from 4.3 to 6.0, measured on a 7-point Likert scale
(see Table 1 below). Valence had the highest standard deviation of 1.0, which could be
mainly attributed to one question in the scale (whether one perceived to get a higher salary
after the change). This item had an average of 2.6, which equals 1.8 standard deviations below
the average of the scale. However, all original scale items were kept since the scale still had
an acceptable reliability (α = .72).
All scales except one, had an acceptable reliability of a Cronbach’s alpha above .7
(criteria applied from Field, 2009). Trust in management had a rather low reliability (α = .50).
Efficacy had the highest correlation with the outcome of a transparent tool usage (r = .41, p <
.001) followed by principal support (r = .37, p < .001). Appropriateness (r = .20, p = .02) and
valence (r = .18, p = .03) both had small sized correlations with a transparent tool usage.
These significant relationships provided initial support for the hypothesized relationships.
However, discrepancy (r = .02, p = .44) and trust (r = .02, p = .43) did not correlate well with
a transparent tool usage. Trust generally correlated weakly with the other predictors
(correlations below r < .05, p = ns), except for principal support (r = .34, p < .001) and
efficacy (r = .16, p = .05).
Table 1
Summary of Descriptives
Change Beliefs*
Mean Std. Deviation CI
Discrepancy 6.0 0.8 [5.9, 6.2]
Appropriateness 5.7 0.9 [5.5, 5.9]
Valence 4.4 1.0 [4.2, 4.6]
Efficacy 5.5 0.8 [5.3, 5.6]
18
Note. *Measured on a 7-point Likert scale. **Measured on a 5-point Likert scale. CI = Confidence Intervals of
Mean.
Can Change Beliefs and Trust in Management Predict Implementation of Transparency?
The initial model, including all hypothesized predictors, significantly predicted a
transparent tool usage (p < .001). The model had two significant predictors, efficacy (b = .27,
β = .39, t = 2.91, p = .004) and principal support (b = .19, β = .27, t = 2.08, p = .04). 95 %
confidence intervals for the b-values of efficacy (CI = 0.09, 0.46) and principal support (CI =
0.01, 0.37) were broad and might thus not be representative of the true population values, but
still significant and not crossing zero. Except for efficacy and principal support, the other
parameters were close to zero and non-significant. Three highly influential cases were found
(Mahalanobis distance, p < .01). Removing these cases reduced the significance level of
principal support (p = .15), as well as reduced the parameters of the predictor to a beta-value
of .20. Further removing the 10 most influential cases (Mahalanobis distance, p < .05)
reduced the beta-value of principal support to .13 (p = .35). A table of initial model
parameters, without influential cases, is found in the appendix (Appendix C). The power of
the initial parameters and significance of principal support was thus attributed to a few
influential cases rather than a reliable predictive power of the variable.
Table 2
Descriptives for Final Multiple Regression Model
Note. R2
adjusted = 0.16 (p < .001). CI = Confidence Intervals for B.
A final regression model was thus presented, excluding non-significant predictors. As
seen in Table 2, efficacy could significantly predict successful implementation of
transparency (p < .001). A small difference between R2 (.168) and adjusted R2 (.16) indicates
Principal support 5.1 0.8 [5.0, 5.3]
Trust in Management* 4.3 0.9 [4.1, 4.4]
Transparency** 3.4 0.6 [3.3, 3.5]
B 95 % CI SE B β t p <
Constant 1.97 [1.30, 2.64] .34 5.81 .001
Efficacy .29 [0.17, 0.41] .06 .41 4.67 .001
19
the model did not have a poor cross-validity and thus generalizes well. The beta-value
indicated a prediction of medium effect size (β = .41). Efficacy remained a significant
predictor (t = 5.63, p < .001) when controlling for the potential effect of age, gender, and
tenure in a hierarchical regression, nor did the control model reduce the beta-value of efficacy
but rather increased it (β = .51). However, the model including only demographics (age,
gender, & tenure) did not predict any significant variance in transparent tool usage (p = .44)
and none of the control variables were significantly correlated with the outcome.
No troublesome violations of assumptions could be found. In general, the assumptions
of multiple regression were found to have been met. Except for a potential heteroscedastic
variance of residuals of principal support and with a few more cases with higher standardized
residuals than would have been expected in a normal distribution. However, the violations
were minor, and the somewhat troublesome variance of principal support did not affect the
final regression model since this variable was not included. Further details of the controls of
assumptions can be found in the appendix (Appendix B).
Discussion
The results indicate one of the proposed predictors, efficacy, had a significant
correlation with a successful implementation of transparency. Thus, partial support can be
found for the proposed hypothesis. Although initial support could be found for the predictive
power of principal support, this relationship decreased below significant levels, as well as
decreased the model parameters to indicate only a potential predictive power of small effect,
when removing highly influential cases. Due to the ambiguity of the predictive power of this
variable, it was not included in the final model. One reason for this can tentatively be related
to the construction of the scale measuring management support as well as support from
change agents and peers, which might increase the heterogeneity of responses. Another
potential reason can be found in qualitative comments in the questionnaire as well as
discussions with the teams, indicating insecurity about management support.
No indication of a relationship could be found between the three other change beliefs
(discrepancy, appropriateness, & valence), trust in management, and a successful
implementation of transparency. A somewhat surprising result considering the proposed
importance of the variables in the agile change literature (e.g., Barroca et al., 2019; Dikert et
al., 2016; Kalenda et al., 2018; Korhonen, 2013; McHugh et al., 2012). However, this study
intended to measure direct relationships to the intended behavioral outcome of the change
among employees, through a transparent use of the tool implemented to increase
transparency. It could thus be argued that the relationships are rather of an indirect nature
20
such as through readiness for a change (as proposed in Hameed et al., 2019) or related to
resistance to change (as in Abdel-Ghany, 2014; Armenakis et al., 2007). This would then
potentially explain the lack of direct relationships between the variables. Another potential
explanation is how well the change initiative has been shared and understood by change
recipients. It might be that the intentions of implementing the tool have not been properly
understood, leading to a possible high perception of change recipients’ beliefs and trust, while
still scoring low on the intended outcome, simply because there is a lack of understanding or
confusion of the intended outcome of the change and the intended use of the tool for
transparency. Indications of this have been found in discussions with the teams, indicating a
confusion of how to use the tool for some of the stated purposes.
Another potential reason for a lack of relationships can be the measured outcome.
Although the items intended to measure transparent tool usage on an individual level, it can
be argued that individual use of the tool for transparency is dependent on how others use the
tool and share information. Even if the respondent might be willing to use the tool in the
intended way, this might then not be possible if the right information is not shared in the tool.
However, although it could be argued that others’ tool usage is a potentially confounding
variable, it could also be argued that the predictive power of change beliefs as a measurement
of motivation would be related to the perceived benefits of using the tool for transparency.
This would then be perceived as less if not the right information was available in the tool. The
found relationship between efficacy and transparent tool usage would also contradict this.
With perceived high dependence on other parties to successfully implement the change, it
would potentially be less likely for an individual’s efficacy to predict a transparent use of the
tool since this would then rather be dependent on others’ tool usage. Whether a recipient
perceives themselves as capable of successfully implementing the change would then have
less impact on the outcome of the change.
Lastly, the lack of relationship between trust in management and a transparent tool
usage is not surprising considering the low reliability of the trust scale (α = .50). Even though
guidelines of Cronbach’s alphas below 0.70 have been proposed for psychological constructs
that are more challenging to measure (Breakwell, Smith & Wright, 2012; Field, 2009), low
reliability reduces the chance of finding any relationships even though they might exist
(Caruso, 2000). Since the scale had acceptable psychometric properties when developed
(Mayer & Gavin, 2005), the low reliability might rather be attributed to attributes of this study
or potentially the heterogeneous sample, as will be discussed in Limitations.
21
Efficacy was found to be a significant and robust predictor of successful
implementation of transparency, even when controlling for influential cases and demographic
variables. The relationship can be explained by previous findings of a relationship between
change efficacy and resistance versus support for change (e.g., Abdel-Ghany, 2014), as well
as commitment to change (Sundborg, 2019). A high efficacy would indicate the recipient is
more likely to be supportive of and committed to the change and less resistant, as would be
demonstrated in high usage of the tool in the intended way rather than resisting to be a part of
the implementation. Efficacy has also been found to predict technology acceptance and
personal initiative of adopting new technological systems (Brown, 2009), which would be
aligned with a transparent tool usage. The relationship can also be understood from the found
relationship between self-efficacy and motivation (e.g., Latham, 2012), where a higher usage
could be understood from an increase in motivation due to a higher sense of change-efficacy.
Lastly, a reversed causality might also be likely. It would be logical to conclude that an
already high usage of the tool in the intended transparent way would be followed by a
perception of a high capability to succeed in the change, simply because these respondents
already perceive they are managing the change successfully.
No clear violations of assumptions and only a small difference between R2 and
adjusted R2 support the generalizability of the predictive power of efficacy for the success of
agile change initiatives especially targeted at transparency. However, generalizations of
efficacy as a significant predictor of a successful implementation of transparency should still
be made tentatively. This since the results from this study are based on the specific premises
of the change journey made within one organization. The results found in this study should
thus be followed by replication in other contexts to ensure generalizability to other
organizations going through an agile change.
Limitations
Response Rate
The response rate deserves further reflection. A low response-rate might indicate a
biased sample where respondents potentially more invested in the change (positively or
negatively) might be more prone to participate compared to those less interested in the
change. As previously mentioned, 220 employees were initially estimated to use the tool
actively, while responses could only be collected from 118 active users. However, as became
apparent during team presentations as well as comments in the survey, several of the teams
estimated to be using the tool did not perceive themselves to be active users and thus
restrained the sample frame. One work role was lower represented in the sample than would
22
have been estimated. Agile methodologies aim to include business representatives,
stakeholders, more actively in IT operations and development. However only 13 stakeholders
participated in the survey, and of these 4 perceived themselves to be using the tool, thus also
indicating a restriction in the sample frame. However, difficulties in reaching stakeholders in
the study might not be surprising considering previously found difficulties of involving
customers in agile transitions (e.g., Hoda, Noble, & Marshall, 2011).
Heterogeneous Sample
In the studied organization, employees with different roles used the tool being
implemented. This might affect how the tool is used. It might also affect how the change is
perceived since more managerial roles might have different beliefs about the change. A
heterogeneous sample could thus also be a potential explanation of the lack of found
relationships, as well as the low reliability of the scale for trust in management. Although
somehow problematic, it was considered of interest to measure increases of transparency on
all organizational levels involved in the change since measuring only one organizational level
would not give a fair picture of the success of the change. Transparent use of the tool is
needed on both a managerial, team, and individual level. Due to a possible risk to anonymity,
no questions about specific roles were included but instead on a more general organizational
level. Potential differences in usage were taken into consideration when designing the
measurement of transparency. This was done together with the change agents in the
organization to ensure all roles were considered and only measuring parts of the tools which
were intended to be used by all roles. Further, other studies measuring change beliefs have
included a wide range of organizational roles such as administrative workers, technicians,
nurses, and managers (e.g., Mazur et al., 2011).
The teams consist of consultants and directly hired employees. This might affect
responses and response rates. There could also be a higher risk of social desirability since
consultants’ continued work might to a higher extent be dependent on a positive image
towards the firm. Since this risk might have increased if consultants were not allowed to
answer the survey anonymously, this was not controlled for. However, due to a high rate of
consultants in the organization, it was not considered beneficial to exclude these from the
sample. To increase the response rate in both groups, information was directed at both groups,
and participants were ensured of their anonymity in upcoming presentations of the results. It
can still be argued that consultants might, for example, be less involved in a change compared
to employees. However, this confounder might to some extent be reduced by the fact that
23
many consultants had been in the company for years. Furthermore, only respondents who
indicated they had been in the company for at least half a year were included in the analysis.
Future Research
As previously discussed, one aspect worth further investigating is potential indirect
relationships between the proposed predictors and successful change implementation. It could
be that no direct relationship can be found to transparent tool usage due to a rather indirect
relationship through, for example, resistance or support for a change. It could also be of
interest to test relationships with other potential transparency outcomes of a change and not
only a transparent tool usage. This could, for example, be employees’ willingness to share
information and more attitudinal measures as these might give other indications. It would be
interesting to look further into potential relationships, or lack of relationships, between change
support and outcome of an agile change. This since management support has been one of the
more frequently reported key factors for success in the agile literature (e.g., Barroca et al.,
2019; Dikert et al., 2016; Kalenda et al., 2018).
It would have been interesting to analyze transparency on a team level (as has been
done by Palanski et al., 2011). This since how the tool is used might differ between teams
rather than within teams. Unfortunately, this was not possible since some of the teams were
too small to ensure the anonymity of the respondents. However, as has been previously
mentioned a transparent use of the tool might also differ based on differing work roles, thus
making a team aggregation less applicable in this context. And although differences are likely
to exist, all teams had the same agents driving the change and the same tool and agile
framework being implemented since one of the reasons for the proposed change was to
improve collaborations between teams. This would thus potentially increase the similarities
between teams.
Lastly, more research is needed to further study the generalizability of change-efficacy
as a significant predictor of a successful implementation of transparency. No change journey
is like another and this change was mainly driven through the implementation of a tool to
increase the transparency of one’s work. However, similar relationships are not necessarily
found in other contexts and this research thus needs to be followed by other studies to further
investigate the generalizability and predictive power of efficacy in agile change initiatives.
Practical Implications
The results from this study might be used for more practical implications within
organizations going through agile changes. Efficacy was in this study found to be
significantly related to a successful implementation of transparency. This might therefore be
24
used to guide practitioners in their agile change initiatives to target efficacy in their
interventions. It could also be interpreted as support for the previously found importance of
related factors such as training, coaching and knowledge during agile changes (Asnawi et al.,
2012; Dikert et al., 2016; Mohan, 2018). However, this study also indicates that some caution
might be needed before applying results from previous research to develop a change strategy.
This since there is a need for more quantitative support to further investigate the
generalizability of success factors described in more qualitative research.
Conclusion
This study indicates a lack of quantitative support for several of the previously cited
success factors in the agile change literature. A need for more quantitative and research-
driven literature can therefore be noted to balance and further investigate quantitative support
for previously reported themes and variables in qualitative literature and case studies in the
agile literature. No support could be found for a relationship between discrepancy,
appropriateness, valence, principal support, trust in management, and successful
implementation of transparency. However, the results indicate efficacy is a significant and
robust predictor of a successful implementation of transparency. However, more research is
still needed to ensure the generalizability of efficacy as a predictor of the outcome of an agile
change initiative. Hence, this study could only find partial support for the applicability of
psychological change theories to predict the outcome of an agile change.
25
References
Abdel-Ghany, M. M. M. (2014). Readiness for change, change beliefs and resistance to
change of extension personnel in the New Valley Governorate about mobile
extension. Annals of agricultural Sciences, 59(2), 297-303.
Armenakis, A. A., Bernerth, J. B., Pitts, J. P., & Walker, H. J. (2007). Organizational change
recipients' beliefs scale: Development of an assessment instrument. The Journal of
applied behavioral science, 43(4), 481-505.
Armenakis, A. A., & Harris, S. G. (2002). Crafting a change message to create
transformational readiness. Journal of organizational change management, 15(2),
169-183.
Armenakis, A. A., & Harris, S. G. (2009). Reflections: Our journey in organizational change
research and practice. Journal of change management, 9(2), 127-142.
Asnawi, A. L., Gravell, A. M., & Wills, G. B. (2012). Emergence of agile methods:
Perceptions from software practitioners in Malaysia. In Guerrero, J. E. (Eds.),
AgileIndia 2012 (pp. 30-39). The Institute of Electrical and Electronics Engineers.
Bandura, A. (1997). Self-efficacy: the exercise of control. W. H. Freeman.
Bang, T. J. (2007). Introducing agile methods into a project organisation. In G. Concas, E.
Damiani, M. Scotto, & G. Succi (Eds.), Agile Processes in Software Engineering and
Extreme Programming (pp. 203-207). Berlin: Springer.
Barroca, L., Sharp, H., Dingsøyr, T., Gregory, P., Taylor, K., & AlQaisi, R. (2019). Enterprise
agility: A Balancing Act-a local government case study. In P. Kruchten, S. Fraser, &
F. Coallier (Eds.), Agile Processes in Software Engineering and Extreme
Programming (pp. 207-223). Cham: Springer.
Bayraktar, S., & Jiménez, A. (2020). Self-efficacy as a resource: a moderated mediation
model of transformational leadership, extent of change and reactions to change.
Journal of Organizational Change Management.
Beck, K., Beedle, M., Bennekum, A., Cockburn, A., Cunningham, W., Fowler, M., …
Thomas, D. (2001). Manifesto for Agile Software Development. Retreived from
http://agilemanifesto.org/
Björkholm, T. & Brattberg, H. (2016). Prioritera, fokusera, leverera: din snabbguide till
Lean, Agile, Scrum, Kanban och XP: version 1.4. Stockholm: Vulkan.
Breakwell, Glynis, M., Smith, Jonathan, A., & Wright, Daniel, B. (2012). Research methods
in psychology. London: SAGE Publications.
26
Brown, S. (2009). Technology acceptance and organizational change: An integration of
theory (Doctoral thesis). Auburn, Alabama: Auburn University. Retrieved from
http://etd.auburn.edu/handle/10415/1991
Bruckman, J. C. (2008). Overcoming resistance to change: Causal factors, interventions, and
critical values. The Psychologist-Manager Journal, 11(2), 211-219.
Burnes, B. (2020). The Origins of Lewin’s Three-Step Model of Change. The Journal of
Applied Behavioral Science, 56(1), 32-59.
Caruso, J. C. (2000). Reliability generalization of the NEO personality scales. Educational
and Psychological Measurement, 60(2), 236-254.
Chreim, S. (2006). Managerial Frames and Institutional Discourses of Change: Employee
Appropriation and Resistance. Organization Studies, 27(9), 1261–1287.
https://doi.org/10.1177/0170840606064106
CollabNet VersionOne. (2019). 13th Annual State Of Agile Report. Retrieved from
https://www.stateofagile.com/#ufh-i-521251909-13th-annual-state-of-agile-
report/473508
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.
Cohn, M. (2010). ADAPTing to Agile for Continued Success. Retrieved from
https://www.mountaingoatsoftware.com/presentations/adapting-to-agile-for-
continued-success
Conboy, K. (2009). Agility from first principles: Reconstructing the concept of agility in
information systems development. Information systems research, 20(3), 329-354.
Dikert, K., Paasivaara, M., & Lassenius, C. (2016). Challenges and success factors for large-
scale agile transformations: A systematic literature review. Journal of Systems and
Software, 119, 87-108.
Dingsøyr, T., Dybå, T., Gjertsen, M., Jacobsen, A. O., Mathisen, T. E., Nordfjord, J. O., Røe,
K., & Strand, K. (2019). Key lessons from tailoring agile methods for large-scale
software development. IT Professional, 21(1), 34-41.
Dirks, K. T., & Ferrin, D. L. (2002). Trust in leadership: Meta-analytic findings and
implications for research and practice. Journal of applied psychology, 87(4), 611-628.
Doyle, O., McNamara, K., & Logue C. (2009). Readiness for Change: Baseline Data on
Readiness to Implement the Síolta Framework. UCD Geary Institute.
Ertürk, A. (2008). A trust-based approach to promote employees' openness to organizational
change in Turkey. International Journal of Manpower, 29(5), 462-483.
Field, A. (2009). Discovering statistics using SPSS. London: SAGE.
27
Figalist, I., Elsner, C., Bosch, J., & Olsson, H. H. (2019). Scaling agile beyond organizational
boundaries: coordination challenges in software ecosystems. In P. Kruchten, S. Fraser,
& F. Coallier (Eds.), Agile Processes in Software Engineering and Extreme
Programming (pp. 189-206). Cham: Springer.
Gandomani, T. J., & Nafchi, M. Z. (2016). The essential prerequisites of Agile transition and
adoption: A grounded theory approach. Journal of Internet Computing and Services,
17(5), 173-184.
Giangreco, A., & Peccei, R. (2005). The nature and antecedents of middle manager resistance
to change: Evidence from an Italian context. The international journal of human
resource management, 16(10), 1812-1829.
Gregory, P., Barroca, L., Sharp, H., Deshpande, A., & Taylor, K. (2016). The challenges that
challenge: Engaging with agile practitioners’ concerns. Information and Software
Technology, 77, 92-104.
Hameed, I., Khan, A. K., Sabharwal, M., Arain, G. A., & Hameed, I. (2019). Managing
successful change efforts in the public sector: An employee’s readiness for change
perspective. Review of Public Personnel Administration, 39(3), 398-421.
Hoda, R., Noble, J., & Marshall, S. (2011). The impact of inadequate customer collaboration
on self-organizing Agile teams. Information and Software Technology, 53(5), 521-
534.
Horlach, B., Schirmer, I., Böhmann, T., & Drews, P. (2018). Agile portfolio management
patterns: a research design. In S. Wagner, R. Hoda, & A. Aguiar (Eds.), Proceedings
of the 19th International Conference on Agile Software Development: Companion.
New York: Association for Computing Machinery.
Iivari, J., & Iivari, N. (2011). The relationship between organizational culture and the
deployment of agile methods. Information and software technology, 53(5), 509-520.
Kalenda, M., Hyna, P., & Rossi, B. (2018). Scaling agile in large organizations: Practices,
challenges, and success factors. Journal of Software: Evolution and Process, 30(10),
1-24.
Korhonen, K. (2013). Evaluating the impact of an agile transformation: a longitudinal case
study in a distributed context. Software Quality Journal, 21(4), 599-624.
Kotter, J. P. (2012). Leading Change. Harvard Business Review Press.
Laanti, M., & Kettunen, P. (2019). SAFe adoptions in Finland: a survey research. In R. Hoda
(Eds.), Agile Processes in Software Engineering and Extreme Programming -
Workshops (pp. 81-87). Cham: Springer.
28
Laanti, M., Salo, O., & Abrahamsson, P. (2011). Agile methods rapidly replacing traditional
methods at Nokia: A survey of opinions on agile transformation. Information and
Software Technology, 53(3), 276-290.
Latham, G.P. (2012). Work motivation: history, theory, research, and practice. (2nd ed.)
SAGE Publications, Inc.: California.
Long, K., & Starr, D. (2008). Agile Supports Improved Culture and Quality for Healthwise. In
G. Melnik, P. Kruchten, & M. Poppendieck (Eds.), Agile2008 Conference (pp. 160-
165). The Institute of Electrical and Electronics Engineers.
Lunenburg, F. C. (2010). Forces for and resistance to organizational change. National Forum
of Educational Administration and Supervision Journal. 27(4), 1-10.
Mayer, R. C., & Davis, J. H. (1999). The effect of the performance appraisal system on trust
for management: A field quasi-experiment. Journal of applied psychology, 84(1), 123-
136.
Mayer, R. C., & Gavin, M. B. (2005). Trust in management and performance: who minds the
shop while the employees watch the boss?. Academy of management journal, 48(5),
874-888.
Meyer, J. P., Hamilton, L. K. (2013). Commitment to organizational change: theory, research,
principles, and practice. In Oreg, S., Michel, A., Todnem, R. (Eds.), The Psychology
of Organizational Change: Viewing Change from the Employee’s Perspective (pp. 66-
91). Cambridge: Cambridge University Press.
Mazur, L. M., Rothenberg, L., & McCreery, J. K. (2011). Measuring and Understanding
Change Recipients' Buy-in during Lean Program Implementation Efforts. In T.
Doolen, & E. Van Aken (Eds.), Proceedings of the 2011 Industrial Engineering
Research Conference. Institute of Industrial and Systems Engineers (IISE).
McHugh, O., Conboy, K., & Lang, M. (2012). Agile practices: The impact on trust in
software project teams. Ieee Software, 29(3), 71-76.
Mohan, A. (2018). Agile Project Management Challenges: Analyzing and Exploring Agile
Project Management Challenges from a Practitioner Perspective: A case study on
HMS (Master’s Thesis). Halmstad: Industrial Innovation Management, Halmstad
University.
Murphy, P., & Donnellan, B. (2009). Lesson Learnt from an Agile Implementation Project. In
P. Abrahamsson, M. Marchesi, & F. Maurer (Eds.). Agile Processes in Software
Engineering and Extreme Programming (pp. 136-141). Berlin: Springer.
29
Oprins, R. J., Frijns, H. A., & Stettina, C. J. (2019). Evolution of Scrum Transcending
Business Domains and the Future of Agile Project Management. In P. Kruchten, S.
Fraser, & F. Coallier (Eds.), Agile Processes in Software Engineering and Extreme
Programming (pp. 244-259). Cham: Springer.
Oreg, S. (2006). Personality, context, and resistance to organizational change. European
journal of work and organizational psychology, 15(1), 73-101.
Oreg, S., Michel, A., & By, R. T. (Eds.). (2013). The psychology of organizational change:
Viewing change from the employee’s perspective. Cambridge University Press.
Palanski, M. E., Kahai, S. S., & Yammarino, F. J. (2011). Team virtues and performance: An
examination of transparency, behavioral integrity, and trust. Journal of Business
Ethics, 99(2), 201-216.
Pedrycz, W., Russo, B., & Succi, G. (2011). A model of job satisfaction for collaborative
development processes. The Journal Of Systems & Software, 84, 739-752.
doi:10.1016/j.jss.2010.12.018
Pries-Heje, J., & Baskerville, R. (2017). The translation and adaptation of agile methods: a
discourse of fragmentation and articulation. Information Technology & People, 30(2),
396-423.
Rafferty, A. E., & Griffin, M. A. (2006). Perceptions of organizational change: A stress and
coping perspective. Journal of applied psychology, 91(5), 1154-1162.
Rousseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Not so different after all: A
cross-discipline view of trust. Academy of management review, 23(3), 393-404.
Scaled Agile. (n.d.). SAFe® for Lean Enterprises. Retrieved from
https://www.scaledagileframework.com/
Scaled Agile. (2019a). Team and Technical Agility. Retrieved from
https://www.scaledagileframework.com/team-and-technical-agility/
Scaled Agile. (2019b). Core Values. Retrieved from
https://www.scaledagileframework.com/safe-core-values/
Scaled Agile. (2020). Lean-Agile mindset. Retrieved from
https://www.scaledagileframework.com/lean-agile-mindset/
Schnitter, J., & Mackert, O. (2011). Large-scale agile software development at SAP AG. In L.
A. Maciaszek & P. Loucopoulos (Eds.), Evaluation of Novel Approaches to Software
Engineering (pp. 209-220). Berlin: Springer.
Schwaber, K., & Sutherland, J. (2017). The Scrum Guide: The Definitive Guide to Scrum: The
Rules of the Game. Retrieved from https://www.scrum.org/resources/scrum-guide
30
Seger, T., Hazzan, O., & Bar-Nahor, R. (2008). Agile orientation and psychological needs,
self-efficacy, and perceived support: a two job-level comparison. In G. Melnik, P.
Kruchten, & M. Poppendieck (Eds.), Agile2008 Conference (pp. 3-14). The Institute
of Electrical and Electronics Engineers.
Sidky, A. S. (2007). A structured approach to adopting agile practices: The agile adoption
framework (Doctoral thesis). Virginia: Faculty of the Virginia Polytechnic Institute
and State University.
So, C., & Scholl, W. (2009). Perceptive agile measurement: New instruments for quantitative
studies in the pursuit of the social-psychological effect of agile practices. In P.
Abrahamsson, M. Marchesi, & F. Maurer (Eds.). Agile Processes in Software
Engineering and Extreme Programming (pp. 83-93). Berlin: Springer.
Stajkovic, A. D., & Luthans, F. (1998). Social Cognitive Theory and Self-Efficacy: Going
Beyond Traditional Motivational and Behavioral Approaches. Organizational
Dynamics, 26(4), 62-74.
Stettina, C. J., & Hörz, J. (2015). Agile portfolio management: An empirical perspective on
the practice in use. International Journal of Project Management, 33(1), 140-152.
Sundborg, S. A. (2019). Knowledge, principal support, self-efficacy, and beliefs predict
commitment to trauma-informed care. Psychological Trauma: Theory, Research,
Practice, and Policy, 11(2), 224-231.
Suomalainen, T., Kuusela, R., Teppola, S., & Huomo, T. (2015). Challenges of ICT
companies in lean transformation. Advances in Computer Science: an International
Journal, 4(2), 49-56.
Tabachnick, B. G., Fidell, L. S., & Ullman, J. B. (2007). Using multivariate statistics (5th ed.).
Boston, MA: Pearson.
Tessem, B., & Maurer, F. (2007). Job satisfaction and motivation in a large agile team. In G.
Concas, E. Damiani, M. Scotto, & G. Succi (Eds.), Agile Processes in Software
Engineering and Extreme Programming (pp. 54-61). Berlin: Springer.
Totten, J. (2017). Critical Success Factors for Agile Project Management in Non-Software
Related Product Development Teams (Doctoral thesis). Michigan: Western Michigan
University. Retrieved from https://scholarworks.wmich.edu/dissertations/3178
Tripp, J. F. (2012). The impacts of agile development methodology use on project success: A
contingency view (Doctoral thesis). Michigan: Michigan State University. Retrieved
from
https://d.lib.msu.edu/islandora/object/etd:1068/datastream/OBJ/download/The_impact
31
s_of_agile_development_methodology_use_on_project_success__a_contingency_vie
w.pdf
Tripp, J. F., Riemenschneider, C., & Thatcher, J. B. (2016). Job Satisfaction in Agile
Development Teams: Agile Development as Work Redesign. Journal of the
Association for Information Systems, 17(4), 267-307.
Turetken, O., Stojanov, I., & Trienekens, J. J. (2017). Assessing the adoption level of scaled
agile development: a maturity model for Scaled Agile Framework. Journal of
Software: Evolution and process, 29(6), 1-18.
van Dam, K. (2005). Employee attitudes toward job changes: An application and extension of
Rusbult and Farrell's investment model. Journal of Occupational and Organizational
Psychology, 78(2), 253-272.
Van den Broeck, A., Ferris, D. L., Chang, C. H., & Rosen, C. C. (2016). A review of self-
determination theory’s basic psychological needs at work. Journal of Management,
42(5), 1195-1229.
van Manen, H., & van Vliet, H. (2014). Organization-wide agile expansion requires an
organization-wide agile mindset. In A. Jedlitschka, P. Kuvaja, M. Kuhrmann, T.
Männistö, J. Münch, M. Raatikainen (Eds.), Product-Focused Software Process
Improvement (pp. 48-62). Cham: Springer.
32
Appendix A – Transparency Scale
I use Azure DevOps Services to easily...
TA1 ...see WHAT needs to be done
TA2 ...see WHEN things need to be done
TA3 ...see the CURRENT status of things
TA4* ...find information of HOW things are solved
TA5 ...see WHOM I can contact when I need more information
TA6 ...see how my work RELATES to the work of others
Note. *Excluded due to low correlations with TA 1 (r = 0.11), TA3 (r = 0.19) and TA6 (r = 0.24). Reliability of
initial 6-item scale: α = 0.72. Reliability of 5-item scale: α = 0.71. Items were measured on a 5-point Likert scale
ranging from strongly disagree to strongly agree.
33
Appendix B - Tests of Assumptions, Multiple Regression
In general, the assumptions of a multiple regression can be argued to have been met.
With a potential heteroscedastic variance of residuals in principal support and with a few
more cases with higher standardized residuals than would have been expected in a normal
distribution. However, the violations were minor, and the somewhat troublesome variance of
principal support did not affect the final regression model.
The assumption of an unbounded outcome variable seems to have been met since the
outcome variable varied from 1.8-5 (on a scale from 1-5). There was variance in the variables.
No multicollinearity could be found with predictors highly correlated with one another (r >
.9). No VIF value was over 10 (initial model: average VIF = 1.95), no tolerance value was
below 0.2 and no predictors were loading highly on the same small eigenvalues.
Although impossible to test all external variables, significant predictors were checked
for correlation with potentially influencing demographics (age, tenure, & gender), especially
since a skewness could be found in the demographic variables. No correlations were found.
Visually controlling for homoscedasticity, no clear violations were found in partial plots or
residual plots, except for a potentially heteroscedastic variance of residuals in principal
support, forming a weak s-shape. This could potentially be attributed to the nature of the scale
measuring support from both managers (top & immediate) and peers, which do not
necessarily have to be perceived in the same way. However, this predictor had other reasons
to not be included in the final analysis. Except for potentially principal support, no violations
of linearity could be seen by inspecting partial plots.
With a Durbin-Watson value close to 2 (2.05, initial model, 2.13 final model) the
assumption of independent errors can be argued to have been met. Residuals seemed to be
normally distributed around a mean of zero. Normal Probability Plot of residuals looked OK
and followed the line quite well. A histogram of residuals seemed normally distributed. Less
than 5 % of the cases had a standardized residual of more than +/- 2 in the initial model.
However, 3 cases had a standardized residual of more than +/- 2.5, which is more than the
expected 1 % (= 1.1 cases). However, all had a standardized residual within +/- 3 and one of
the three cases was just above the threshold (2.55). In the final model, 6.4 % of the cases (7
cases) had a standardized residual of more than +/- 2. However, only one case (< 1 %) had a
standardized residual of more than +/- 2.5.
34
Appendix C – Initial Regression Model
Table
Descriptives of Initial Multiple Regression Model
B 95 % CI SE B β t p
Constant 2.23 [1.17, 3.28] .53 4.19 .000
Valence .04 [-.12, .21] .08 .07 .51 .61
Appropriateness -.06 [-.31, .19] .13 -.09 -.5 .62
Efficacy .33 [.09, .56] .12 .41 2.76 .01
Discrepancy -.08 [-.26, .09] .09 -.11 -.94 .35
Principal support .10 [-.11, .31] .11 .13 .94 .35
Trust in Management -.08 [-.21, .06] .07 -.11 -1.08 .28
Note. Excluding the 10 most influential cases (Mahalanobis distance, p < .05) since these had a high influence on
model parameters and significance levels. CI = Confidence Intervals for B.