TEMPORAL AND SPATIAL VARIABILITY OF THE DETERMINANTS OF SATISFACTION WITH PUBLIC TRANSPORT IN SWEDEN

MASTER THESIS

TRANSPORT AND GEOINFORMATION

TECHNOLOGIES

FERNÁNDEZ ABENOZA, Roberto

Supervisors: Yusak O. Susilo

Oded Cats Division of Transport and Location analysisTransport Science DepartmentRoyal Institute of Technology, Stockholm

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

BackgroundWorld trends: Increase population, larger need of mobility and accessibility, longer trips, road safety, environmental problems, etc. PTx2 Strategy (2009-2025 increase use transport by 100%)

Europe: EC White Paper 2011 sets Tranport sector goals by 2050

Sweden issues: Need for socially and environmental sustainable society (transport act 2009); Swedish doubling project (PT modal share up to 24% by 2020); Increase of contribution/subsidies to landsting

Increasing Satisfaction with Public Transport (PT) service very important to increase ridership.

Important to razionalize costs while maximising it benefits.

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Customer satisfaction• Overall fullfilment with customer’s expectations (Tyrinopoulos and Antoniou,

2008)

• Outcome of cumulative satisfaction through time or single-experiences, or the two. Negative Single encounters: NCI (Friman et al, 2001; Friman and Fellesson, 2009; Redman et al, 2013)

• Increase of customer satisfaction leads to a growth in customer loyalty (Barabino et al, 2012)

Adaptation of ECSI customer satisfaction model

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS D.STATS

C. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Purpose

Which are the most important QoSA that contribute to customer satisfaction with PT services in Sweden?

andAre they consistent over time and space?

- Overview of drivers of satisfaction with PT services focusing on supply characteristics of the service

- Knowledge whether the determinants of satisfaction remain unchanged over time. Detect future trends and stability/variability of certain determinants (QoSA)

- Insight into the uniformity or nonuniformity of the determinants across geographical areas

- Ascertain and Identify priority areas for different regions and years, in terms of importance, current performance and trend.

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Groups interested in outcome

Why is it important?

It is of capital importance for service performance monitoring, market analysis, benchmarking and the identification of priority areas

For whom?

Study is instrumental in supporting PT providers, Regional PT authorities as well as County councils and Municipalities to provide a PT service to match provision with their user’s needs and thus foster customer satisfaction and ultimately increase PT ridership

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Scope and limitation

LIMITATIONS

A. Focus on PT users users keeping them happy and increasing ridership . No investigation of only car users

B. Variability over space limited to 5 different county regions levels.

Geographical Scope

Time frame = 2001-2013

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Work-flow

Diagram

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS D.STATS

C. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Quality of Service Attributes (QoSA)

SERVICE QUALITY

AREAS QUALITY OF SERVICE ATTRIBUTES

Time Length of Trip time Adherence to schedule

Information General information Information on planned

changes

Information on unplanned

changes Accessibility Accessibility Ticket Easiness of travel

Availability Operation Network Availability

Network

Comfort On-board

conditions Ride

Comfort

Ambient off-board

conditions

Use of Travel Time

Traffic Stress

avoidance Customer care Staff and assistance Customer interface

Security Freedom from crime Environmental

impact Environmentally friendly

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Application of Weights

Map of Sweden per County

Cartogram regarding Population per county

in 2001

Cartogram regarding sample size from dataset in 2001

Weighted by county, urban area (Tätorter) and age

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Socio-demographic & mobility profile of the weighted sample

Gender

Age groups

Occupation

Disability (2001,2007-2013)

Distance work/school

Urban Area size(in thousands)

Density Population(hab/km2)

Car availability

Frequency travel car(2010-2013)

Frequency travel by PT

Most used PT mode(2011-2013)

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Male Female

15-24 25-34 35-44 45-54 55-64 65-75

Retired / Sick leave Student

Yes No

<1km 1.0 - 3.0 km 3.1 - 5.0 km 5.1 - 10.0 km 10.1 - 30.0 km 30.1 - 50.0 km

>50km

Stockholm-Göteborg-Malmö 100-150

75-100

50-75 <50

<10 10,1-21 21,1-41 41,1-56 51,1-116 >116

Yes No

Daily Weekly Monthly More rarely Never

Daily Weekly Monthly More rarely

City bus Regional bus Commuter train Metro Regional

trainTram

Employed / Self-employed / Employee

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE.STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Overall Satisfaction & QoSA over time

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Cross-correlations• Intercorrelations between QoSA and Overall Satisfaction all positive: Halo

Effect• Highest intercorrelation between Quality and Popularity and Overall

Satisfaction. Explains why higher Overall Satisfaction in Medium/Small counties.

• Strong intercorrelation between Frequency of travel by PT and Relevance implies that a useful and functional PT service favours to stimulate ridership.

• Very high intercorrelation between Loyalty and Overall Satisfaction in consonance with that high OS leads to higher loyalty

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONS

FACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Factor Analysis (PCA)

The component solution explains a total of 57.3% of the variance.

Soft component contributing a 37.5% and Functional and Information components explaining 11.3% and 8.5%, respectively.

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE.STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Models specification

MODELS

Explanatory variables included Time and Geographical frame

3 factors

from FA

12 QoSA

3 Frequency

PT dummy

12 Year

dummy

4 County dummy

Entire database

Per yearPer

County region

Joint Factors x   x x x x    

Joint QoSA   x x x x x    

Year-specific   x x   x   x  

Region-specific   x x x       x

• None of the socio-demographic and travel habits variable systematically significant over time and space. Low explanatory power, thus removed

• No interaction variables were found

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

S.C. t-stats S.C. t-stats

(Constant) ,474 42,012 3,544

Soft Factor ,324 178,455

Ride comfort ,020 10,855

On-board conditions ,044 23,515

Staff and assistance ,083 45,277

Freedom from crime ,136 70,470

Functional Factor ,304 163,500

Network ,099 50,171

Operation ,165 84,950

Length of Trip time ,156 79,551

Information Factor ,278 148,678

Accessibility Ticket ,043 25,078

Information on planned changes

,044 22,206

Information on unplanned changes

,066 34,006

Customer interface ,183 99,119

General information ,047 27,758

Daily ,042 22,342 ,035

One Few time week ,054 29,563 ,059

OneFew time month ,035 20,085 ,048

Västra Götalands ,024 13,284 ,018 9,639

Skåne -,007 -3,485

Medium size ,016 8,04 ,006 3,008

Small size ,034 17,063 ,022 11,115

2002 0,005* 2,048

2003 ,007 2,723 ,010 3,478

2004 ,016 5,814 ,017 6,163

2005 ,011 4,172 ,011 3,798

2006 ,008 2,903 ,010 3,494

2007 ,016 5,232 0,006* 2,003

2008 ,018 6,039 0,007* 2,323

2009 ,017 5,701 0,005** 1,742

2010 ,015 4,939 0,008* 2,347

2011 (-)0,006** -1,887

2012 -,013 -4,045 -,023 -6,756

2013 -,014 -4,250

R-square

Adjusted R-square

Number of observations 397861 397861

Joint QoSA Joint Factors

,502 ,505,502 ,505

Joint ModelsJoint Factors Model

Factors have similar influenceSequence: Soft, Functional, Information

Joint QoSA Model

Customer interface

Operation

Length of trip time

Freedom from crime

Ride Comfort

Joint contribution of QoSA into factors.

Largest differences. Sequence: Functional, Information, Soft

* Sig. 0.05 ** Sig. 0.1 Otherwise Sig. 0.01

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Priority map per Year (2001-2013)

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22

0.70

0.75

0.80

0.85

0.90

0.95

1.00

1.05

1.10

1.15

1.20

1.25

1.19213236263961

1.0286275728068

0.9331786418753030.923416448979771

1.05503612548912

1.09216371118337

1.06885907068224

0.982606055600007

1.11281152974792

0.899928488940797

0.780314245497073

0.830556137558788

1.19616939840976

1.04649714733439

0.9378109514248190.922581683797955

1.05190567270466

1.09240909521426

1.06610943940248

0.986073363245494

1.11366574224464

0.891296441316355

0.758909224570582

0.832671460207872

1.188714938239

1.03680475213205

0.9299340758978690.920372313516002

1.0391093128259

1.088859526192041.06194869606625

0.989672429859854

1.11678514424096

0.903876470943654

0.76669863352118

0.848590958554591

1.1731082110579

1.03468743179023

0.9212663115619480.909632534906267

1.04539827904512

1.09176244337643

1.06348716340652

0.98655124280051

1.10730582482253

0.927637852232024

0.782487479315381

0.852386325032001

1.159440637599

1.06430430901602

0.9134698003523110.897016994166968

1.0543610155894 1.09393584832375

1.06123167459492

0.977759319459731

1.097746027727

0.93563697285288

0.783482363928668

0.852941692097737

1.16201260507429

1.09188605219324

0.9091586962845720.897905691534066

1.05381004010091

1.09185974578491

1.06069556720771

0.978196305073195

1.09006004619455

0.938890261510303

0.776576471973253

0.859209721327449

1.15021905266384

1.12528800799704

0.903424507869520.893072979867936

1.04939726525025

1.09487988289315

1.05593737454196

0.973486720133154

1.08643658063261

0.93744212512532

0.782115039575069

0.85444429459576

1.105323528176011.11600542637411

0.9209957413058150.932539963749078

1.03581603547094

1.08026582866986

1.04222991830668

0.996010619998903

1.08692684267162

0.932656573742018

0.789739182627867

0.852078415220661

1.11095963937683

1.13754319527372

0.9240781436634250.936296886127767

1.03396489878186

1.07790134189942

1.037156549893230.996377143812037

1.09156327459066

0.927970713966176

0.77645059465367

0.840506211116248

1.097

1.137

0.9310.946

1.044

1.080

1.040

0.994

1.097

0.923

0.766

0.839

1.078

1.145

0.9300.948

1.052

1.083

1.044

0.994

1.100

0.921

0.758

0.837

1.064

1.141

0.931

0.952

1.049

1.081

1.040

0.995

1.097

0.931

0.774

0.843

1.062

1.145

0.923

0.953

1.046

1.080

1.051

1.000

1.101

0.933

0.768

0.835

Customer's Priorities

Sati

sfac

tion

inde

x: 1

Ave

rage

Sas

tisf

acti

on S

core

=3,5

2Information on unplanned changes

Customer interface

Freedom from crime

Information on planned changes

Network

General informa-tion

Staff and assistance

Operation

Length of trip time

Ride Comfort

Accessibility ticket

On-board conditions

Possible overkill

Lower Priority

PRIORITY AREA

Keep up the good

work

1 colour per Quality of Service Attribute. Ex. Red for Network 1=2001; 2=2002…13=2013

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22

0.70

0.75

0.80

0.85

0.90

0.95

1.00

1.05

1.10

1.15

1.20

1.25

1.177

1.141

1.035

0.979

0.970

1.065

1.029 0.993

1.065

Travel Information on planned changes

Travel Information on unplanned changes

0.804

Stockholms Västra Götalands Skåne Medium size Small size

Customer's Priorities

Satis

facti

on in

dex:

1

Aver

age

Sasti

sfac

tion

Scor

e=3,

53

PRIORITY AREA

Possible overkill

Lower Priority

Keep up the good

work

Priority map per County-region

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

ConclusionsDeterminants of satisfaction with PT services in Sweden : Customer Interface, Freedom from Crime, Length of Trip Time, Operation, Network

Do the determinants of satisfaction remain unchanged over time?

Overall, the importance of the determinants of Satisfaction with PT service varies over time as can be observed by the horizontal movements of the QoSA The QoSA perdure in the same quadrants.

Do the determinants remain unchanged across regions in Sweden?

Yes, in general. Some QoSA that behave differently depending on the county-region, likewise Stockholm and Skåne for Network and Length of Trip Time.

How can service providers identify priority areas? Priority areas can be identified by adopting priority map which identify importance-priority areas.

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

RecommendationsGeneral• Customer satisfaction surveys: Maintain same set of QoSA over time

Increase key QoSAPrioritization of Customer Interface (Below satisfaction threshold)

Length of trip time (Not in Skåne)

Operation and Network (Not in Stockholm)

Keep security (Freedom from crime)

Watch Information on Planned changes

Information on Unplanned Changes. Might be dangerous.

Ticket Accessibility: Decreasing Satisfaction. Study why.

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

QoSATransport Operators

PT Regulators

Customer Interface   XX

Operation X X

Length of trip time X X

Network X X

Freedom from Crime X X

Staff and Assistance XX  

Information on unplanned changes XX  

Information on Planned Changes XX  

On-board Conditions XX  

Ticket Accessibility   XX

General Information XX  

Ride Comfort XX  

Degree of responsibility with the increase in satisfaction of the QoSA

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Future research

1. Investigation of different geographical areas:- 25 Swedish Concession areas- Generate another groups of homogeneus regions

2. Inclusion of other QoSA - Application of a shorter time frame:

3. Investigation of determinants for different target groups

4. Analysis of seasonal variations (only available from 2010)

5. Analysis of Panel data (data not available)

PROBLEM DESCRIPTION

OBJECTIVE

SCOPE AND LIMITATION

METHODOLOGY

RESULTS DESCRIPTIVE STATSC. CORRELATIONSFACTOR ANALYSISREGRESSION M.

PRIORITY MAPS

CONCLUSIONS & RECOMMENDATIONS

Thanks for your attention!

Any questions?

rfa@kth.se