The study seeks to assess the influence of perceived justice, service failure attribution, and recovery disconfirmation on recovery satisfaction among subscribers of mobile money transfer services in Kenya. The specific objectives of the study are to determine the influence of perceived justice on recovery satisfaction; assess the mediating effect of recovery disconfirmation on the relationship between perceived justice and recovery satisfaction; establish the moderating influence of service failure attribution and to determine the joint effect of the three variables on recovery satisfaction.
Customer satisfaction is crucial in maintaining a mutually beneficial relationship between service providers and their customers. While service organizations make great efforts to ensure customer satisfaction, unique features of services such as variability and intangibility make service failure inevitable. Service failure comprises any real or perceived problem that occurs during a customer’s interaction with the service provider. It may include unavailability, poor delivery, inconsistent outcomes or any incident where a service fails to meet customer expectations.
The negative disconfirmation following a service failure leads to dissatisfaction with the likely consequence of increased customer complaints, negative word-of-mouth and erosion of patronage of the service provider. Organizations strive to develop and implement service recovery strategies in order to restore satisfaction. Customers evaluate service recovery from a perspective fairness of process, interaction and outcome in line with the dimensions of perceived justice.
The population of the study comprised mobile money transfer service subscribers in Kenya. A descriptive cross-sectional survey is used. Primary data is collected using semi-structured questionnaires and a final sample of 803 respondents was realized. Reliability and validity tests are conducted using data from a pilot study. Data is analysed using descriptive statistics, factor analysis, correlation, and regression analysis. The study is relevant for regulators in setting policy for service providers with regard to service recovery standards and redress systems. The results will serve as a point of reference for service providers and managers in assessing recovery satisfaction for effective strategic marketing decision-making.
ii
DECLARATION
ThisDoctoralThesison:
Perceived Justice, Service Failure Attribution, Disconfirmation and
Recovery Satisfaction among Subscribers of Mobile Money Transfer
Services in Kenya
By
Catherine Ngahu
Wasapprovedin2016
Bythe School of Business
Department of Business Administration
University of Nairobi
Kenya
This Doctoral Thesis was approved on 4.11.2016
University Supervisors:
Professor Francis N. Kibera
Professor Peter K'Obonyo
iii
DEDICATION
This Thesis is dedicated to my late Dad, Peter Ngahu, who totally believed in the
education of girls; my Mum Mary, whose love of education continues to inspire me;
and my son, who has been a shining example of dedication.
iv
ACKNOWLEDGMENT
Writing this thesis was a demanding and stimulating experience. I thank God for giving
me the energy, focus and ability to work on this thesis to completion. There are several
people I want to acknowledge for the contribution and support they accorded me during
this process. My profound gratitude goes to my University supervisors Professor Francis
N. Kibera and Professor Peter Kobonyo I cannot thank you enough for the great guidance,
support and encouragement throughout the time I worked on this thesis. I am immensely
grateful to Professor Kibera for his responsiveness, assistance in developing the conceptual
ideas of this thesis and for progressively teaching me to focus on the bigger picture without
losing sight of the small details. From Professor Kobonyo, I learned the importance staying
focused and revisiting my initial assumptions.
I want to thank Dr. Florence Muindi for steadfastly holding my hand, sharing ideas and
encouraging me to keep going thought-out the process. The advice and recommendations
from the Doctoral Studies Committee, resource persons and those who chaired the various
presentations are greatly appreciated. Their valuable comments, suggestions and
recommendations helped improve my work. On this account, I want to sincerely express
my gratitude to Dr Mirie Mwangi, Professor Zack Awino, Professor Munyoki, Dr Joseph
Owino, Dr Muindi, Dr Kate Litondo and Dr Mary Kinoti.
Deep appreciation is extended to my colleagues at University of Nairobi and particularly
Professor Martin Ogutu and Professor Evans Aosa who kept pushing me to complete my
doctoral studies over the years thank you for keeping the pressure on. I am grateful to my
colleagues at SBO Research -Reuben, Boniface, Edwin, Cicily and Caroline who served
as a sounding board on mobile money transfer issues and for assistance in data analysis,
editing and printing. I appreciate all my friends who contributed in one way or another
towards the completion of this thesis.
To my dear parents Mary and Arthur, thank you for your untiring love and spiritual support
throughout this journey. To my Son Andrew - your love, optimism and philosophical
outlook inspired me greatly. To my partner Owens your love, commitment and care kept
me energized as I wrote this thesis. To my sisters and brothers and all members of `Ngahu
Choir', your words of encouragement and musical expression truly motivated me. May
God bless you abundantly!
v
TABLE OF CONTENTS
ACKNOWLEDGMENT ... iv
LIST OF TABLES ... ix
LIST OF FIGURES ... x
ABBREVIATIONS AND ACRONYMS ... xi
ABSRACT ... xiii
CHAPTER ONE ... 1
INTRODUCTION ... 1
1.1
Background of the Study ... 1
1.1.1
Perceived Justice ... 4
1.1.2
Service Failure Attribution ... 6
1.1.3
Recovery Disconfirmation ... 8
1.1.4
Recovery Satisfaction ... 9
1.1.5
Mobile Money Transfer Service ... 11
1.1.6
Mobile Money Transfer Service Subscribers in Kenya ... 13
1.2
Research Problem ... 16
1.3
Research Objectives ... 22
1.4
Value of the Study ... 22
CHAPTER TWO ... 25
LITERATURE REVIEW ... 25
2.1
Introduction ... 25
2.2
Theoretical Foundation of the Study ... 25
2.2.1
Equity Theory ... 26
2.2.2
Attribution Theory ... 27
2.2.3
Expectancy Disconfirmation Theory ... 30
2.2.4
Cognitive Dissonance Theory ... 31
2.3
Perceived Justice and Recovery Satisfaction ... 32
2.4
Perceived Justice, Service Failure Attribution and Recovery Satisfaction
... 40
2.5
Perceived Justice, Disconfirmation and Recovery Satisfaction ... 46
2.6
Perceived Justice, Service Failure Attribution, Disconfirmation and
Recovery Satisfaction ... 54
2.7
Conceptual Framework ... 60
2.8
Conceptual Hypotheses ... 61
CHAPTER THREE ... 62
RESEARCH METHODOLOGY ... 62
3.1
Introduction ... 62
3.2
Research Philosophy ... 62
3.3
Research Design ... 63
3.4
Population of the Study ... 64
3.5
Sample Design ... 66
3.6
Data Collection ... 67
3.6.1
Reliability of the Research Instrument ... 69
vi
3.6.2 Validity of the Research Instrument ... 70
3.7
Operationalization of the Study Variables ... 72
3.8
Data Analysis ... 74
3.9
Chapter Summary ... 79
CHAPTER FOUR ... 80
DATA ANALYSIS, INTERPRETATION AND DISCUSSION ... 80
4.1
Introduction ... 80
4.2
Respondents' Screening and Response Statistics ... 80
4.3
Test of Reliability ... 83
4.4
Validity Tests ... 84
4.5
Respondents Characteristics ... 85
4.5.1
Mobile Money Transfer Service Used ... 85
4.5.2
The Number of Years Respondent had used the Mobile Money Transfer
Service ... 85
4.5.3
Failures for which Service Recovery was sought by Respondents ... 86
4.5.4
Distribution of respondents by Region ... 87
4.5.5
Respondents' Highest Level of Education ... 88
4.5.6
Age of Respondents ... 89
4.5.7
Gender of Respondents ... 91
4.5.8
Respondents' Occupation ... 91
4.6
Assessment of Perceived Justice ... 92
4.6.1
Perception of Procedural Justice ... 94
4.6.2
Perception of Interactional Justice ... 94
4.6.3
Perception of Distributive Justice ... 97
4.7
Assessment of Service Failure Attribution ... 98
4.8
Assessment of Recovery Disconfirmation ... 100
4.9
Assessment of Recovery Satisfaction ... 101
4.10
Test of Normality, Linearity, Collinearity and Homoscedasticity ... 102
4.10.1 Tests for the Pertinent Assumptions ... 102
4.10.2
Collinearity Statistics ... 103
4.11
Correlation Analysis ... 103
4.12
Regression Analysis and Hypotheses Testing ... 104
4.12.1
Perceived Justice and Recovery Satisfaction ... 104
4.12.2
Dimesions of Perceived Justice and Recovery Satisfaction ... 107
4.12.3
The mediating effect of recovery disconfirmation ... 110
4.12.4
The Moderating Influence of Service Failure Attribution ... 116
4.12.5
The Moderation Effect of Dimensions of Service Failure Attribution .. 120
4.12.6
The Joint Effect of Perceived Justice, Disconfirmation and Service
Failure Attribution on Recovery Satisfaction ... 121
4.13
Summary of Research Objectives, Hypotheses and Conclusions ... 124
4.14
Discussion of the Results ... 128
4.14.1
The Relationship between Perceived Justice and Recovery Satisfaction
128
4.14.2
The Mediating effect of Recovery Disconfirmation ... 129
4.14.3
The Moderating effect of Service Failure Attribution ... 130
4.14.4
Perceived Justice, Service Failure Attribution, Recovery Disconfirmation
and Recovery Satisfaction ... 130
4.15
Chapter Summary ... 131
CHAPTER FIVE ... 132
vii
SUMMARY, CONCLUSION, RECOMMENDATIONS AND IMPLICATIONS
132
5.1
Introduction ... 132
5.2
Summary ... 132
5.3
Conclusions ... 135
5.4
Implications ... 137
5.4.1
Theoretical Implications ... 138
5.4.2
Policy Implications ... 139
5.4.3
Implications for Practitioners ... 140
5.5
Limitations of the Study ... 142
5.6
Suggestions for Further Research ... 143
REFERENCES ... 145
APPENDICES ... 153
MOBILE MONEY TRANSFER SERVICES QUESTIONNAIRE ... 153
APPENDIX 1B ... 161
HOJAJI YA HUDUMA ZA UHAMISHAJI WA FEDHA KWA NJIA YA SIMU
... 161
APPENDIX 2 ... 168
LIST OF MOBILE MONEY TRANSFER SERVICE PROVIDERS USING MNO-
LED MODEL ... 168
APPENDIX 2(A) ... 169
CBK MONTH ON MONTH MOBILE MONEY TRANSFER STATISTICS 2007
- 2016 ... 169
APPENDIX 3 ... 179
SAMPLE SIZE TABLE ... 179
APPENDIX 4 ... 180
Principal Component Analysis of Perceived Justice Scale (Rotated Component
Matrix) ... 180
APPENDIX 5 ... 181
Distribution of respondents by County ... 181
APPENDIX 6 ... 182
APPENDIX 7 ... 183
APPENDIX 8 ... 184
Regression Results for the Mediating Effect of Recovery Disconfirmation (Steps 1-
3) ... 184
APPENDIX 9 ... 187
Moderating Influence of Locus of Attribution on the Relationship between
Perceived Justice and Recovery Satisfaction ... 187
APPENDIX 10 ... 188
The moderating influence of stability attribution on the relationship between
perceived justice and recovery satisfaction ... 188
viii
APPENDIX 11 ... 189
Regression Results for assessing the moderating influence of controllability
attribution on the relationship between perceived justice and recovery
satisfaction ... 189
ix
LIST OF TABLES
Table 1.1: Mobile Money Transfer Services Growth statistics
14
Table 2.1: Summary of Knowledge Gaps
57
Table 3.1: Description of the Population of the Study
65
Table 3.2: Sample Design
67
Table 3.3: Operationalization of the Study Variables
.73
Table 3.4: Summary of Analytical methods and Interpretation
78
Table 4.1: Study Response Statistics
82
Table 4.2: Reliability Analysis
83
Table 4.3: Number of Years Respondent has used service
86
Table 4.4: Failures for which Service Recovery was sought by Respondents
87
Table 4.5: Distribution of Respondents by Region
88
Table 4.6: Highest Level of Education
89
Table 4.7: Age of Respondents
89
Table 4.8: Respondents' Gender
91
Table 4.9: Occupation of Respondents
92
Table 4.10: Descriptive Statistics for Perceived Justice Dimensions
93
Table 4.11: Descriptive Statistics for the Indicators of Perceived Justice
96
Table 4.12: Descriptive Statistics for the Indicators of Service Failure Attribution
99
Table 4.13: Statistics for Indicators of Recovery Disconfirmation
100
Table 4.14: Statistics for the Indicators of Recovery Satisfaction
101
Table 4.15: Correlation analysis for the Study Variables
103
Table 4.16: Regression Results for Perceived Justice and Recovery Satisfaction
106
Table 4.17: Regression Results for Dimensions of Perceived Justice and Recovery
Satisfaction
108
Table 4.18: Regression Results for the Mediating Effect of Recovery Disconfirmation
113
Table 4.19: Summary of Mediating Effect of Recovery Disconfirmation
114
Table 4.20: Regression Results for Assessing the Moderating Influence of Service
Failure Attribution on the Relationship between Perceived Justice and
Recovery Satisfaction
118
Table 4.21: Regression Results for Determining the Joint Effect of Perceived Justice,
Disconfirmation and Service Failure Attribution on Recovery Satisfaction
123
Table 4.22: Summary of Research Objectives, Hypotheses and Conclusions
125
x
LIST OF FIGURES
Figure 2.1: Conceptual Model ... 60
Figure 3.1: Moderation Path Diagram ... 75
Figure 3.2: Model for Testing Mediating Effect ... 76
Figure 4.1: Path Diagram for Mediation Effect of Recovery Disconfirmation ... 115
Figure 4.2: Modified Conceptual Model ... 127
xi
ABBREVIATIONS AND ACRONYMS
AFI
Alliance for Financial Inclusion
B2P
Business to Person
BoP
Base of the Pyramid
CATI
Computer Assisted Telephone Interview
CBK
Central Bank of Kenya
CA
Communications Authority of Kenya
CCK
Communications Commission of Kenya
CGAP
Consultative Group to Assist the Poor
EPG
Electronic Payment Guidelines
FSD
Financial Sector Deepening
GoK
Government of Kenya
GSMA
Global System for Mobile Communications Association
ICT
Information and Communication Technology
IFRC
International Federation of the Red Cross and Red Crescent Societies
IEA
Institute of Economic Affairs
KMO
Kaiser-Meyer-Olkin
MNO
Mobile Network Operator
MVNO
Mobile Virtual Network Operator
MMTS
Mobile Money Transfer service
NCSB
Norwegian Customer Satisfaction Barometer
NGO
Non-Governmental Organization
NPSA
National Payment systems ACT
P2B
Person to Business
P2M
Person to Merchant
xii
P2P
Person to Person
PJ
Perceived Justice
POS
Point of Sale
RD
Recovery Disconfirmation
RS
Recovery Satisfaction
SFA
Service Failure Attribution
SIM
Subscriber Identification Module
VIF
Variance Inflation Factor
WOM
Word of Mouth
xiii
ABSRACT
The study sought to assess the influence of perceived justice, service failure attribution, and
recovery disconfirmation on recovery satisfaction among subscribers of mobile money
transfer services in Kenya. The specific objectives of the study were to determine the
influence of perceived justice on recovery satisfaction; assess the mediating effect of recovery
disconfirmation on the relationship between perceived justice and recovery satisfaction;
establish the moderating influence of service failure attribution on the relationship between
perceived justice and recovery satisfaction and to determine the joint effect of perceived
justice, disconfirmation and service failure attribution on recovery satisfaction. Four pertinent
hypotheses were derived from the objectives based on the literature reviewed. The population
of the study comprised mobile money transfer service subscribers in Kenya. A descriptive
cross-sectional survey was used. Primary data were collected using semi-structured
questionnaires and a final sample of 803 respondents was realized. Reliability and validity
tests were conducted using data from a pilot study. Data were analyzed using descriptive
statistics, factor analysis, correlation, and regression analysis. The pertinent results indicated
that perceived justice had a positive and statistically significant relationship with recovery
satisfaction. The results also showed that the dimensions of procedural justice, interactional
justice and distributive justice have a positive influence on recovery satisfaction. In addition,
the results revealed that service failure attribution had a significant moderating influence on
the relationship between perceived justice and recovery satisfaction while recovery
disconfirmation had a significant mediating influence on that relationship. Finally, the joint
effect of perceived justice, service failure attribution and disconfirmation on recovery
satisfaction was statistically significant. The results of this study add to theory by establishing
a linkage between perceived justice, recovery disconfirmation, service failure attribution and
recovery satisfaction. The study is relevant for regulators in setting policy for service
providers with regard to customer service recovery standards and redress systems. The results
will serve as a point of reference for service providers and managers in assessing recovery
satisfaction for effective strategic marketing decisions. The limitations of this study include
the non-exhaustive selection of variables, focus on one sector and the use of recall based
survey. Future research should consider inclusion of other relevant variables and replication
of the study in other sectors.
Key words: Perceived Justice; Service Failure attribution; Recovery Disconfirmation;
Recovery satisfaction; Mobile money transfer
1
CHAPTER ONE
INTRODUCTION
1.1
Background of the Study
Customer satisfaction is crucial in maintaining a mutually beneficial relationship
between service providers and their customers. While service organizations make great
efforts to ensure customer satisfaction, unique features of services such as variability
and intangibility make service failure inevitable (Kotler & Keller, 2012). Service
failure comprises any real or perceived problem that occurs during a customer's
interaction with the service provider (Maxham, 2001). It may include unavailability,
poor delivery, inconsistent outcomes or any incident where a service fails to meet
customer expectations.
The negative disconfirmation following a service failure leads to dissatisfaction with
the likely consequence of increased customer complaints, negative word-of-mouth
(WOM) and erosion of patronage of the service provider (Kau & Loh, 2006).
Organizations strive to develop and implement service recovery strategies in order to
restore satisfaction. Customers evaluate service recovery from a perspective fairness of
process, interaction and outcome in line with the dimensions of perceived justice.
Based on equity theory, the tendency of customers to complain when they experience
service failure stems from a perceived injustice arising from an imbalance or inequity
in the relationship between the customer and service provider. The customer will expect
the company to offer a recovery to compensate for the imbalance. However to get this
compensation the customer has to spend time and effort (Chebat & Slusarczyk, 2005).
This complaining and recovery process is evaluated from a viewpoint of fairness based
2
on the procedural, interactional and distributive dimensions of perceived justice
(Ellyawati et al., 2012).
Recovery satisfaction is associated with the quality of the corrective action taken after
a service failure. On the other hand, disconfirmation, that is, the discrepancy between
expectations and perceived performance is also expected to have an influence on
satisfaction. Furthermore, the cause of a service failure is expected to influence the
level of dissatisfaction, with greater anticipation of redress in situations where the
failure is attributed to the service provider's negligence.
The study of customers' perceived justice, attribution, disconfirmation and recovery
satisfaction falls under theories of equity (Adams, 1965), disconfirmation (Oliver,
1980), cognitive dissonance (Festinger, 1957) and attribution (Weiner, 2000). The
notion of equity suggests that people seek fairness in exchange relationships.
Accordingly, customers expect fairness from service providers in service recovery
situations and commonly base their evaluation on the process, interaction and outcome.
Disconfirmation theory explains post purchase satisfaction as a function of comparing
expectations and actual performance. Attribution theory relates to the perception of the
cause of the service failure based on three dimensions of locus, stability and
controllability. On the other hand, cognitive dissonance theory asserts that people seek
consistency between their beliefs and perceptions and tend to experience psychological
discomfort or doubt following disconfirmation of expectations.
Financial sector deepening (FSD) is considered critical to economic growth and
contributes significantly to the overall functioning of the economy. In emerging
3
economies FSD helps reduce poverty and inequality by broadening access to finance
to the poor and vulnerable groups (World Bank, 2013). Digital platforms contribute by
offering the opportunity to speedily scale up access to financial services using mobile
phones. Mobile money payment platforms are of great interest to the base of the
pyramid (BoP) scholars and practitioners as they seek to develop models for enhancing
access to formal financial systems to the poor. They provide an opportunity for the poor
in the informal sector to enter into formal credit, savings and insurance transactions.
Mobile money transfer service (MMTS) is an important component of FSD which is
now viewed as a great opportunity to increase financial access in emerging markets by
offering convenient and affordable financial services to unbanked and under-banked
individuals around the world (GSMA, 2013). The resulting electronic money
ecosystem is expected to improve livelihoods of those at the bottom of the pyramid
earning less than two dollars a day (AFI, 2010). The development of MMTS in Kenya
benefited from active encouragement by the Government of Kenya (GoK) which
sought to enhance financial inclusion as outlined in the Vision 2030 blueprint (GoK,
2007).
The evolution of mobile money in Kenya has created increasing competition among
service providers and greater expectations among customers thus increasing the
significance of service recovery management. Despite the central role played by mobile
money, subscribers experience service failures related to occasional network
breakdown, agent system weaknesses and service menu issues. Some of these service
failures are of such severity that a customer will seek recovery from the service
provider.
4
Service providers need to understand the key variables that influence recovery
satisfaction and their relative significance. However, a broad review of the literature
has revealed a scarcity of empirical studies focused on this critical area of service
marketing in Kenya and particularly with reference to mobile money transfer services.
There is therefore a need to establish the influence of perceived justice, attribution and
disconfirmation on recovery satisfaction in the Kenyan setting.
1.1.1
Perceived Justice
The concept of perceived justice relates to the perception of fairness. In a service
recovery context, perceived justice refers to the customer's perception of fairness in the
recovery effort made by the service provider following an unsatisfactory initial service
(Tan, 2014). It is based on the customer's subjective interpretation of the overall
recovery experience including process, interaction and outcome. When customers
experience a service failure, for example a delay in confirming money transfer for
electricity bill payment, they will feel aggrieved and contact the service provider to
seek recovery or correction of the problem.
Customers perceive service failure as an injustice based on an imbalance in the
relationship between them and the service provider (Chebat & Slusarczyk, 2005).
There is a perceived imbalance in the exchange because the customer has invested time,
effort, money and trust, yet the service provider has not delivered the service as
promised. Further, the assessment of fairness may also be influenced by their
expectations of the service provider and the knowledge of how other customers in
similar situations were treated.
5
Perceived justice is significant in service recovery management because a perceived
lack of fairness impacts satisfaction, repurchase intention, word-of-mouth and loyalty
(smith et al., 1999; Nibkin et al., 2010). Perceived justice is a multi-dimensional
concept comprising three dimensions, namely distributive, procedural, and
interactional justice. Procedural justice refers to the perceived fairness of procedures
applied to rectify the service failure (Río-Lanza et al., 2009). Service organizations are
expected to develop and implement effective systems for addressing service failures
including accessible processes which enable frontline employees to react with speed.
The second dimension of perceived justice is interactional justice which focuses on
both interpersonal treatment and the relevance of the information provided during the
service encounter. Interpersonal treatment concerns the behavior of frontline
employees as they interact with customers in a service recovery situation. An apology
is a key aspect of interactional fairness that seems particularly important in
implementing service recovery. Informational justice focuses on the perceived
adequacy, truthfulness and usefulness of the information provided in explanation of the
problem during service recovery. Interactional justice highlights the importance of
honesty, empathy and respect.
Distributive justice refers the perceived fairness of the redress offered in the form of
assignment of tangible resources by the offending party to rectify a service failure with
a view to restoring equity (Tan, 2014). Distributive justice is frequently evaluated from
the point of view of compensation which may be in the form of exchange or monetary
refund. Customers tend to expect the service provider to compensate them for any
tangible loss they suffered because of the specific service failure. The level of
6
compensation expected may be tied to the severity or magnitude to the problem (Smith
et al., 1999). These dimensions were adopted in the current study.
1.1.2
Service Failure Attribution
Attribution is the perceived cause of an occurrence. It relates to largely subjective
inferences about the causes of an event such as service failure. The perceived reason
for a service failure is important because it influences the degree of dissatisfaction as
well as consumer expectations for service recovery. Service failure occurs when a
service fails to meet customer expectations. This may be based on reality or perception
and could be related to the quality of the service delivery process, interaction with
frontline employees or outcome (Maxham, 2001). A service failure is viewed as
different from the failure of a tangible product because a service is associated with a
psychological and largely personal outcome (Hocutt et al., 1997).
It has been suggested that individuals tend to perform attribution searches for most
negative occurrences (Weiner, 2000). The cause of a service failure is of interest to
customers and many are likely to attempt to find out why it happened (Folkes, 1984).
Attribution is significant because it serves to provide the customer with the incentive
to act or seek redress. However, if an appropriate explanation is provided and an
effective recovery strategy executed the focus on attribution can be minimized.
Customers who experience a service failure tend to form attributions along dimensions
of locus, control, and stability (Weiner, 2000).
Locus relates to who is considered to be responsible for the failure incident or who is
to blame for the problem, between the customer and the service provider (Swanson
Hsu, 2011). The interactive nature of many services today with customers participating
in self-service procedures impacts on customer allocation of responsibility for service
7
failure. A study by Harris et al. (2006) found that online customers tend to expect less
of a recovery than offline customers following a service failure mainly because they
consider themselves to blame for many service failures.
The stability dimension relates to the extent to which a cause is viewed as temporary
or permanent. It is assessed by considering whether the service failure is likely to occur
very often. The controllability dimension concerns whether a service failure could have
been avoided and is assessed by considering the degree to which a cause is perceived
to be within or outside the service provider's control. It has been suggested that the
more customers associate a service failure with external locus (service provider or
agent), stability (likely to occur frequently) and controllability (avoidable), the more
they are likely to be dissatisfied (Swanson Hsu, 2011).
Attribution search is impacted by service failure severity and criticality. A more severe
or critical failure will lead to a different customer perception of harm or loss than a
minor one (Weun et al., 2004). For instance, a failed mobile money transfer transaction
is more likely to motivate attribution if the money was to pay for hospital admission
than if it was for buying household goods. Consequently, it is important service staff
to listen to the customer keenly in order to appreciate the perceived harm as they work
to build an appropriate recovery strategy. Weiner (2000) noted that admission of
responsibility decreases conflict while enhancing controllability attribution. The use of
compensation as a recovery strategy is taken to imply admission of responsibility by
the service provider an increases controllability attribution
Stability attribution causes customers to expect similar service failures in future while
controllability attribution is associated with enhanced anger and lower recovery
8
satisfaction. External attribution is associated with negative emotions of anger and
disgust while internal attribution is associated with emotions of guilt or shame (Smith
Bolton, 2002). It is the role of management of the service provider firm to ensure
that appropriate recovery strategies are applied in rectifying service failure with a view
to restoring customer satisfaction. This study sought to investigate the perception of
service failure attribution particularly as it relates recovery satisfaction.
1.1.3
Recovery Disconfirmation
Service recovery has been recognized as a critical process in service marketing since
service failures are inevitable owing to the unique nature of services and varied
customer expectations. Recovery disconfirmation refers to the incongruity between
recovery expectations and recovery performance (McCollough et al., 2000). Service
recovery is the action taken by the service provider to remedy the service failure.
Recovery expectations relate to the level of effort anticipated by the customer based on
the service provider communications and previous experience.
The expectancy disconfirmation model proposes a process by which consumers
compare their prior expectations with a service performance and how the comparison
result leads to satisfaction or dissatisfaction (Oliver, 1980). This applies in recovery
situations in a similar manner, with a performance that is consistent to expectations
leading to recovery satisfaction. However, if the performance is lower than expected
then a negative disconfirmation occurs, which leads to recovery dissatisfaction
(Serenko Stach, 2009).
Based on the disconfirmation paradigm, following the evaluative process, the customer
may have three possible conclusions which are: disconfirmation which is a neutral state
9
which occurs when the service is performed as expected; negative disconfirmation
which occurs when the service performance fails to match prior expectations; or
positive disconfirmation where the service performance is better than prior
expectations (Oliver, 1980). A service failure implies a discrepancy between expected
service and actual service provider's performance, hence a negative disconfirmation.
Following service failure, customers are likely to complain because of the need to
reduce the discrepancy.
Recovery expectation arises after a service failure encounter and is different from the
initial expectation which a customer has before a service failure occurs. Since not all
customers who are dissatisfied complain, it is argued that those who do are motivated
by recovery expectations. Recovery disconfirmation is also affected by service failure
severity and criticality as they influence customer perception of the level of harm or
loss incurred as a result of the failure (Smith et al, 1999; Weun et al, 2004). The
literature highlights apology, problem solution and compensation as important
attributes of the service recovery expectation (Krishna, et al., 2011). This study
considered the role of recovery disconfirmation in the recovery satisfaction judgment.
1.1.4
Recovery Satisfaction
Customer satisfaction is a measure of how services provided meet or surpass customer
expectations (Kotler Keller, 2012). According to Oliver (1997) customer satisfaction
is a subjective judgment that a service provides fulfillment. Satisfaction is associated
with service performance that adds value to the customer and is based on the appraisal
of service from cognitive as well as affective perspectives. Recovery satisfaction
relates to the favorability of a customer's subjective evaluation of the corrective action
taken following service failure. Hence, it reflects customer satisfaction with the service
10
provider's recovery effort including the perception of the process, interaction and
solution.
Recovery satisfaction has an important effect on customer evaluation of the business
and may contribute to retention and loyalty. An effective recovery promotes brand
evangelism, which is an extension of word of mouth communication and positive
referrals (Rashid Ahmed, 2014). Service recovery is the corrective action taken by
the service provider to correct a service failure and is aimed at returning the customer
to satisfaction. It's a well thought out and strategic approach to service problems which
is different from complaint management in that it focuses on the service providers'
immediate response to service failures.
Service recovery is a proactive strategy whose goal is to anticipate and address service
failures without waiting for customers to complain (Smith et al., 1999). As a proactive
strategy it implies that any service problems that occur are appropriately addressed,
analyzed, and documented to ensure that service delivery systems and procedures are
revised to prevent a recurrence. In a service recovery context, satisfaction is viewed as
the result of appropriate action taken by the service provider in resolving the problem.
Kau and Loh (2006) observed that recovery satisfaction or dissatisfaction affects the
customer's attitude towards the service provider and determines repurchase intention
as well as willingness to recommend the organization to others. Further, Smith and
Bolton (2002) suggested that when initial service failures were addressed effectively,
customers were likely to react positively. Still, although dissatisfied customers may
contemplate brand switching, structural factors such as switching costs and availability
11
alternatives may deter them. As such, even after a failed recovery, some customers may
choose to stay with the provider.
The service provider has various options for resolving a service failure including
offering an apology, rectifying the error, offering a discount, or compensating the
customer for the service failure. Service recovery efforts may have important
implications for levels of satisfaction which in turn influence repurchase intent and
positive WOM (Maxham, 2001). Recovery satisfaction indicators include service level,
solution provided, outcome, and overall satisfaction. In this study, the term recovery
satisfaction was used to refer to the level of overall fulfilment following the corrective
action with the consequence that the customer continues to patronize the service
provider.
1.1.5
Mobile Money Transfer Service
The mobile money transfer subsector is a significant component of the financial
services sector and plays a critical role in the economy by enhancing financial access.
Mobile money transfer service (MMTS) originated from the convergence of mobile
telecommunication and financial sectors. The growing significance of mobile money
led the Central Bank of Kenya (CBK) to change its regulatory strategy of applying
traditional banking law to MMTS and to develop Electronic Payment Guidelines (EPG)
in 2011. This was later followed by enactment of the National Payment Systems ACT
(NPSA) in 2011 (Buku Meredith, 2013).
Vodafone through Safaricom pioneered commercial MMTS globally in 2007 through
M-Pesa. The amount of money transferred through MMTS in Kenya grew to 2.3 trillion
shillings (US$26 billion) in 2014 compared to 473 billion shillings (US$5.2billion) in
12
2009 (CBK 2015). There was a month on month growth in transactions and value since
introduction of MMTs to the end of 2015 (Appendix 2A). The growth in transaction
value is attributed to the increasing interface with banks which has eased transactions
and enhanced access.
Omwansa and Sullivan (2012) observed that increasing transactions have led to Kenya
being recognized as the leading MMTS market in the world with the model being
replicated in various countries. Further, mobile money has served to catalyze the digital
ecosystem which has reinforced the growth of digital entrepreneurship in Kenya,
demonstrated by the emergence of many tech start-ups using mobile money pay
systems (GSMA, 2015).
Mobile money involves transferring money using the infrastructure of the mobile
network operators (MNOs) through the cellular phone using a network of agents for
deposit and collection. A highly reliable network of cash-in and cash-out agents is a
critical bridge between physical cash and e-float. These agents convert cash to e-float
(cash-in) and e-float to cash (cash-out). These services are largely provided by MNOs,
generally referred to as the MNO-led model (IEA, 2011). The licensed MNOs who
provide mobile money services in Kenya are Safaricom, Airtel and Telkom (Orange).
Yu closed operations in 2014 and sold her infrastructure and customers to Safaricom
and Airtel respectively.
The Communication Authority of Kenya (CA) issued Mobile Virtual Network
Operator (MVNO) licenses in 2014 to three firms namely Zioncell, Tangaza and Fincell
(owned by Equity Bank) to provide services using the infrastructure of MNOs. The
introduction by Equity Bank of the Thin SIM which is overlaid on the existing SIM
13
card to allow the customer to access the mobile money services of two operators
simultaneously has raised some controversy. While some view it as an excellent
opportunity to expand the offering of mobile money services in Kenya, others argue
that it may compromise the security of mobile money transactions and violate the
privacy of subscribers. Ultimately, this is expected to increase competition in the
subsector.
The large number of MMTS transactions means that some service failures are bound
to occur, leading to disconfirmation and dissatisfaction. A survey of 3000 M-Pesa
customers commissioned by CBK (AFI, 2010) identified various service failure types
including agency lack of cash or e-float, occasional network breakdown and dishonesty
by agent. Other disappointments arise from delay or failure by the service provider to
process transaction reversals requested by subscribers and delay in confirming
transaction.
Customers perceive an injustice when service recovery fails to meet expectations and
are likely to seek redress regardless of service failure attribution. In order to enhance
recovery satisfaction, the providers need to understand the key variables that influence
satisfaction in MMTS in Kenya and their relative significance.
1.1.6
Mobile Money Transfer Service Subscribers in Kenya
Mobile money subscribers are mobile phone or SIM card holders who are registered
for mobile money service with a provider. At over twenty six million subscribers Kenya
has the highest mobile money penetration in the world. Of the registered subscribers,
half are considered active, having had a transaction within a thirty day period (GSMA,
2015). The high subscription rate in Kenya is attributed to new uses of mobile money
and increasing interface with banks which has led to enhanced efficiency. Generally,
14
MMTS subscribers in Kenya use the service to send money, receive money, and pay
bills as well as for airtime top up.
Table 1.1 below presents mobile money transfer services growth statistics. It shows a
significant growth in the number of subscribers.
Table 1.1: Mobile Money Transfer Services Growth Statistics
June 2012
June 2013
June 2014
Mobile phone subscribers (Million)
29.2
30.5
32.2
Mobile phone penetration (%)
75
77.3
79.2
Mobile money transfer subscribers
19,501,702
24,840,404
26,611,077
Mobile money transfer Agents
49,079
81,422
109,286
Source: Communications Authority of Kenya, Quarterly Sector Statistics Report
(2014)
Subscribers have the opportunity to use various MMTS offerings including person-to-
person transfer (P2P), business-to-person transfer (B2P) and person-to-business
transfer (B2P) (USAID, 2011). P2P is a transfer between individual mobile money
subscribers. B2P entails the transfer of money from an organization to an individual as
is the case with transactions originating from government, business and
nongovernmental organization (NGO).
P2B allows an individual subscriber to pay an institution through MMTS and is
generally used as a bill payment service. It is commonly used by customers to pay
utilities bills for electricity and water supply as well as other predefined payments such
as school fees and pay television charges. This is done by using the company's pay-bill
number as a reference for the recipient and the subscriber's account number as a
15
reference for the customer. In case of donations for some churches where they need a
record of contributors yet they do not have account numbers, the platform allows for
use of the institutions pay bill number and the customer name as account number. The
customer receives a confirmation of payment from the service provider and a short
message service (SMS) receipt from the receiving organization.
Another offering of MMTS used by subscribers is a modified version of the bill
payment service which may be termed person-to-merchant (P2M). This enables
customers to pay for goods and services at outlets such as supermarkets, shops, fuel
stations, hair salons and restaurants. This was introduced by Safaricom under the name
`Lipa na M-Pesa' which is Swahili for pay with M-Pesa. The mobile service provider
issues merchants with a till number which the customer uses to make the payment
directly to the merchant's account. This may be linked to the merchant's bank account
allowing the payment to be deposited to the bank instantly and can also be integrated
to point of sale (POS) systems to create a formal record of sale and generate receipt for
the customer.
Another variant of B2P is B2C or bank to customer which is used by subscribers to
withdraw money from their bank accounts to their mobile wallets. Notably, MMTS is
now generally used as an alternative delivery channel by banks for deposit, withdrawal,
savings and loan services. The increase in the number of services available through
mobile money transfer has improved customer convenience and increased the
proportion of business transacted through that mode in the Kenyan economy.
Issues of subscriber recovery satisfaction with MMTS have been given new attention
in recent times. Recognizing the importance of customer satisfaction for the success of
16
MMTS, the Central bank of Kenya provided detailed guidelines on customer protection
and service recovery management in the National Payment System Regulations (CBK,
2014). The guidelines require mobile money service providers to establish a customer
redress and complaints handling mechanism and to inform MMTS subscribers of
procedures for lodging complaints including how to escalate the complaint if the
customer is not satisfied with initial service recovery effort.
The payment guidelines further require the service provider to provide clear
information on expected outcomes and timelines for complaints resolution. Similarly,
the GSMA (2014) code of conduct for mobile money providers globally requires them
to develop mechanisms to ensure that complaints are effectively addressed and
problems are solved in a timely manner and that personal data is collected, processed
and transmitted fairly and securely (GSMA, 2014).
1.2
Research Problem
The performance of a transaction by a service provider is perceived to meet, exceed or
fall below customer expectations resulting in satisfaction, delight or dissatisfaction
respectively. Dissatisfied customers expect the service provider to engage in a recovery
process to correct the service failure and restore satisfaction. A recovery creates a new
service loop with a new evaluation process.
The literature suggests various factors that are expected to influence recovery
satisfaction including perceived justice, service failure attribution and disconfirmation.
The assumed impact of these factors is based on the theories of equity, disconfirmation,
cognitive dissonance and attribution. Whereas several propositions exist regarding the
17
influence of various factors on recovery satisfaction, there are still significant gaps and
unanswered questions regarding their individual roles and interactions.
Although service recovery has received attention in the literature and has been
recognized as a major factor in customer retention, there are major inconsistencies in
the literature with regard to the relationship between perceived justice and recovery
satisfaction as well as the role of service failure attribution and disconfirmation in that
relationship. Different scholars have evaluated the concept differently resulting in
diverse interpretation and conclusions. Maxham and Netemeyer (2002) observed that
some of the contradictions are associated with different treatment of transactional and
overall satisfaction in evaluating service recovery efforts.
There is still considerable debate on perceived justice with some authors analyzing its
dimensions in aggregate while others analyze them separately asserting that the relative
effect of each dimension is different (Ellyawati et al. 2012). Moreover the studies have
used different methodological approaches with reference to research design, sampling
and analysis making comparison difficult. These conflicting perspectives and diverse
methodological approaches have led to a situation where the relationships between
perceived justice, service failure attribution, disconfirmation and recovery satisfaction
have not been clearly expounded.
The MMTS sub-sector is a major part of the financial services sector and has become
a payment and transaction tool of growing importance to subscribers and the Kenyan
economy as a whole. There are now more mobile money accounts than bank accounts
and more mobile money agents than bank branches in Kenya (GSMA, 2013). MMTS
has contributed to financial sector deepening by enhancing financial access to the
18
unbanked and underbanked. Further, it has contributed to employment creation and the
establishment of new businesses in the mobile money value chain including
applications developers and MMTS agents who now number over one hundred
thousand.
Many donors, NGOs and micro finance institutions (MFIs) have integrated mobile
money transfer services in their programs and consider it an important platform for
enhancing efficiency and reach. Moreover, mobile money is viewed as the next frontier
for competition among mobile network operators (MNOs) as it is linked to improving
new customer attraction, increasing non-traditional telecommunication revenues and
reducing churn (IEA, 2011).
Despite the remarkable growth of MMTS, there are concerns regarding service failure
related to agent system weaknesses, network breakdown and service system issues. The
service failures include agent's lack of cash or e-float, unavailable service and delayed
reversal of wrong transactions (CGAP, 2009; Intermedia, 2014). Further, other service
failures relate to customer error such as forgetting PIN number or sending money to
the wrong number. In addition, there are numerous prison scams targeting MMTS
users.
Balasubramanian and Drake (2015) found that mobile money customers attribute
service failure responsibility differentially based on type of problem encountered.
Further, customers regularly seek corrective action for various types of failures
regardless of who they blame for the occurrence. The negative disconfirmation arising
when service recovery does not meet customer expectations leads to perceived injustice
and dissatisfaction. There was inadequate information regarding customers' perceived
19
justice, attribution, disconfirmation and recovery satisfaction with respect to MMTS in
Kenya. There is therefore an urgent need to expand knowledge in this area.
Previous theoretical and empirical studies on the relationship between perceived justice
and recovery satisfaction have produced mixed findings with the result that the specific
dynamics in this relationship are not conclusive. Further, these studies have used
different combination of variables making comparison problematic. Maxham and
Netemeyer (2002) conducted a longitudinal survey of 208 customers of an American
bank and concluded that there is a significant positive relationship between perceived
justice and recovery satisfaction. However, the study did not consider the relative
effects of perceived justice dimensions or the influence of disconfirmation.
Weber and Sparks (2010) used experimental design in a study of 331 randomly selected
airline customers in Australia and concluded that locus of attribution affects recovery
satisfaction. However, they did not consider the effect of disconfirmation. In addition,
the experiments were based on hypothetical scenarios which may not fully fit actual
customer experience with service failure and recovery encounters.
Andreassen (2000), analyzed data from the Norwegian Customer Satisfaction
Barometer (NCSB) and concluded that disconfirmation of expectations of service
recovery has an impact on satisfaction. The study did not consider the role of service
failure attribution. In a study at an American University, Hess et al. (2003) using an
experimental design found that customers with higher expectation of relationship
continuity were more likely to report higher satisfaction with an average recovery effort
than those without such expectations. However, the study did not consider the influence
of attribution.
20
In examining the effect of perceived justice on recovery satisfaction in Philippines Tan
(2014) used a descriptive survey customers recruited through purposive and snowball
sampling and concluded that distributive justice has the strongest effect on recovery
satisfaction. This corroborated the findings of an earlier descriptive survey conducted
with airline customers in Malaysia (Nibkin et al., 2010). However, both studies did not
examine the role of service failure attribution and disconfirmation in recovery
satisfaction. On the other hand, in a study of cellphone customers in Spain, Rio-Lanza
et al. (2009) reported that procedural justice has the strongest relative influence on
satisfaction. Likewise, the study did not factor the influence of disconfirmation or
service failure attribution.
There was paucity of empirical studies on perceived justice and satisfaction in the
regional and local context. Some of those available embraced two variables and were
narrow in geographic scope. For instance, Chepkwony et al. (2012) examined the
effects of distributive justice as a strategy for resolving complaints in the banking sector
in Eldoret, Kenya through a descriptive survey and found it to have a significant
influence on recovery satisfaction. Besides the narrow geographical scope, the study
did not consider the other dimensions of perceived justice or the role of disconfirmation
and attribution.
Komunda and Osarenkhoe (2012) conducted a descriptive survey of bank customers in
Makerere University, Uganda and reported that service recovery impacts recovery
satisfaction. Further, Smith and Mpinganjira (2015) conducted a survey among retail
bank customers in South Africa and concluded that procedural, interactional and
distributive dimensions of perceived justice influence recovery satisfaction. However,
21
this study did not consider the role of service failure attribution or recovery
disconfirmation.
Based on the foregoing discussion, it was clear that previous studies had not provided
sufficient evidence on the relationship between perceived justice, attribution,
disconfirmation and recovery satisfaction. There was a lack of an integrated framework
linking perceived justice, service failure attribution and recovery disconfirmation to
recovery satisfaction. Conceptual frameworks that sought to explain the recovery
satisfaction judgement, tended to focus on single explanatory variables leading to
partial explanations. Equally, previous studies had focused on direct relationship
between variables with minimal attention on the influence of moderating and mediating
variables.
Moreover, the role of perceived justice on recovery satisfaction had not been
adequately investigated in the Kenyan context. Consequently, there was urgent need
for a rigorous study to fill the gaps highlighted by applying an integrated approach that
comprises perceived justice, disconfirmation, attribution and recovery satisfaction. The
results of this study have provided support for an integrated framework that links
perceived justice, recovery disconfirmation and service failure attribution to recovery
satisfaction. The study was guided by the following research question: To what extent
do service failure attribution and disconfirmation influence the relationship between
perceived justice and customer recovery satisfaction with mobile money transfer
services in Kenya?
22
1.3
Research Objectives
The broad objective of the study was to assess the influence of perceived justice, service
failure attribution and recovery disconfirmation on recovery satisfaction among
subscribers of mobile money transfer services in Kenya. The specific objectives were
to:
(i) Determine the influence of perceived justice on recovery satisfaction among
subscribers of mobile money transfer services in Kenya.
(ii) Assess the effect of recovery disconfirmation on the relationship between
perceived justice and recovery satisfaction.
(iii) Establish the influence of service failure attribution on the relationship
between perceived justice and recovery satisfaction.
(iv) Determine the joint effect of perceived justice, disconfirmation and service
failure attribution on recovery satisfaction.
1.4
Value of the Study
The results of the current study have contributed to theory, policy and practice in
service marketing by offering insight into the relationship between perceived justice,
service failure attribution, and disconfirmation and recovery satisfaction. This is
important at this stage of mobile money growth given the limited scholarly work
focusing on service failure and recovery issues.
The result provides information on the common problems that cause customer
dissatisfaction with services and help focus management efforts towards proactive
service recovery strategies. It further provides a rationalization for documenting service
failures with a view using the information in the re-design of service protocols in order
23
to minimize repeated failures. Additionally, it offers insights into the key drivers of
recovery satisfaction and serves as a point of reference for service provider
organizations and managers in the development of service recovery strategies.
The research results demonstrate the importance of ensuring fairness in the service
recovery process, interaction and outcome within the MMTS sub-sector in Kenya. The
results demonstrate the influence of perceived justice on satisfaction and repurchase
intention. In addition, the study provides practitioners with the relevant indicators for
assessing recovery satisfaction for effective strategic marketing decision-making.
The results of this study have added to theory by providing a framework that links the
influence of perceived justice, recovery disconfirmation and service failure attribution
on recovery satisfaction. This is a significant contribution to service marketing theory
as it integrates equity theory, attribution theory, cognitive dissonance and
disconfirmation model in the assessment of customer satisfaction in service failure and
recovery encounters. In addition, this study has contributed to existing literature and
served to extend the prevailing discourse on the role of customer justice perceptions,
attribution and disconfirmation in the recovery satisfaction judgment.
As Kenya is the recognized global leader in MMTS and other countries are seeking to
replicate its model, this study has supplemented the scanty evidence available on the
perceived justice and recovery satisfaction relationship in this sector. It offers details
on the mobile money service systems that can be used by new entrants into the market.
The study further provides new perspectives on the endless opportunities for
integrating mobile money payment systems with a wide array of services and products.
24
Furthermore, it provides a new angle for the evaluation of the mobile money services
for those seeking to invest in the sector.
The insights generated are of value to both financial services and telecommunications
regulators as they develop policies for entrenching mobile money transfer services in
order to enhance financial inclusion. The study provides pertinent information for
mobile communication associations, policy makers, financial inclusion advocates,
government ministries and regulators as they seek to develop and implement
appropriate policies to promote and enforce standards of service delivery in the mobile
money subsector.
1.5 Organization of the Thesis
This study report is organized in five chapters. This chapter presents the conceptual
background on perceived justice, service failure attribution, recovery disconfirmation
and recovery satisfaction. Further, it covers the contextual background on mobile
money transfer services, the statement of the research problem, objectives and the value
of the study. The second chapter presents a detailed review of pertinent literature,
theoretical foundations of the study, summary of the knowledge gaps identified in the
literature, the conceptual framework and hypotheses of the study.
Chapter three presents the research methodology adopted in the study. It covers the
research philosophy, research design, population and sample of the study. It explains
the data collection methods, measurement of research variables and the data analysis
techniques used in the study. Chapter four presents data analysis, findings and
interpretation of results. Chapter five summarizes the entire study and provides a
summary of findings, discussions, conclusions and implications.
25
CHAPTER TWO
LITERATURE REVIEW
2.1
Introduction
This chapter reviews the literature on the variables of the study, namely perceived
justice, service failure attribution, disconfirmation and recovery satisfaction. It reviews
various theoretical perspectives including equity theory, attribution theory and
expectancy disconfirmation paradigm as they relate to recovery satisfaction and their
application to service marketing practices. The chapter also presents a summary of
selected studies, the identified knowledge gaps as well as the conceptual framework
and hypotheses to be used in the study.
2.2
Theoretical Foundation of the Study
There has been great interest among marketing scholars on the process customers' use
in evaluating services and eventually making the satisfaction judgment. As such, many
theories have been proposed to explain the satisfaction/dissatisfaction question. The
overarching theory for this study is the theory of consumer behavior. This theory
postulates that consumer needs, behavior and satisfaction are a factor of the interplay
between psychological and sociological influences and marketing activities (Schiffman
Kanuk, 2007).
Although several theories provide various perspectives regarding satisfaction, equity
theory viewpoint is considered particularly relevant in a service recovery context
(Ellyawatti et al., 2012). Other theories that support the perceived justice and recovery
satisfaction relationship include attribution theory, cognitive dissonance and
expectancy disconfirmation paradigm.
26
The equity perspective focuses on the expectation of fairness in an exchange
relationship, attribution theory relates to the allocation of responsibility for a negative
occurrence. On the other hand, disconfirmation theory concerns the discrepancy
between expectation and actual performance while cognitive dissonance theory
explains post-purchase dissatisfaction from the perspective of inconsistency between
beliefs and perceptions. The current study sought to integrate the four theoretical
perspectives in order to provide a more comprehensive assessment of the recovery
satisfaction judgment.
2.2.1
Equity Theory
Equity theory is adapted from social phycology is considered one of the justice theories
(Greenberg, 1990). It was proposed by Adams (1965) and asserts that there is a
universal human drive to pursue justice in exchange relationships. It endeavors to
explain relational satisfaction based on perception of fairness or equity between inputs
and outputs. It suggests that equity is assessed by comparing the ratios of contributions
and benefits of each person within the relationship. The theory posits that in an
exchange relationship, people compare their inputs against outcomes and relate that to
the other exchange party and determine whether there is a balance between them
(Adams, 1965).
The theory applies to recovery satisfaction because of the assumption that customers
consider the fairness of distribution of resources in the evaluation of service recovery.
Based on their perception of balance or lack of it, customers may deem an exchange as
either fair or unfair. Exchanges that are perceived to be unfair lead to feelings of
annoyance and discontent with the consequence that the individual seeks to restore
equity. When applied to a service failure and recovery encounters, customer inputs
27
include economic, time, energy and psychic costs (Hoffman Kelly, 2000). The output
includes the service provider's recovery process, apology and compensation or other
form of redress (Smith et al, 1999).
Equity theory was originally applied in human resource context to explain the impact
of perception of equity in the workplace but has recently gained popularity among
marketing scholars particularly with reference to service recovery evaluations. Spector
(2008) suggests that perceived negative inequity may lead to anger and frustration
while positive inequity would cause feelings of remorse and embarrassment. This
implies that customers who experience service failure view it as a negative inequity
and are likely to seek recovery to restore equity. The service recovery evaluation is
based on procedural, interactive and distributive dimensions of perceived justice
(Andreassen, 2000).
The main criticism directed to equity theory relates to the assumptions and practical
application of the theory with some scholars arguing that the model is too simplistic.
Questions have been raised regarding its ability to comprehensively capture the many
psychological and sociological considerations that influence people's interpretations of
fairness (Huseman et al., 1987). Another criticism has to do with the applicability to
real world situations of many equity studies which were based on laboratory settings.
2.2.2
Attribution Theory
Attribution theory on the other hand, is concerned with understanding the perceptions
of the causes of an occurrence. Attribution is the process by which individuals describe
the causes of behavior and events (Weiner, 2000). It involves linking outcomes to
possible causes or allocating blame for negative events. The theory seeks to explain
28
how individuals make causal explanations and is mainly concerned with how people
answer the question why. It suggests that people perform attributional searches for
causes of most negative events because such occurrences tend to arouse more causal
retribution than positive ones (Swanson Kelly, 2001).
Attribution theory focuses on causal references along the dimensions of locus, stability
and controllability with different implications for recovery satisfaction. The locus
dimension is split into internal and external locus based on who or what is viewed as
responsible for the negative incident. The relevance of locus of causality is in
identifying the cause of a problem and guiding on where to direct corrective action
(Ford, 1978). In the case of service failure, understanding the locus of causality may
support the decision on service recovery strategy.
The stability dimension explains whether the cause is viewed as temporary or
permanent and indicates the relative duration attached to a cause (Weiner, 2000). The
significance of stability dimension relates to its implications on future performance
expectations. Stability attributions tend to motivate change. Controllability dimension
relates to whether the cause of a service failure is perceived to be within the service
provider's control or not (Swanson Hsu, 2011).
Controllability ranges on a continuum from high to low and reflects the perception that
the service provider is able to influence the cause. Uncontrollable causes relate to what
the service provider has no ability to determine or prevent while controllable causes
are preventable by the organization. The significance of controllability is based on its
motivating influence on service managers to take action to restore perception of control
29
(Ford, 1985). Prior studies have suggested that there is attribution bias towards
controllable factors in situations where there is ambiguity or doubt.
The theory of attribution has been used in the context of recovery satisfaction to explain
how customers' perception of the cause of service failure influences satisfaction or
dissatisfaction (Swanson Kelly, 2001). Recovery satisfaction appears to be more
affected by stable causes as customers tend to be more forgiving when a service failure
occurs infrequently. When customers attribute a failure to a stable cause, they tend to
expect recurrence and they conclude that the organization is aware of potential
recurrence of such failures and they should have policies and procedures in place to
manage the issue (Wirtz Mattila, 2004). Accordingly, when a failure occurs
repeatedly, customers will be more frustrated and less forgiving as they might assume
that the service provider is aware of the problem that led to the service failure but has
not bothered to address it.
In a study on the impact of perceived justice and intention to complain (Hocutt et al.,
1997) concluded that customers who have service failures they attribute to their own
mistakes satisfactorily corrected perceive higher justice and indicate higher levels of
satisfaction than those who attribute the failure to the service provider. The main
criticisms for attribution theory relate to its simplistic assumption that human beings
evaluate their experience rationally and that their actions are influenced by causal
inferences (Folkes, 1984). Questions have also been raised regarding its failure to
consider the likely influence of social-cultural and historical factors on locus of
attribution. Another concern relates to failure to consider the risk of fundamental
attribution error.
30
2.2.3
Expectancy Disconfirmation Theory
Another relevant theory for this study is the expectancy disconfirmation theory which
has been used over the years to explain customer satisfaction/dissatisfaction with the
initial service delivery as well as in recovery situations (Oliver, 1980). The theory seeks
to explain post-purchase satisfaction as a function of expectations, perceived
performance, and disconfirmation of beliefs. Its core focus is the idea that customers
compare perceived service performance to expectations. If a performance meets
customer expectations it is confirmed; one which fails to meet expectations is
negatively disconfirmed; and one that exceeds customer expectations is positively
disconfirmed (McCollough et al., 2000).
Based on the disconfirmation theory, satisfaction is conceptualized as an individual's
feeling of pleasure or displeasure resulting from comparing a perceived service
performance to expectations whereby the performance is perceived to fall below, match
or exceed expectations leading to dissatisfaction, satisfaction and delight respectively
(Oliver, 1980; Kotler Keller, 2012). The confirmation/disconfirmation process
entails three stages of evaluation which start with a prior formation of expectations,
followed by a comparison of service experience and expectations and concludes with
a satisfaction/dissatisfaction judgement based on the size of the gap between
expectations and performance (Donna, 1986). Perceived performance is then
summarily rated as better, same, or worse that than expected.
Service recovery situations are viewed as creating a new service loop which then
becomes subject to recovery disconfirmation (Magnini et al., 2007). From a
disconfirmation perspective, customer satisfaction with a service recovery experience
is influenced by both performance and how that performance measures up to
31
expectations. Spreng (1996) proposed that overall satisfaction is comprised of two
dimensions based on attribute and information satisfaction where attribute satisfaction
refers to a customer's evaluation of the recovery experience itself while information
satisfaction is based on satisfaction with the information on which the expectations
were based.
The expectancy disconfirmation theory implies that satisfaction is related to size and
direction of disconfirmation. Although the disconfirmation paradigm has become the
dominant theory in the assessment of customer satisfaction, criticisms have been
focused the suitability of assessing expectation versus value. The theory is linked to
cognitive dissonance in that it relates to situations where observations are not aligned
to beliefs (Oliver, 1997).
2.2.4
Cognitive Dissonance Theory
The dissonance theory suggests that human beings seek consistency in their beliefs and
perceptions. A discrepancy between the two leads to significant psychological
discomfort which the individual seeks to minimize in order to restore balance
(Festinger, 1957). Individuals attempt to minimize cognitive dissonance by focusing
on factors that are in harmony with their expectations or by adjusting the conflicting
perceptions so that they are consistent with other beliefs or behaviors.
This implies that a customer who expects a certain standard of service and perceives a
deviation from that expectation will engage in a process of cognitive repositioning to
reduce dissonance (Oliver, 1980). Hence customers who experience dissonance
constantly engage in perceptual distortion in an effort to bring harmony between their
perception and performance. Some critics of this theory argue that even when
32
presented with overpowering evidence to the contrary, many people will still not
change their behavior.
Sweeny et al. (2000) found that consumers who are highly satisfied were less dissonant
than those who were satisfied and concluded that dissonance is more associated with
dissatisfaction than satisfaction. Cognitive dissonance theory has been used in
marketing to explain customer dissatisfaction arising from the disparity between
expectation and actual service experience. Carrel and Dittrich (1978) suggested that
individuals experience greater cognitive dissonance in situations of inequity than those
in equitable situations.
When there is a discrepancy between beliefs and behavior something must change in
order to eliminate or reduce the dissonance. The growing competition in service
businesses challenges service providers to anticipate dissonance and facilitate
customers to minimize it by providing information to reinforce customer decisions in
order to enhance satisfaction.
2.3
Perceived Justice and Recovery Satisfaction
The concept of perceived justice and its effect on recovery satisfaction has been the
subject of scholarly research and practitioner's debate over the years. Perceived justice
with service recovery interprets the customer's assessment of the fairness of an
organization's service recovery strategy and impacts satisfaction. In a study on the
role of recovery satisfaction in the relationship between service recovery and brand
evangelism (Rashid Ahmad, 2014) acknowledged the influence of perceived justice
on customers as they form evaluative judgment on service situations involving conflict.
33
The concept of fairness is based on social psychology, and has been extensively used
to examine individual responses to service encounters involving failure and recovery
(Ellyawati et al, 2012). The justice perspective is considered a suitable basis for
gaining insights on customer evaluation of service encounters involving failure and
recovery because it has consequences on post recovery attitudes and behavior. The
importance of fairness evaluation in the recovery satisfaction judgement is associated
with the interpretation that customer suffers a deficit (Oliver, 1980) following service
failure and may therefore seek redress through service recovery with a view to
obtaining restoration of justice.
Previous studies have reported a relationship between perceived justice and recovery
satisfaction in a wide range of settings including hotels, restaurants, airlines, retail and
mobile telecommunications services (Blodgett et al., 1997; Spark McColl-Kennedy,
2000; Kau and Loh, 2006; Nibkin et al., 2010; Ellyawati et al., 2012). Many researchers
have found that perceived justice in the handling of complaints has implications for
customer satisfaction as well as post recovery behavior (Tan, 2014; Smith
Mpinganjira, 2015).
Negative consequences such as spreading negative word-of-mouth communications,
increasing complaints and switching to competitors have been associated with
perceived injustice in service recovery. Extant literature supports three dimensions of
justice, namely distributive, procedural, and interactional justice. To establish the gap
between customer expectations and perception of service recovery satisfaction, Smith
et al. (1999) proposed a model for assessing encounters involving failure and recovery
that integrated the three dimensions of perceived justice. The study showed that service
34
recovery influenced customer satisfaction indirectly through the perceived justice of
the process, the nature of interactions and the final outcome.
Procedural justice dimension relates to the perceived fairness of the processes and
principles of service recovery. Since customers who have suffered a service failure are
frustrated, they are likely to be more alert during service recovery and will pay extra
attention to the fairness of policies and procedures applied in a conflict situation
(Zeithaml et al., 1993; Kau and Loh, 2006). Accordingly, the process of recovery is
expected to be fair in terms of application of policies and procedures to all customers
who experience a similar problem (Rio-Lanza et al, 2009).
The service personnel ability to listen attentively to the customers' viewpoint during
recovery is an important part of process fairness which impacts recovery satisfaction.
The evaluation of procedural justice involves not only recovery strategy applied but
also the speed of the process (Kelley et al., 1993). It is therefore possible that a customer
who is satisfied with fairness of procedure may be dissatisfied with the speed of
implementation, particularly if they deduce that the recovery could have been
accomplished more rapidly.
The perception of procedural justice also implies that the application of the recovery
procedure takes into account the interests of all parties concerned and is consistent and
unbiased. Previous findings indicate that recovery satisfaction is associated with the
perception that the procedure is based on accurate information and ethical principles
(Blodgett et al., 1997). Further, recovery satisfaction may be impacted by procedures
applied in service failure resolution without reference to whether it meets the desired
35
outcome. Rio-Lanza et al., 2009 concluded that procedural justice is the most important
predictor of recovery satisfaction.
Interactional justice interprets customers' perception of fairness of the behavioral
element during recovery process and impacts satisfaction. The manner in which the
customer is treated by the frontline staff during the service recovery process affects
recovery satisfaction (Tan, 2014). An assessment of interactional justice includes the
attitude of the service organization's frontline personnel with reference to the
politeness, courtesy and consideration with which they handle interactions with the
customers during the recovery process.
Blodgett et al., (1997) proposed that interactional justice is demonstrated by honesty,
clear explanation, hospitality, sensitivity, attentiveness, empathy and openness. In a
study of service recovery in restaurants in the United States Namkung and Jang (2009)
found that interactional justice played an important role in satisfaction and customer
retention. Similar findings were reported by Collie et al. (2000) in a study on the
hospitality industry where they concluded that perceived interactional justice impacted
the level of recovery satisfaction.
Although speed of recovery may enhance satisfaction by demonstrating that the
company cares for the customer's time, the delivery might be too fast as to deny
employees the opportunity to send interactional cues of courtesy and politeness which
are critical to interactional justice (Chebat and Slusarczyk, 2005). Furthermore,
Zeithaml et al. (1993) pointed out that customers do not view issues of fairness in
service failure and recovery from an economic perspective only, because the
intangibility of services heightens customers' emotional sensitivity.
36
In service recovery assessment, customers may rely heavily on the conduct and
attitudes of the organization's frontline staff (Tan, 2014) particularly because service
failure tends to amplify their attention. From an interactional justice perspective,
satisfaction is influenced by the perception of fairness of the interpersonal relationship
and communication during recovery. This is interpreted based on perception of respect,
courtesy and consideration (Davidow, 2003). Affiliation cues such as a smile and
attentiveness from the contact personnel can augment interactional justice perception
and impact recovery satisfaction.
Ellyawati et al. (2012) proposed that since customers tend to feel helpless following
service failure, an effective redress approach should seek to minimize negative feelings
during service recovery in order to enhance customer satisfaction. The ability to
honestly explain the cause of a service error in the process of recovery demonstrates
accountability and impacts recovery satisfaction. Ellyawati et al. (2012) concluded that
perceived justice contributed to the recovery satisfaction in a retail setting.
Distributive justice has been mentioned in the literature as a sig Distributive justice has
been mentioned in the literature as a significant element of the fairness evaluation. It is
inferred from equity theory which proposes that distributive fairness is a critical factor
in the evaluation of an exchange relationship (Nibkin et al, 2010). Distributive justice
focuses on the perceived fairness of outcomes in service recovery with an emphasis on
compensation and my affect recovery satisfaction in a more direct way because it is the
most tangible dimension of justice (Smith Bolton, 2002) making customer
assessment less complicated.
The distributive justice principle implies that customers subconsciously conduct a cost-
benefit analysis based on their inputs and the outcomes from the service provider. The
37
evaluation of fairness influences recovery satisfaction and positive post recovery
behaviors such as repurchase and loyalty (Kau Loh, 2006). In a study on how
emotions mediate the effect of perceived justice on loyalty (Chebat Slusarczyk,
2005) concluded that distributive justice is the most effective means of improving
customer satisfaction in situations involving recovery.
Moreover, other studies by Tan (2014) and Chepkwony et al. (2012) have shown
positive effects of distributive justice on recovery satisfaction. Komunda and
Osarenkhoe (2012) in a study on effects of service recovery on satisfaction and loyalty
observed that higher levels of redress independently increase positive customer
response. A study by McCollough et al. (2000) found that post-recovery satisfaction
is predicted by both distributive and interactional justice dimensions. The distributive,
procedural, and interactional justice dimensions have been found to influence recovery
satisfaction in dissimilar ways.
The procedural, interactional and distributive justice dimensions have different effects
on customer evaluation of service recovery. Whereas several studies demonstrate a
positive relationship between perceived justice and customer satisfaction with service
recovery, the findings with reference to the influence of the specific justice elements
have been inconsistent. For instance, in a study on the influence of distributive,
procedural, and interactional justice on post recovery behavior in the retail sector in
North America, Blodgett et al. (1997) concluded that interactional and distributive
justice were more critical in explaining satisfaction than procedural justice.
Smith et al. (1999) in developing a model of recovery satisfaction following service
failure concluded that customers assign a higher fairness value to distributive and
38
procedural justice than interactional justice. On the other hand Tan (2014) concluded
that the distributive justice dimension was the most important of the three in predicting
recovery satisfaction. Furthermore, a study by McCollough et al. (2000) which
investigated customer satisfaction with service recovery concluded that both
interactional and distributive justice components had similar importance in recovery
satisfaction. However, the study did not factor the influence of procedural justice in the
service recovery assessment.
A study by Kau and Loh (2006) investigated the perception of justice in service
recovery and how it affects recovery satisfaction in the mobile phone industry in
Singapore and found that distributive justice was a major factor in the satisfaction
judgement while fairness of process even though significant had a lower impact. The
study further established that dissatisfied complainants showed a lower level of trust
and were more likely to engage in negative word-of-mouth behavior. Further, Smith
Mpinganjira (2015) studied the banking sector in South Africa and concluded that the
three justice dimensions have a positive effect on recovery satisfaction with the
distributive element having the greatest influence.
Although each of the three dimensions of justice is viewed as uniquely important, it is
generally agreed that the blending of the three dimensions determines the customer's
overall perception of fairness and subsequently satisfaction. Blodgett et al., (1997
noted that higher levels of procedural justice may compensate for lower levels of
distributive justice particularly because customers who do not receive the desired
outcomes may still be satisfied with the overall recovery effort if they feel that the that
the process used to determine the outcome was fair.
39
In a similar way, interactional justice and procedural justice may compensate for
weaknesses in each other. For instance if a customer perceives that the manner in which
they are treated by the frontline personnel indicates high respect and courtesy, they may
be willing to overlook a perceived weaknesses in the procedures. This apparent
compensatory behavior in the evaluation of justice dimensions emphasizes the
importance of assessing the overall justice perception.
In a study on perceived justice and behavioral intentions (Ha Jang, 2009) found that
the quality of recovery performance influences the extent of recovery satisfaction.
Further, in a study examining perceived justice, emotional responses and service
recovery, Rio-Lanza et al. (2009) observed that all the three justice dimensions affect
satisfaction, with procedural justice showing the strongest relative influence. However,
in a study on the effects of justice theory on service recovery satisfaction in Philippines,
Tan (2014) found overall support for the influence of perceived justice with different
contributions from the various dimensions.
There has been considerable debate on the nature of the relationship between perceived
justice and recovery satisfaction. Based on a study on effect of customer perceptions
of complaint handling (Maxham Netemeyer, 2002) observed that the procedures,
policies, methods and practices used by the service provider to recover a failed service
are important for satisfaction. While some studies suggest procedural justice is more
important than the outcome in service recovery situations, others differ arguing that the
outcome has the greatest influence on recovery satisfaction (Davidow, 2003). Tan
(2014) found interactional justice to have a weaker link to recovery satisfaction than
the distributive dimension. This corroborated an earlier finding by Nibkin et al. (2010).
40
Previous studies have suggested that the level of influence of each justice dimension
may depend on type of service, nature or severity of the service failure, customer
relationship with the organization and subculture (Smith et al., 1999; Chang et al,
2008). Ghalandari et al. (2012) investigated the effects of perceived justice on post
recovery satisfaction and revisit intention in the Iranian airline industry and concluded
that only interactional justice influenced recovery satisfaction.
Furthermore, different studies indicate dissimilar effects of the dimensions on post
recovery behavior and attitudes. The relationship between the justice dimensions and
recovery satisfaction is rather complex leading to contradictory findings mainly in
relation to which dimension or combination of dimensions has greater influence on
satisfaction. There was, therefore, a need of conducting a rigorous study to explicitly
assess this relationship.
2.4
Perceived Justice, Service Failure Attribution and Recovery Satisfaction
The relationship between perceived justice, service failure attribution and satisfaction
is of interest to scholars and marketing practitioners. Service failure attribution entails
the allocation of blame for the service failure and has been found to influence
perception of fairness as well as the recovery satisfaction assessment (Wirtz Mattila,
2004; Swanson Hsu, 2011). Weiner (2000) indicated that service failure experiences
motivate customers to engage in causal attributions along the dimensions of locus,
stability and controllability. In a study on service failure attribution and firm reputation
(Nibkin et al., 2011) found that both attribution and perceived justice influence the level
of satisfaction with service recovery.
In a study on attribution of service failures and satisfaction in Spain, Iglesias (2009)
concluded that attribution to a service provider causes a systematic reduction on all
41
quality perception measures far beyond the service aspects related to the failure. Wirtz
Mattila (2004) concluded that stability attributions associated with recurring failures
impacts satisfaction negatively. As the number of failures increase, customers tend to
apply stability attributions and therefore intensify the blame on the provider.
In another study on the impact of perceived justice and intention to complain (Hocutt
et al., 1997) found a link between fairness perception, service failure attribution and
recovery satisfaction. Based on a study on service failure and social identity, Weber
and Sparks (2010) submitted that customer evaluations of service recovery are
influenced by external attribution in the context of alliances or partnerships whereby a
service failure attributed to the lesser known entity or the outside partner is rated more
harshly.
From an attribution perspective service failure represents a breakdown in the delivery
of a service with the consequence that customers attempt to explain the occurrence or
to allocate blame. A service failure activates a recovery effort which involves an
economic and social interaction between the customer and the service provider through
which an outcome is allocated to the customer (Hess et al., 2003). The service recovery
effort is evaluated based on attribution and the perceived fairness of the recovery.
Service failure is associated with emotional reactions of anger, resentment and
frustration which lead to behavioral responses of complaining and switching (Sparks
Fredline, 2007). These negative emotions influence customer's perception of justice
as well as service failure attribution. Service failures lead to dissatisfaction, and
customers are inclined to allocate blame even as they expect the company to undertake
a series of activities to restore them to satisfaction. Service failures that are not properly
42
addressed attract more attributional search and are likely to affect satisfaction
negatively. Keaveney (1995) found that switching behavior was commonly driven by
service failure experiences and perception of unfairness in the recovery process.
Taking into consideration the intricacy of services particularly due to intangibility, and
simultaneousness of production and consumption, service failures can vary greatly
(McCollough et al, 2000). This diversity makes it somewhat problematic to classify
service failures even as categorization is important in understanding attribution
processes. Proponents of service failure classification argue that it is useful in
comprehending customer's satisfaction judgement in a recovery situation. Service
literature has classified failures based on severity, criticality and service type (Smith et
al., 1999; Weun et al. 2004; Levesque et al., 2000).
A more severe failure will lead to a different customer response than a minor one
(Weun et al., 2004). The severity of failure relates to the perceived harm or loss caused
to the customer due to the service problem experienced and has an effect on recovery
satisfaction. Service failure ranges from simple to complex with different levels of
damage including those that cause a mere irritation and those that disrupt a customer's
life (Smith et al., 1999). Indeed, it has been noted that even the service occasion can
influence the customer's grievance and affect attribution search, justice perception and
recovery satisfaction.
Service failure attribution may be impacted by the perspective of process because the
dynamics involved in delivery including the frontline staff behavior, abilities and
empowerment may influence on attribution as well as the success of recovery. Prior
studies have indicated that customer satisfaction may be impacted by the stage of
43
service delivery when the failure happens as well as their evaluation of specific
performances (Michel, 2001).
Service failure attribution and recovery satisfaction are also influenced by the number
and frequency of failures. Customers are more likely to attribute the first failure to
accidental causes which are not within the service provider's control, and expect only
moderate recovery while repeated failures are more likely to be attributed to provider's
incapacity or indifference (Maxham Netemeyer, 2002). Zeithaml et al. (1993) noted
that repeated negative encounters can lead to heightened expectation of future
occurrences with the consequence that customers apply stability attribution and report
lower satisfaction.
Given that negative encounters are salient and more memorable, recurring failures tend
to be assigned to stability attribution and are likely to increase dissatisfaction. The
complainant will expect higher recovery efforts in order to compensate the long term
inequity associated with multiple service failure (Wirtz Matilla, 2004). Conversely,
effective recovery performance is associated with attributions of controllability and
may send a signal to customers to adjust their expectations upwards, thereby creating
a challenge in the event of multiple failures (Maxham and Netemeyer, 2002).
Tax et al. (1998) indicated that most customers tend to attribute blame and seek redress
when the service failure related to very important services and where the severity of
failure is high. Recovery satisfaction is impacted by the customer perception that the
harm from service failure is high as it becomes difficult to provide a fair and appropriate
redress because of increasing recovery expectations (Maxham Netemeyer, 2002).
Consequently, recovery satisfaction is expected to be lower for service failures
associated with high severity.
44
Controllability attribution has implications for perceived justice and recovery
satisfaction. The perception that a service failure arose because the service organization
has invested inadequate resources for service delivery will cause greater dissatisfaction
among customers as they expect the organization to have this under control (Nibkin et
al, 2011). When the customer perceives that the service provider was dishonest in the
service process, an ethical issue arises in the evaluation which is attributed to the
service provider. Notably, such occurrences may heighten the perception of injustice
in service recovery.
In a study on the impact of perceived justice and intention to complain (Hocutt et al.,
1997) found a link between fairness perception, service failure attribution and recovery
satisfaction. When a failure that customers attribute to their own mistakes is
satisfactorily corrected customers are likely to indicate higher justice and satisfaction
levels than when they attribute the failure to the service provider. This is also related
to lower expectation for redress when service failure is internally attributed.
Nibkin et al. (2011) concluded that stability and controllability attributions moderated
the relationship between perceived justice and recovery satisfaction suggesting that the
lower the stability and controllability of service failure, the stronger the positive
relationship between perceived justice and recovery satisfaction. A study by Mattila et
al. (2001) revealed that customer values impacted the evaluation of service recovery as
well as failure attribution perception. The results indicated that attribution processes
influence recovery satisfaction through their impact on customer perceptions of the
recovery effort. Kanousi (2005) found that evaluation of service recovery varies across
45
cultures with reference to attribution as well as the role of tangibles in the recovery
process.
The specific dimensions of service failure attributions have different implications for
recovery satisfaction. Hess et al. (2003) concluded that customer attributions about the
service failures they experience influence their satisfaction as well as behavioral
responses toward the service provider. Swanson and Hsu (2011) examined the effect
of locus of attribution and service failure severity on repurchase behavior and observed
that attribution has implications for the level of consumer dissatisfaction, likelihood to
make a complaint to the firm and intention to spread negative word of mouth about the
service provider.
Harris et al. (2006) used an experimental design and considered the effect of shopping
medium on customers' attributions of blame for a service failure in the airline and
banking industries and found that online customers have lower expectations of service
recovery than offline customers because they frequently accept blame for most service
failures. The study concluded that shopping medium (online/offline) had an
intervening effect on expected service failure recovery.
Wirtz and Mattila (2004) found that all the perceived justice dimensions have an
influence on consumers' attribution processes as well as their recovery satisfaction.
However, they concluded that there is a variation across sectors with reference to
consumers' stability and controllability attributions. Smith and Bolton (2002) reported
mixed findings for the influence to of stability attribution on recovery satisfaction. This
study sought to add to the existing literature on the relationship between perceived
justice, service failure attribution and recovery satisfaction.
46
2.5
Perceived Justice, Disconfirmation and Recovery Satisfaction
Expectancy disconfirmation that is, the discrepancy between recovery expectation and
performance has been found to influence overall satisfaction as well as perceived
justice after a service failure. From a disconfirmation perspective service failure
represents a breakdown in the delivery of a service where the service fails to meet
customer expectations. In a study on critical factors of service failure and recovery
(Krishna et al., 2011) found that service recovery efforts influence customer
expectation and satisfaction.
Seawright et al. (2008) examined the elements in service recovery design and observed
that customer expectation is influenced by the magnitude of service failure in the initial
service encounter. The study submitted that recovery disconfirmation has an important
influence on recovery satisfaction. In a study investigating the role of customer
relationship with service provider in recovery situations, (Hess et al., 2003) established
that customers with higher expectancy of relationship continuity reported higher
satisfaction.
Expectancy disconfirmation shows that customer satisfaction increases when a
provider's recovery response exceeds expectations. Justice theory surmises that
recovery satisfaction is higher when customers perceive fairness in the provider's
response to rectify the service failure and particularly with reference to procedures,
interaction and outcome. Prior research has shown that interactional, distributive and
procedural justice elements have an influence on customers' disconfirmation and
satisfaction evaluations during a service failure and recovery encounter (Rio-Lanza et
al., 2009; Smith Mpinganjira 2015).
47
Oliver (1980) suggested that desirability of an occurrence is an important consideration
in the satisfaction evaluation, arguing that perceived performance is rated as
satisfactory in the event of positive disconfirmation of a desirable event or negative
disconfirmation of an undesirable event. Further, Spreng (1996) conducted a consumer
experiment which demonstrated that desires congruency and expectations congruency
influenced customer satisfaction. This implies that an individual will be dissatisfied
when faced with negative disconfirmation of a desirable event or positive
disconfirmation of an undesirable event.
Expectancy-disconfirmation and equity provide different explanations of satisfaction
in situations involving recovery. In a study on justice based recovery expectations (Yim
et al., 2003) integrated equity and expectancy disconfirmation and demonstrated that
recovery satisfaction increases when the providers exceeds customer expectations for
a fair failure recovery response. This study indicated that customers form recovery
expectations based on distributive justice and procedural justice and then use these as
a standard against which they evaluate service recovery performance.
Service failure may lead to dissatisfaction arising from a perception of injustice. The
service provider is expected to resolve the failure quickly and effectively to restore
customer satisfaction. Since the customer has to spend time and effort or even
emotional energy in order to seek redress, the service failure and recovery process
infers an exchange process with the attendant recovery expectations. The customer
expects the service recovery to match the loss incurred as a result of the failed service
(Tan, 2014) and this has implications for recovery disconfirmation and satisfaction.
48
Perceived justice, recovery disconfirmation and satisfaction are impacted by service
failure severity because it has been found to influence customers' perception of loss,
as well as expectations for recovery (Weun et al., 2004). Customers interpret severity
of service failure based on the perceived loss incurred as a result of the occurrence with
the consequence that more severe failures influence fairness, recovery expectations and
satisfaction evaluation. Higher magnitude service failures imply higher expectations
for recovery efforts than lower magnitude failures (Levesque et al., 2000) with
implications for recovery satisfaction.
Recovery disconfirmation and satisfaction are also affected by service failure criticality
because of its influence on recovery expectation. Criticality refers to the importance
the customer attaches to the service sought hence the importance of a successful service
delivery at a specific point (Chen, 2013). Furthermore, Ostrom and Iacobucci (1995)
found that customer recovery expectations and evaluation of service recovery were
affected by the importance of the occasion of purchase. Webster and Sundaram (1998)
examined the effect of service criticality on recovery satisfaction and concluded that it
determines the level of recovery effort required to restore satisfaction.
Ostrom and Iacobucci (1995) suggested that criticality has a different effect on
recovery disconfirmation and evaluation based on the nature of the service, particularly
whether they are credence or experience services. Credence services require a longer
consumption time frame for effective evaluation while experience services are easily
evaluated shortly after consumption. Credence services include tax consultancy,
trichology and financial investment, while experience services include restaurant, hotel
and mobile money transfer services. In a study of various types of experience and
credence services, Ostrom and Iacobucci (1995) found that criticality influenced
recovery satisfaction for both types but had a greater effect on credence services.
49
Expectations are a critical component in the recovery disconfirmation judgment
because they serve as the reference point or standard against which the customers
evaluate service recovery. Specific service attributes, prior experience and
organizations marketing communication have an influence on customer service
expectations (Kotler Keller, 2011). A study by Yim et al. (2003) which integrated
expectancy disconfirmation with perceived justice reported that customers formed
justice based expectations based on their desired or ideal recovery approach. Further
the study noted that customers reported greater satisfaction when they perceived that
the service recovery effort exceeded their expectations.
Based on disconfirmation paradigm, complaining behavior is driven by the need to
reduce the discrepancy between service expectations and performance related to the
service failure. Oliver (1997) proposed that customer complaints are motivated by the
need to reduce deficit associated with the discrepancy between expectation and
performance. As such poor service recovery leads to downward adjustment of future
recovery expectations, with the risk of customer switching owing to anticipated future
injustice. Since multiple failures present a threat to service providers, service recovery
strategies should incorporate systems for preventing their re-occurrence.
A fairness perception of a service recovery is associated with customers' expectation
that the service provider will apply recovery policies uniformly and consistently in
similar situations (Smith Bolton, 2002). Extant literature suggests that service failure
occurrences are associated to negative feelings which lead to heightened expectations
(Ellyawati et al., 2012). Customers with higher expectations are more likely to use a
50
more elaborate process in analyzing the recovery with the consequence that both
procedural and interactional justice elements may be found wanting.
Prior research has shown that expectations are important in recovery evaluation
because service failure elicits a perception of injustice which can lead to psychological
tension, dissonance or anger (Andreassen, 2000). According to Oliver (1997) negative
feelings are an important part of the satisfaction judgment because emotions coexist
alongside cognitive evaluations. Further, Chebat and Slusarczyk (2005) submitted that
customer perception of injustice could influence recovery disconfirmation because it
generates more intense feelings of resentment.
Given the influence of recovery expectations on satisfaction, customers with lower
expectations are likely to show greater satisfaction after experiencing above average
recovery performance while those with higher expectations would report lower
satisfaction (Hess et al., 2003). Prior research suggests that exceptional recovery efforts
can minimize the dissatisfaction associated with failure to the extent that the service
provider may be fully exonerated (Zeithaml et al., 1993).
In situations where the recovery fails to meet or exceed customer's expectations the
outcome is greater negative disconfirmation and dissatisfaction. This is associated with
the double deviation effect (Maxham Netemeyer, 2002) which arises from a double
failure, that is the failure of the initial service and a second failure in the recovery
attempt. Moreover, in cases of extremely severe failures where the damage done is
too much, even a successful recovery may have less effect on satisfaction (Smith
Bolton, 2002). The negative service evaluation becomes magnified when a double
deviation ensues leading to high recovery dissatisfaction.
51
Customer expectations are based on information from various sources including service
provider marketing communications and sales claims with regard to service capabilities
(Kotler Keller, 2012). Since customers expect honesty and reliability from the
service organization, the disappointment when the delivery does not match the
promises made or implied leads to negative disconfirmation (Andreassen, 2000). The
dissatisfaction arising is based on the failure to meet performance expectations and the
offence of providing inaccurate or misleading information.
In a study on disconfirmation and switching Chang et al. (2012) concluded that both
initial and recovery disconfirmation influence recovery satisfaction. The study
suggested that the satisfaction or dissatisfaction explained by disconfirmation paradigm
is a judgement that is transaction specific. Given that service recovery naturally starts
with failure, customer begins the recovery encounter with a deficit based on the initial
dissatisfaction. As such customer's satisfaction evaluation of service may be impacted
by the disconfirmation with initial service experience as well as with the recovery
disconfirmation.
The disconfirmation paradigm infers that satisfaction that is greater than expectations
is positively influenced by superior recovery performance. The concept of recovery
paradox has been used in the literature to refer to a situation where the customer is more
satisfied after a recovery than they were satisfied before the service failure (Magnini et
al., 2007). Andreassen (2000) found that negative feelings associated with the service
failure do not have a significant impact on satisfaction with service recovery,
suggesting that recovery paradox is contrast driven based on the perception of a
successful recovery as a superior experience.
52
McCollough et al. (2000) indicated that when the service provider delivered a very
successful recovery that exceeded customers' recovery expectations they reacted with
positive surprise. The strong positive feelings associated with positive recovery
disconfirmation give the customer a sense of being delighted (Andreassen, 2000). This
feeling of pleasant surprise from superior service recovery goes beyond restored
satisfaction and can inspire customer loyalty (Nibkin et al., 2010; Tan, 2014). This
positive surprise is consistent with the service recovery paradox, which implies that
customers indicate higher satisfaction and repurchase intentions after a service failure
is successfully resolved than before the failure occurred (Smith and Bolton, 2002).
Despite the expression of positive surprise, McCollough et al. (2000) concluded that
consumers would rate a company better for avoiding service failure than for superior
recovery. Moreover, they found that initial disconfirmation had a greater influence on
satisfaction than recovery disconfirmation hence the supposition that initial service
performance is the main predictor of initial disconfirmation, while recovery
performance is the main predictor of recovery disconfirmation.
The recovery paradox effect appears to be supported when the recovery effort is able
to completely assuage the harm caused by a service failure. Magnini et al. (2007)
researched the service recovery paradox and concluded that it is more likely to occur
when the service failure is associated with low severity and the cause is deemed to be
unstable and non-controllable. Smith et al. (1999), provided evidence in support of
a recovery paradox effect, but emphasized that it was contingent on achieving
consistently high-service recovery performance. Prior research has reported mixed
findings on the possibility a recovery paradox effect with some concluding that it can
occur, while others disagree (McCollough et al., 2000).
53
Extremely negative disconfirmation is likely to occur in cases where failures are caused
by service sabotage that is, staff behaviors that are intentionally designed to affect
service negatively (Harris Ogbonna, 2006). Service sabotage represents a wide array
of deliberate actions taken by service personnel calculated to sabotage customers'
service experience. Although service sabotage affects customers directly, it is
commonly targeted at the service organization itself or other coworkers. The calculated
nature and immediacy effects associated with the sabotage of a service encounter imply
sudden negative customer experiences with the consequence of negative
disconfirmation and dissatisfaction. Service sabotage is likely to lead to double
deviation since the employee who deliberately causes service failure is unlikely to
facilitate successfully recovery.
Previous research has shown that disconfirmation has a dominant impact on recovery
satisfaction. Based on a study of customer satisfaction after service failure and recovery
McCollough et al. (2000) suggested that low expectations of service recovery tend to
encourage positive disconfirmation while high expectations encourage negative
disconfirmation. Previous research has shown that the direction and size of the
discrepancy between perception of service and the prior expectations influence
customer satisfaction as well as perception of fairness of the recovery attempt (Smith
Bolton, 2002).
Andreassen (2000) observed that disconfirmation and perceived justice are
conceptually distinct and can be considered as complementary drivers of satisfaction
further indicating that, in contrast to disconfirmation where the satisfaction evaluation
is related to expectations before corrective action is taken, equity is a relative dimension
54
which results in comparing normative standards of performance. Accordingly, the
dissatisfaction caused by initial service failure is associated with negative feelings
which may have a carry-over effect on the recovery process leading to poor satisfaction
judgment.
In a study on satisfaction with service encounters involving failure and recovery Smith
et al. (1999), concluded that both disconfirmation of expectations of service recovery
and perceived justice of the outcome, have an influence on satisfaction. Further,
although they observed that disconfirmation has the lesser influence in predicting post-
recovery satisfaction when compared to perceived justice, they found it to be more
important than the interactional dimension of justice. Similarly, McCollough et al.
(2000) examined airline passengers' satisfaction with service recovery through a
scenario-based experiment and concluded that recovery satisfaction is a function of
recovery disconfirmation together and perceived justice.
Some authors have suggested that decline in memory of expectations may influence
disconfirmation. For instance, Andreassen (2000) observed that the effect of
disconfirmation on recovery satisfaction may be credited to a halo effect arguing that
complaining customers bring overall positive and negative biases to the recovery
evaluation task. There was therefore need to conduct a comprehensive study to clearly
investigate this relationship.
2.6
Perceived Justice, Service Failure Attribution, Disconfirmation and
Recovery Satisfaction
Previous scholars have demonstrated a positive relationship between perceived justice
and recovery satisfaction with others proposing that such a relationship is contingent
55
on other intermediary factors including service failure attribution and disconfirmation
(Smith Bolton, 2002; Nibkin et al., 2011). Recovery satisfaction is a central
determinant of future customer behavior and is critical for retaining customers and
safeguarding a profitable business. As such, service providers need to develop
operational systems for taking corrective action whenever service failure occurs and to
ensure that customers perceive the process, interaction and outcome of recovery to be
fair.
From on the foregoing discussion, it is clear that previous scholars have recognized the
effect of perceived justice on recovery satisfaction. However, the dynamic forces in
this relationship are not conclusive due to varying perspectives and areas of
concentration of the various studies. Prior studies have concentrated on different
variables that influence the relationship between perceived justice and recovery
satisfaction making comparison problematic. For instance, some have focused on
various attitudinal and behavioral outcomes such as re-purchase intention, loyalty,
recommendation, word of mouth communication and trust.
Prior studies have considered the mediating effect of corporate image (Nibkin et al.,
2010); and recommendation behavior (Serenko Stach, 2009); locus of attribution
(Swanson Hsu, 2011); emotions (Ellyawati et al., 2012); cross-cultural influences
(Wang Matilla, 2011) and brand evangelism (Rashid Ahmad, 2014). Further,
others have emphasized the influence of the nature, severity and criticality of service
failure (Ostrom Iacobucci, 1995; Smith et al., 1999; Weun et al., 2004) and recovery
disconfirmation (McCollough et al., 2000) on perceived justice and recovery
satisfaction.
56
Regarding service failure attribution, prior studies report that stability and
controllability elements influence recovery satisfaction (Hocutt et al., 1997; Weber
Sparks, 2010; ) while others place greater emphasis on the locus of attribution (Kau
Loh, 2006). Prior studies reported that attribution to the service provider greatly erodes
all quality perceptions of the service (Iglesias, 2009) while attribution to the customer
(self-attribution) minimizes expectations for recovery (Harris et al, 2006). Service
failure attribution was found to have a mediating influence in the relationship between
perceived justice and recovery satisfaction (Nibkin et al., 2011).
Concerning recovery disconfirmation, though some studies have demonstrated its
influence in the relationship, some have emphasized the severity or criticality of service
failure in the initial service encounter (Smith et al., 1999; Weun et al., 2004) while
others have focused on recovery expectations (Andreassen, 2000). Furthermore, since
those who experience service failures are likely to be more observant than those in an
initial service situation, they are presumed to have more amplified expectations thus
magnifying the role for disconfirmation in a recovery situation.
Despite the fact that many studies have examined the relationship between perceived
justice and recovery satisfaction, their conclusions have been diverse leading to a lack
of clarity on the relationship. Further, the role of causal attribution and recovery
disconfirmation in this relationship, have not been investigated in equal measure. A
summary of knowledge gaps is presented in Table 2.1.
57
Table 2.1: Summary of Knowledge Gaps
Authors
Focus of
the study
Research
Method
Main
findings
and
conclusions
Gaps
Focus of
the
proposed
study
Maxham
and
Netemeyer
(2002)
Examined
the effects
of
perceived
justice (PJ)
on recovery
satisfaction
(RS) and
intent
among
customers
of a South
Eastern US
bank.
Longitudinal
survey of 208
customers
who actively
complained
about the
quality of
banking
services.
RS partially
mediates
the effects
of PJ on
positive
word-of-
mouth
(WOM)
intent and
purchase
intent.
Did not
consider the
relative effects
of PJ
dimensions or
the influence of
(SFA) and
recovery
disconfirmation
(RD) on RS.
The
proposed
study
examines
the relative
dimensions
of PJ and
the role of
RD and
SFA in the
relationship
between PJ
and RS.
Lio-Ranza
et al. (2009)
Investigated
the relative
effects of
the
dimensions
of PJ on RS
and the
emotions
triggered by
service
recovery in
Spain.
Descriptive
study of
cellphone
customers
who had
experienced a
recovery in
previous
year. Used
convenience
sample of
554.
All three
justice
dimensions
affect
satisfaction,
with
procedural
justice
showing the
strongest
relative
influence.
The study did
not factor the
influence of
RD or SFA.
Examines
the role of
RD and
SFA in the
relationship
between PJ
and RS.
Weber and
Sparks
(2010)
The study
assessed the
effects of
SFA and
social
identity in a
strategic
airline
alliance
context in
Australia.
Experimental
scenarios
approach.
Random
sample of
313 travelers
assigned four
scenarios to
rate.
Locus of
attribution
affects
consumers'
post-
recovery
behavior.
Speed of
service
recovery
has an
influence on
RS
What is the
effect of RD on
RS? Used
experimental
scenarios
which may not
capture the
emotions
generated in an
actual service
recovery event.
The
proposed
study also
examines
the role of
RD and uses
a survey of
real service
failure/
recovery
experiences.
Nibkin et al.
(2010)
Investigated
the
relationship
between PJ,
Descriptive
survey of
customers
who had
Distributive
justice and
interactional
justice have
The study was
limited to PJ
and RS and did
not consider
Examines
the role of
RD and
SFA in the
58
corporate
image and
RS among
airline
passengers
in
Malaysia.
experienced
service
failure using
a
convenience
sample of
118.
a positive
relationship
with RS.
role of RD and
SFA
individually or
collectively.
Study used
convenience
sample.
relationship
between PJ
and RS.
Uses
random
sample.
Ellyawati et
al. (2012)
This study
examined
the effect of
PJ on RS
and the
influence of
emotions
on justice
among
retailers in
Indonesia.
Descriptive
survey of 102
retailers who
had
experienced
service
failure in the
previous
year.
PJ
significantly
affects RS.
Positive
emotion
mediates
the effect of
PJ on RS.
Did not factor
the role of SFA
and RD. Used
retrospective
reporting
requiring recall
over a long
period.
The role of
SFA and
RD will also
be
investigated.
Limits recall
period to 6
months to
minimize
recall bias.
Komunda
and
Osarenkhoe
(2012)
Examined
the
relationship
between
service
recovery,
satisfaction
and loyalty
in
commercial
banks in
Kampala,
Uganda.
Descriptive
survey of 120
banks
customers
among
students and
staff of
Makerere
University
Business
School.
Service
recovery
impacts on
RS. Higher
levels of
redress
influence
customer
responses.
Did not
investigate the
role of
perceived
justice, SFA, or
RD in the
recovery
satisfaction
judgment.
Examines
the role of
PJ, RD and
SFA in
customer's
evaluation
of
satisfaction.
Chepkwony
et al. (2012)
The effect
of
distributive
justice
dimension
of PJ on
customer
satisfaction
in the
banking
industry in
Eldoret,
Kenya.
Descriptive
survey of 372
customers
using a self-
administered
questionnaire.
Distributive
justice was
found to
have a
significant
influence on
satisfaction
in a service
recovery
situation.
Did not
examine the
role of
procedural and
interactional
dimensions of
justice and did
not factor the
influence of
SFA and RD in
the
relationship.
Additionally
examines
the role of
procedural
and
interactional
justice as
well the
influence of
RD and
SFA
Tan, T.A.
(2014)
Assessed
the
influence of
PJ on RS
Descriptive
survey of 288
customers
recruited
The effect
of
distributive
justice on
Did not
investigate the
role of SFA
The study
incorporates
the role of
SFA and
59
on dine-in
experiences
in Metro
Manila,
Philippines.
through
purposive
and snowball
sampling.
recovery
satisfaction
is stronger
than
interactional
justice.
and RD in the
RS judgment.
RD in the
relationship
between PJ
and RS.
Mayombo,
M. B.
(2014).
Examined
the
influence of
customer
complaint
behavior,
firm
responses,
and service
quality on
customer
loyalty of
mobile
telephone
subscribers
in Uganda.
Descriptive
survey of 336
respondents
selected
through
stratified
random
sampling.
The joint
effect of
customer
complaint
behavior,
firm
responses
and service
quality on
customer
loyalty was
statistically
significant.
Did not
investigate the
influence of
perceived
justice, SFA or
RD on RS.
This study
investigates
the role of
PJ, RD and
SFA in
customers'
evaluation
of a service
failure and
recovery
encounter.
Smith and
Mpinganjira
(2015)
Studied the
role of
perceived
justice in
recovery
satisfaction
and
behavioral
intentions
among
banking
customers
in South
Africa.
Descriptive
survey of 281
retail banking
customers
who had
reported a
service
failure six
months
before the
study.
A self-
administered
structured
questionnaire
was used.
Found that
procedural,
interactional
and
distributive
dimensions
of PJ
influence
recovery
satisfaction
and
behavioral
intentions.
The study did
not factor the
role of SFA
and RD in the
recovery
satisfaction
judgement.
This study
examines
the
relationship
between PJ
and RS and
the role
played by
RD and
SFA in that
relationship.
Source: Current Author
60
2.7
Conceptual Framework
The literature reviewed here above revealed that the relationship between perceived
justice and recovery satisfaction may be influenced by many factors including service
failure attribution and disconfirmation. However, the nature of the influence of service
failure attribution and recovery disconfirmation in that relationship is unclear.
Conceptual gaps in the literature have raised questions regarding the treatment of the
variables that influence the satisfaction judgment in a service recovery. This study
developed a conceptual model (Figure 2.1) based on the envisaged inter-relationships
among the four variables.
Figure 2.1: Conceptual Model
Moderating Variable
Independent Variable
Dependent Variable
Mediating variable
Source: Current Author
Service Failure
Attribution
-Stability
-Controllability
-Locus
Perceived Justice
-Procedural Justice
-Interactional Justice
-Distributive Justice
Recovery Satisfaction
-Service encounter
satisfaction
-Overall satisfaction
-Repurchase intention
Recovery
Disconfirmation
-Expectations
-Performance
H1
H3
H2
H4
61
Based on the literature reviewed and objectives of this study the conceptual model
depicted in Figure 2.1 was configured around the hypothesized relationships between
the independent variable (perceived justice) and the dependent variable (recovery
satisfaction), as well as the moderating variable (service failure attribution) and
mediating variable (recovery disconfirmation). It also presents the indicators for each
of the variables.
Recovery satisfaction was expected to be directly influenced by perceived justice (H1).
It was hypothesized that recovery disconfirmation mediates the relationship between
perceived justice and recovery satisfaction (H2) while service failure attribution was
expected to moderate that relationship (H3). Lastly, it was hypothesized that perceived
justice, recovery disconfirmation, service failure attribution jointly influence recovery
satisfaction (H4).
2.8
Conceptual Hypotheses
The study had four hypotheses based on the expected relationships between the
variables. In line with the literature reviewed and the relationships depicted in the
conceptual model in Figure 2.1, the following hypotheses were framed and tested:
H1: There is a significant relationship between perceived justice and recovery
satisfaction.
H2:
Recovery disconfirmation has a significant mediating influence on the
relationship between perceived justice and recovery satisfaction.
H3:
Service failure attribution significantly moderates the relationship between
perceived justice and recovery satisfaction.
H4: The joint effect of perceived justice, service failure attribution and
disconfirmation on recovery satisfaction is statistically significant.
62
CHAPTER THREE
RESEARCH METHODOLOGY
3.1
Introduction
This chapter provides a detailed description of the methodology that was deployed in
conducting the study. It presents the research philosophy, research design, population
of the study and the sampling plan. The chapter further describes the relevant data
collection method and operationalization of research variables. Finally, it explains the
measures of reliability and validity as well as how the pertinent data was analyzed.
3.2
Research Philosophy
Stevens et al. (2006) highlighted the main elements to consider in choosing a research
philosophy by reviewing the principal features of the two paradigms namely,
positivism and constructivism (phenomenological). Positivism is based on the belief
that the world is external, the observer is independent and that science is value free. It
views the role of the researcher as to focus on facts, look for causality, and formulate
and test hypotheses. The researcher plays the role of an objective analyst to analyze
quantitative data collected and produce the results in line with the objectives of the
research (Saunders, 2003). The preferred methods include the use of large samples and
operationalizing concepts so that they can be measured. Positivism is concerned with
theory (hypotheses) testing.
Phenomenological paradigm holds the belief that the world is socially constructed and
subjective. The observer is part of what is observed and that science is driven by human
interests. Accordingly, the researcher should focus on meanings and seek to understand
what is happening while looking at the totality of the situation and developing ideas
63
through induction. The preferred methods include the use of small samples and
triangulation, which involves the use of multiple methods to establish different views
of phenomena. The main emphasis is on depth of investigation. Phenomenological
paradigm is therefore concerned with theory building.
Given the research problem investigated in this study, positivism approach was
appropriate as it applies a quantitative approach to the investigation of social objects
and aims to identify causal explanations while developing generalizations regarding
the issues under study. According to Cacioppo et al. (2004), positivism is suitable in
studies involving systematic observation and description of phenomena contextualized
within a model, the presentation of hypotheses, and the use of inferential statistics to
test hypotheses. The current study satisfied these requirements.
3.3
Research Design
The research design for the current study was a descriptive cross-sectional survey
which involved data collection from a population or a representative sample at one
specific point in time without manipulating the environment (Stevens et al., 2006). The
descriptive design is used when the purpose is to describe what exists and to uncover
new facts and meanings. It is used to provide information about naturally occurring
behavior, attitudes and other characteristics associated with a study population
(Bryman, 2012). It was considered suitable for this study as it sought to provide an
accurate and valid representation of the factors and processes pertaining to the research
question. Further, it effectively supported the gathering of data from a large population
about prevailing conditions for the purpose of description and interpretation.
Cross sectional design was appropriate for this study as it sought to establish
associations among different variables based on a single interaction with respondents
64
(Aaker et al., 2007). It simplified data collection from a diverse population and enabled
patterns of convergence to emerge hence facilitating the interpretation of the
relationships between the study variables. It was suitable for the purpose of the current
study which sought to establish the relationship between perceived justice, service
failure attribution, disconfirmation and recovery satisfaction. This design was used
successfully in similar studies (Tan, 2014; Mayombo, 2014; Ndungu, 2012).
3.4
Population of the Study
The population of the study comprised all mobile service subscribers who were
registered for MMTS services with companies providing money transfer services in
Kenya under the Mobile Network Operators (MNO) led model. These were users of
M-Pesa, Airtel Money and Orange Money. There were over twenty five million MMTS
subscribers registered with MNOs in Kenya by June 2014 (CA, 2014). This study
focused on customers who obtain services directly from MNOs. While there are other
mobile money providers such as those using the mobile virtual network operator license
(MVNO), they operate on a different model and were not included in this study. The
composition of the population is presented in table 3.1.
65
Table 3.1: Description of the Population of the Study
Mobile Money Transfer
Service (MMTS) providers
Registered MMTS
subscribers
Percentage of the
population
Safaricom M-Pesa
19,776,056
78
Airtel Money +Yu
5,385,893*
21.3
Telkom Orange Money
185,463
0.7
Total
25,347,412
100
Source: Based on Communications Authority of Kenya, Quarterly Sector Statistics Report
(2014).
*Yu Cash customers were under Airtel management.
The target population comprised users of MMTS who had experienced a service
failure/recovery encounter within the previous six months. The period of six months
was considered appropriate for minimizing recall bias. Previous studies on recovery
satisfaction have used a recall period of six months to one year (Ellyawati et al, 2012;
Tan, 2014).
In a study on financial inclusion in Kenya, Intermedia (2014) using a random sample
of 3000 users found that 50% of the mobile money subscribers sampled had
experienced a service failure related to network or agent system breakdown with 24%
experiencing such a problem at least three times in a six month period. However, there
was no comprehensive list available for the customers who had sought service recovery
in the last one year. As such, screening questions were used to select respondents to
participate in the study.
66
3.5
Sample Design
The study focused on the two main providers of MMTS in Kenya namely, Safaricom
and Airtel. This was justified because the two MNOs controlled more than ninety nine
percent of mobile money transfer subscriber's accounts. A proportionate random
stratified sampling technique was used to ensure representativeness based on the
number of registered subscribers with each provider. This was justified by the disparity
of subscriber concentration within the two networks. Thus, stratified sampling ensured
that estimates and comparisons could be made with equivalent precision. The sample
was distributed on the basis of the customer base of each MNO.
Simple random sampling was used within each stratum. A final sample of 803
respondents was realized for this survey against a target of 784 proposed in accordance
with the recommendations of Krejcie and Morgan (1970) when the population is larger
than 100,000 (Appendix 3). This sampling approach was used based on the assumption
that each observation was independent of all the others and that the margin of error
indicated was acceptable for this type of research. This Table 3.2 contains the sample
structure.
67
Table 3.2: Sample Design
Mobile
money
transfer
service
(MMTS)
providers
MMTS
subscribers
registered with
MNOs
Sample
proportion of
the
population
(%)
Targeted
sample
size
Achieved
Sample
%
M-Pesa
19,776,056
78.6
616
629
78.3
Airtel Money
+ Yu Cash
5,385,893
21.4
168
174
21.7
Total
25,161,949
100
784
803
100
Source: Based on Communications Authority of Kenya, Quarterly Sector Statistics Report
(2014).
Since there was no complete list available for customers who had experienced service
failure and sought service recovery in the last six months, screening questions were
used to select respondents for the study. The questions asked the respondents in the
contact sample first; whether they had experienced any service failure with their mobile
money transfer service in the previous six months and second; if they had reported the
same to the service provider with expectation of recovery. The interview was
terminated for those who responded negatively and continued for those who responded
positively.
3.6
Data Collection
Primary data was collected using a recall based survey. This was suitable for this study
as it sought to focus on actual customer service failure and recovery encounters. This
approach required respondents to answer questions with reference to past occurrences
and is commonly used in service failure/recovery studies. A semi-structured
questionnaire in both English and Kiswahili languages (appendix 1a and 1b) was
68
considered appropriate for this study because of ease of administration to a large
number of respondents spread out nationally as well as the nature of the data sought.
The data was collected from customers of MMTS who had experienced a service
failure/recovery encounter within the previous six months. The respondents were
designated using a computer assisted number management system from a list of
numbers allocated by Communication Authority to each MNO. The participants to the
survey were selected through screening questions in section 1 of the questionnaire.
These questions sought to identify those who had experienced a service
failure/recovery encounter within the preceding six months.
The questionnaire was administered through telephone interviews. A Computer
Assisted Telephone Interview (CATI) system was used. This is a front-end telephone
interviewing system which enables an interviewer to ask questions over the phone
while keying in the responses into the computer. Interviewers were trained on how to
conduct the interviews. This method was considered suitable for this study as it used
mobile phone which is the same platform consumers use to obtain mobile money
services; to complain to the service provider; and to seek service recovery.
CATI is recommended for studies with a short questionnaire using multiple choice
formats as was the case with this study and is commonly used in marketing research
(Stevens et al., 2006). The interviews for the current study took about fourteen minutes.
Chebat and Slusarczyk (2005) collected data for a study on the influence of emotions
on perceived justice in service recovery in Canada through a telephone survey that took
fifteen minutes using CATI system. This study included the use of Likert-type scale
ranging from `strongly agree' to `strongly disagree'. Mobile phone-based surveys have
69
been used in Kenya by the International Federation of the Red Cross and Crescent
Societies in surveys on health issues, disaster management and water and sanitation
(IFRC, 2012).
Stevens et al. (2006) observed that the use of CATI improves data quality as it allows
for interviewing and data entry simultaneously as well as automatically controlling
questionnaire branching and skipping. Further, it supports random sampling from a
large population as it uses an automated number management system for selecting and
dialing phone numbers. The system allows for greater standardization as it prompts the
interviewer with the next question hence minimizing interviewer bias. In the event of
nonresponse or failure to cooperate, the system searches for the next customer or
replacement automatically based on set rules (Bassili Fletcher, 1991; Bryman, 2012).
The study questionnaire was divided into seven sections. The first section covered
screening questions to determine whether the respondent experienced a service
failure/recovery encounter with the previous six months. Section two captured data on
the nature of service failure/recovery encounter with MMTS. Sections three, four and
five were used to gather data on service failure attribution, perceived justice, and
disconfirmation respectively. Section six was used to elicit data on respondent's overall
satisfaction with the recovery. Section seven was used gather data on the respondent's
profile.
3.6.1
Reliability of the Research Instrument
Reliability refers to the degree to which a measurement technique can be expected to
secure consistent results upon repeated application. Reliability of the research
instrument relates to the extent to which it is expected to yield consistent results in
70
repeated trials assuming no major changes in conditions surrounding the measurement
(Carmines Zeller, 1979). The questionnaire was pre-tested with 52 MMTS customers
through a telephone survey using computer assisted telephone interviewing system
(CATI). This was meant to address any difficulties in understanding the statements or
responding to questions.
To assess internal consistency of the research instrument, a common measure of
reliability, Cronbach's Alpha, was used (Sekaran, 2003). Cronbach's alpha is
frequently used as an estimate of reliability. It is an internal consistency index
considered suitable for assessments with no right or wrong answer such as those asking
respondents to agree or disagree with a statement on a particular scale (Carmines
Zeller, 1979) as was the case with the questionnaire for this study.
The higher the internal consistency of the items, the closer the Cronbach's Coefficient
Alpha is to 1. The lower the internal consistency of the items, the closer the Cronbach's
Coefficient Alpha is to zero. Reliability of the four variables in research instrument was
established at above the cut-off point of 0.7 which is recommended as the minimum
acceptable level by Nunnally Bernstein (1994). Previous studies have used
Cronbach's alpha coefficient to measure reliability (Nibkin et al., 2010; Kinoti, 2012;
Ndungu, 2013).
3.6.2 Validity of the Research Instrument
Validity refers to the degree to which the research instrument measures what it is meant
to measure. It is concerned with the extent to which the assessment tool actually
measures the underlying outcome of interest. Carmine and Zeller (1979) assert that
validity can be assessed by use of expert opinion or judgment. The content validity of
71
the instrument was enhanced by the use of measurement items that have been used by
other scholars in previous studies and have been confirmed as the appropriate measures
of the behavior or values that were being measured.
In addition, the questionnaire was reviewed by university professors who are experts
in service marketing with the aim of improving its content validity. Following revision,
it was pre-tested with 52 customers from the study population using telephone survey
with a view to identifying any ambiguities after which the final questionnaire was
finalized.
Factor analysis using principal component analyses was applied to test for validity. The
questionnaire was then revised and piloted with 52 respondents through a telephone
survey. This was meant to identify and remove any ambiguities. The questionnaire was
then modified based on the pilot study results and then used to collect data from the
larger sample. Factor analysis using principal component analyses was applied to test
construct validity by identifying underlying variables, or factors, that explain the
pattern of correlations within the observed variables. An assessment tool has construct
validity if it exhibits an association between the test scores and the prediction of the
underlying attribute of interest.
Principal components analysis was used because the primary purpose was data
reduction. Before performing principal components analysis, the fitness of data for
factor analysis was assessed using Kaiser-Meyer-Olkin (KMO) which produced a value
of 0.918 exceeding the recommended value of 0.6 and Bartlett's Test of Sphericity
reached statistical significance supporting the factorability of the data. Factorial
72
evidence has been used in previous studies to determine the validity of the research
instrument (Ellyawati et al., 2012; Ndungu, 2013; Njeru, 2014).
3.7
Operationalization of the Study Variables
Perceived justice was operationalized based on the literature with a categorization of
three dimensions of procedural, interactional and distributive justice (Smith et al.,
1999). Service failure attribution was operationalized as in the literature to include
stability, controllability and locus based on customer perception (Swanson Kelly,
2001). Recovery disconfirmation was operationalized through expected performance
versus perceived performance (Oliver, 1997). Recovery satisfaction, the dependent
variable, was operationalized as transaction satisfaction, overall satisfaction and
intention to repurchase (Davidow, 2000; Maxham Netemeyer 2002).
Multi-item measures were adapted from previous studies and slightly modified to fit
the current context and research setting and used to measure each of the variables in
the study. The items in all scales were measured using a five point Likert-type scale
ranging from `not at all' (1) to `to a very large extent' (5) except for recovery
satisfaction where the terms `Not at all satisfied' to `very much satisfied' were applied.
Perceived justice dimensions were measured using items adapted from the study by
Davidow (2003) and Nibkin et al. (2010). The attributes for assessing recovery
disconfirmation were adapted from Andreassen (2000) and Magnini et al. (2007).
Measures for assessing service failure attribution were adapted from Swanson and
Kelly (2001) and Hocutt et al. (1997) while the measures of recovery satisfaction were
from Maxham and Netemeyer (2002) and Davidow (2003). Table 3.3 summarizes the
operationalization of the variables.
73
Table 3.3: Operationalization of the Study Variables
Variable
Natur
e
Indicator
Specific measure Suppo
rting
Literat
ure
Scale
Questi
on No
Perceived
Justice
Independent
-Procedural
justice
-Interactional
justice
-Distributive
justice
-Accessibility,
speed of service,
fairness of
practices.
-Courtesy,
empathy,
apology,
listening,
information
- Fairness of
solution, reward
and outcome.
Nibkin
et al.,
(2010);
Davido
w
(2003)
Rating
scale
5 point
Likert
type
Sectio
n IV
Q5
Recovery
Disconfirma-
tion
Media
ting
- Expectations
- Perceived
performance
Performance:
-Better than
expected
-Same as
expected
-Worse than
expected
Magnin
i et al.
(2007);
Andrea
ssen
(2000).
Rating
scale
5 point
Likert
type
Sectio
n V
Q6
Service Failure
Attribution
Moderating
-Locus
-Stability
-Controllability
Whether service
failure:
-was caused by
service provider
or customer
-is expected to
recur
-could have been
avoided
Swanso
n and
Kelly
(2001);
Hocutt
et al.,
(1997).
Rating
scale
5 point
Likert
type
Sectio
n III
Q7
Recovery
Satisfaction
Dependent
-Transaction
satisfaction
-Overall
satisfaction
-Repurchase
intention
-Recommen-
dation
intention
-An assessment
on the recovery
transaction
-An overall
assessment
service
- Willingness to
use the service
from the same
provider again
- Willingness to
recommend the
service provider
Maxha
m and
Neteme
yer
(2002);
Davido
w
(2003)
Rating
scale
5 point
Likert
type
Sectio
n VI
Q8
Source: Current Author
74
3.8
Data Analysis
The data collected was cleaned, edited, coded and then analyzed using descriptive and
inferential statistics and reported in line with the objectives of the study. The Statistical
Software Package for the Social Sciences (SPSS) version 20 was used to facilitate the
analysis. The descriptive statistics including mean scores, standard deviation,
percentages, and frequencies were used to describe and summarize characteristics of
the variables under study and to provide evidence about the central tendency and
dispersion of data.
Factor analysis using principal component analysis was conducted before testing the
analytical model. This helped to reduce data into smaller set of factors that explain most
of the variance observed. Tests of assumptions including linearity of data, normality
of distribution, outliers, homogeneity of variance and multicollinearity were conducted
to confirm that the assumptions of parametric data were acceptable.
Pearson Product Moment Correlation (r) was used to examine the relationships
between variables, particularly to explore the direction and strength of the relationship
between variables. Further, the correlation matrix was useful in determining whether
multicollinearity existed before doing further analysis. This was important as
multicollinearity which occurs when independent variables are highly correlated (r=0.9
and above) leads to a poor regression model (Green et al., 1988).
Since the study sought to determine the influence of the independent variable
(perceived justice) on the dependent variable (recovery satisfaction), and the
moderating and mediating effects of recovery disconfirmation and service failure
attribution respectively, regression analysis was used to test the hypothesized
75
relationships. The F-test was used to test overall model fit while T test was computed
to test individual parameters. For each hypothesis, the coefficient of determination (R
2
)
was used to measure the amount of variation between the variables. Regression
coefficient () was computed to determine prediction level of regression models and
assess the relationship between the independent variable and dependent variables.
To test for moderation a step-by-step process in line with the proposition of Baron and
Kenny (1986) and Fairchild and MacKinnon (2009) was followed. Hierarchical
regression analysis was used to assess the moderating influence of service failure
attribution on the relationship between perceived justice and recovery satisfaction. In
testing for moderating effect, an interaction term was created and entered in the
regression equation to assess whether the moderator alters the strength of the
relationship. A moderator variable alters the strength of the causal relationship. While
moderation typically implies a weakening of a causal effect, a moderator variable can
magnify or even reverse that effect (Baron Kenny, 1986).
Figure 3.1 presents the moderation path diagram.
Figure 3.1: Moderation Path Diagram
Z
H
3
X Y
X
1
Z
2
Y
XZ
3
76
X=Dependent variable; Z=Moderator variable; XZ=the Interaction term.
Source: Fairchild and MacKinnon (2009)
The regression coefficient for the interaction term XZ provides an estimate for
moderation effect. If 3 is statistically different from zero, there is a significant
moderation effect between perceived justice and recovery satisfaction.
To test for mediation effects, a step by step process in line with the proposition of Baron
and Kenny (1986) and Fairchild and MacKinnon (2009) was followed. Hierarchical
regression analysis was conducted and the significance of coefficients in each step
examined. The last step involved multiple regression analysis with the independent
variable (PJ), mediating variable (RD) and dependent variable (RS). A mediating
variable transmits the effect of an independent variable on a dependent variable and it
represents the addition of a third variable to the X Y relation, whereby X causes the
mediator, M, and M causes Y (X M Y).
Figure 3.2: Model for Testing Mediating Effect
Figure 3.2 presents a graphical representation of the model for testing mediating effect.
In the model, the variable X is the causal variable perceived justice while the variable
Y is the outcome, recovery satisfaction while path c is the total effect. The figure below
reflects the un-mediated model.
Part A: Overall Direct effect Unmediated model
Path c
X
Y
77
The effect of X (perceived justice) on Y (recovery satisfaction) may be mediated by
variable M (recovery disconfirmation). The mediated model is presented below.
Part B: Path diagram for the mediation effect
M
X
c'
Y
Source: Fairchild and MacKinnon (2009)
The mediation model implies a causal relationship where the mediator is presumed to
cause the outcome and not the other way round. Path c' is referred to as the direct effect
while the mediator is an intervening between M. Where variable X no longer affects Y
after M has been controlled and hence path c' is zero then there is complete mediation.
Partial mediation is the reflected where the path from X to Y is reduced in absolute size
but is still different from zero when the mediator is introduced. Table 3.4 represents a
summary of the relevant analytical models used in the study.
78
Table 3.4: Summary of Analytical Methods and Interpretation
Objectives
Hypotheses Analytical method
Hypothesi
s testing
Interpretation
Objective 1:
Determine the
influence of
perceived
justice on
recovery
satisfaction
among
subscribers of
mobile money
transfer
services in
Kenya.
H1: There is
a significant
relationship
between
perceived
justice and
recovery
satisfaction.
RS = + 1X1 + e
= + 11X1 + 12X2
+ 13X3 + e, where:
RS represents
dependent variable
recovery satisfaction;
= regression constant
(intercept); 11, 12,
13 represents
regression coefficients
and X1, X2, X3 =
perceived justice (PJ)
dimensions
X1=Procedural justice,
X2 =Interactional
justice, X3=
Distributive justice; e
is the error term.
Regression
analysis
Beta coefficient to
determine
prediction level of
models.
Zero = no
relationship;
If positive =
positive
relationship;
If negative =
negative
relationship.
R
2
to assess how
much of the
dependent
variable's variation
is due to its
relationship with
the independent
variable.
Reject hypothesis
if p.05
Objective 2:
Assess the
mediating
effect of
recovery
disconfirmatio
n on the
relationship
between
perceived
justice and
recovery
satisfaction.
H2:
Recovery
disconfirmat
ion has a
significant
mediating
influence on
the
relationship
between
perceived
justice and
recovery
satisfaction.
Testing for mediating
effect:
Step 1: Test the direct
relationship between
PJ and RS
RS = + 21PJ + e
Step 2: test if PJ
predicts recovery
disconfirmation (RD)
RD = + 31PJ + e
Step 3: test if RD
predicts RS
RS = + 41RD+ e
If the relations is
significant proceed to
step 4
Step 4: multiple
regression with PJ and
RD predicting RS
RS == + 51PJ +
52RD + e
Hierarchic
al
regression
analysis
If both coefficients
for paths a, and b
are significant,
then M mediates
the relationship
between X and Y
and cl is assessed
to check the link
strength.
Where variable X
no longer affects Y
after M has been
controlled and path
c' = zero then there
is complete
mediation. If the
path from X to Y is
reduced in absolute
size but is still
different from zero
when the mediator
is introduced then
there is partial
mediation.
Reject hypothesis
if p.05
79
Objective 3:
Establish the
moderating
influence of
service failure
attribution on
the
relationship
between
perceived
justice and
recovery
satisfaction.
H3: Service
failure
attribution
has a
significant
moderating
influence on
the
relationship
between
perceived
justice and
recovery
satisfaction.
Testing for moderation
effect:
RS= + 61PJ +
62SFA + cPJ*cSFA
+ e, where
=constant, =
regression coefficients,
cPJ*cSFA interaction
of PJ and SFA and e is
error term.
Hierarchic
al
regression
analysis
3 is the regression
coefficient for the
interaction term
and if it is
statistically
different from zero
then Z moderates
the relationship
between X and Y.
Reject hypothesis
if p.05
Objective 4:
Determine the
joint effect of
perceived
justice,
disconfirmatio
n and service
failure
attribution on
recovery
satisfaction.
H4: The
joint effect
of perceived
justice,
service
failure
attribution
and
disconfirmat
ion on
recovery
satisfaction
is
statistically
significant.
Testing for joint
influence
RS= + 71PJ+
72RD + 73SFA +e
Regression
analysis
Beta coefficients
determine the
strength of the
joint influence of
predictor variables
on the independent
variable. Zero
means no
relationship;
Positive means
positive
relationship;
Negative means
negative
relationship.
Reject hypothesis
if p.05
Source: Current Author
3.9
Chapter Summary
This chapter has explained the research methodology used to implement the study. The
chapter has explained the research philosophy, research design, study population and
sampling plan adopted for this study. It has also explained the method used for data
collection and the measures of reliability and validity. Further, it has documented the
operationalization of study variables and the statistical analysis techniques which
comprised descriptive statistics, correlation and regression analyses. Additionally, it
has summarized the analytical models used for data analysis and hypotheses testing.
The next chapter presents data analysis, findings and interpretation of results.
80
CHAPTER FOUR
DATA ANALYSIS, INTERPRETATION AND DISCUSSION
4.1
Introduction
The objective of this study was to investigate the relationship between perceived
justice, service failure attribution, disconfirmation and recovery satisfaction among
subscribers of mobile money transfer services (MMTS) in Kenya. This chapter
provides data analysis, findings and discussions of the results based on the objectives
of the study. Primary data was used for this research and was collected from a survey
of MMTS subscribers who had experienced a service failure and recovery encounter in
the six months preceding the survey. Questionnaires administered through telephone
were used to seek the respondents' opinion of service recovery with regard to perceived
justice, disconfirmation, service failure attribution, and recovery satisfaction.
4.2
Respondents' Screening and Response Statistics
The study targeted Subscribers of MMTS who had experienced a service failure and
recovery encounter within the preceding six months. Since there was no comprehensive
list of the defined population, a screening process was used to determine qualified
respondents from among a list of MMTS subscriber numbers. Screening questions were
used and qualified respondents were selected based on their responses. First,
respondents were asked whether they had experienced any service failure during the
past six months. The interview was terminated for those who responded in the negative.
Those who responded in the positive were asked a follow up question as to whether
they had reported the specific issue to the mobile money service provider with
81
expectation of a solution. Those who answered `no' were terminated while those who
answered `yes' were qualified to continue.
From a total of 2663 outgoing calls made to randomly selected numbers, 374 calls had
no response, 235 requested to be called back while 542 were screened out because the
respondents had only experienced service failure but not service recovery. Out of the
1512 qualified respondents who had experienced service failure and recovery, 709 did
not wish to be interviewed while 803 agreed to participate in the study and completed
the interview indicating a response rate of 46%. The final sample of 803 respondents
was satisfactory as it exceeded the target sample of 784 respondents targeted before the
study. Table 4.1 presents the response statistics.
82
Table 4.1: Study Response Statistics
OUTGOING CALL STATISTICS
No
%
Phone not answered (not answering,
switched-off, busy)
374
14.04%
Requested to be called back
235
8.82%
Experienced service failure but not recovery 542
20.35%
Experienced service failure and recovery
but did not wish to be interviewed
709
26.62%
Experienced service failure and recovery
and completed the interview
803
30.15%
Total Outgoing Calls
2,663
100%
RESPO
N
SE
RATE
Total qualified respondents *
1512
100%
Completed interviews / Response rate
803
46%
INTERVIEWS
English
332
41.3%
Kiswahili
471
58.7%
Total Interviews
803
100.0%
Average length of interview
13.42 Minutes
Source: Primary Data
*Made up of completed interviews and those qualified but did not wish to be interviewed.
Based on the digital minute counter, the interviews took an average of 13.42 minutes
which is comparable to a previous study on perceived justice in service recovery by
Chebat and Slusarczyk (2005) where the average telephone interview took fifteen
minutes.
83
4.3
Test of Reliability
Reliability refers to the consistency, stability or dependability of the data. Reliability is
a measure of the degree to which a research instrument yields consistent results or data
after repeated trials (Mugenda and Mugenda, 2003). Reliability was established
through computation of the Cronbach's alpha coefficient for each of the constructs of
data collection instruments (Sekaran, 2003). This study used a cut-off point of 0.7
which is recommended as the minimum acceptable level by Nunnally Bernstein
(1994). Table 4.2 indicates the reliability statistics for the four variables, that is,
recovery satisfaction, perceived justice, service failure attribution and recovery
disconfirmation.
Table 4.2: Reliability Analysis
Scale
Items
Cronbach's Alpha ()
Recovery Disconfirmation
3
0.972
Recovery Satisfaction
6
0.899
Perceived Justice
15
0.849
Service Failure Attribution
6
0.715
Source: Primary Data
All the variables tested for this study were quite reliable with a Cronbach's alpha
reliability coefficient greater than the pre-set cut-off point of 0.7. Perceived justice had
a Cronbach's alpha of 0.849. Recovery disconfirmation reported a Cronbach alpha of
0.972, service failure attribution 0.715 and recovery satisfaction 0.899 indicating a high
internal consistency.
84
4.4
Validity Tests
To assess validity a pilot study was conducted to help confirm the appropriateness of
the questionnaire with reference to comprehension and clarity. The questionnaire was
piloted with 52 respondents through a telephone survey in order to identify and remove
any ambiguities. Factor analysis using principal component analyses was applied to
test construct validity. Factor analysis was used to identify underlying variables, or
factors, that explain the pattern of correlations within the observed variables in the
perceived justice scale.
Principal components analysis was used because the primary purpose was to reduce the
information in the scale into a smaller set of components. These factors were used as
variables for further examination using multiple regression analysis to test the
hypothesis formulated in this study. Prior to performing principle components analysis,
the suitability of data for factor analysis was assessed. Inspection of the correlation
matrix revealed the presence of many coefficients of .3 and above. An assessment of
the suitability of perceived justice scale for factor analysis produced Kaiser-Meyer-
Olkin (KMO) value of 0.918, exceeding the recommended value of 0.6 and Bartlett's
Test of Sphericity reached statistical significance (P-Value0.05) supporting the
factorability of the correlation matrix.
The 15 items of the perceived justice scale were subjected to principal component
analysis using SPSS version 20. Principal components analysis revealed the presence
of 3 components with eigenvalues exceeding 1. The components accounted for 60.26%
of the cumulative variance in perceived justice. To aid in the interpretation of these
three components, Varimax rotation with Kaiser Normalization was performed. The
rotated solution revealed the presence of a simple structure, with the three components
85
showing a number of strong loadings of variables substantially on only one component.
Inspection of the rotated component matrix showed that interactional justice explained
44.4% of the variance in perceived justice, procedural justice contributed 8.5% and
distributive justice accounted for 7.4% of the variance in perceived justice. The detailed
results are presented in Appendix 4.
4.5
Respondents Characteristics
This section presents the respondents characteristics including demographics, service
failure experienced and usage information. On demographic characteristics, the study
sought information on county of residence, highest level of education, age and gender.
Usage information gathered included mobile money service provider and number of
years of usage. Concerning service failure, the study sought information on type of
service failure experienced and whether service recovery was sought.
4.5.1 Mobile Money Transfer Service Used
The respondents were asked to state which mobile money transfer service they used
most frequently. A majority of the respondents (78.3%) mentioned M-Pesa mobile
money transfer service while 21.7% of the respondents were using Airtel Money.
4.5.2
The Number of Years Respondent had used the Mobile Money Transfer
Service
The respondents had been asked to state how long they had subscribed to the mobile
money transfer service. Their responses are summarized in Table 4.3.
86
Table 4.3: Number of Years Respondent has used service
No of Years Respondent has used service
Frequency
Percentage (%)
Up to 6 months
14
1.7
7-12 months
36
4.5
1-5 Years
220
27.4
5 years
533
66.4
Total
803
100.0
Source: Primary Data
The results in table 4.3 show that a majority of respondents (66.4%) had used the
service for more than 5 years while another 27.4% had used the service for between 1
to 5 years. This is realistic as mobile money has matured in Kenya with over twenty
six million registered customers and subscription rates on the mobile network operator
model slowing down (CA, 2015). Most of the new subscriptions are with the mobile
virtual network operator licensees' such as Equity Bank.
4.5.3
Failures for which Service Recovery was sought by Respondents
The respondents said they faced many types of service failures during the six months
preceding the survey. They reported these problems through the designated phone line
of the MMTS provider and obtained service recovery. Most of the service failures
related to network error, customer error, agent failure and phone failure. The main
service failures for which recovery was sought by respondents are presented in Table
4.4.
87
Table 4.4: Failures for which Service Recovery was sought by Respondents
Frequency %
Receipt of mobile money transfers from strangers or conmen
204
28
Transaction taking long to confirm/delay in processing transaction
102
14
Sending money to the wrong number
95
13
Agent lack of cash/float
66
9
No notification message after transferring money to another
number or after buying airtime or bundles
51
7
Unavailability of mobile network service
51
7
Sending the same transfer twice unintentionally
44
6
No response when trying to use/ access the money transfer menu
37
5
Delayed reversal of funds (after wrong transaction)
29
4
Delay in confirming bill payment for KPLC, DSTV, Water etc.
29
4
Others
95
13
Total
803
100
Source: Primary Data
The three main failures namely, receipt of mobile money from strangers, delay in
processing transaction and sending money to wrong number constituted over half
(55%) of the problems for which customers sought service recovery.
4.5.4
Distribution of respondents by Region
The study sought information on the respondent's county of residence. The pertinent
results are summarized in Table 4.5.
88
Table 4.5: Distribution of Respondents by Region
Region
Frequency
Percentage (%)
Rift Valley
196
24.4
Coast
126
15.7
Nairobi
118
14.7
Eastern
100
12.5
Nyanza
97
12.1
Central
95
11.8
Western
57
7.1
North Eastern
14
1.7
Total
803
100.0
Source: Primary Data
As shown in Table 4.5, about half of respondents (54.8%) reside in three regions,
namely Rift Valley, Coast and Nairobi. Thus the use of telephone interviews allowed
for easy reach to respondents in a wide geographical area within a short period at
relatively low cost.
The survey obtained responses from MMTS users from 45 of the 47 counties including
those in far off areas of Kenya that are difficult to access by physical infrastructure.
These include Turkana, West Pokot, Lamu and Wajir counties. The distribution of
respondents by county is presented in Appendix 5.
4.5.5 Respondents' Highest Level of Education
Respondents were asked to state their highest level of education. The results are
presented in Table 4.6.
89
Table 4.6: Highest Level of Education
Highest Education Level
Frequency
Percentage (%)
Primary
164
20.4
Secondary
327
40.7
College/diploma
213
26.5
Graduate +
99
12.3
Total
803
100.0
Source: Primary Data
The results reveal that majority of respondents (40%) had obtained secondary
education, 26.5% had college/diploma education, while 20.4% had primary level
education. This is consistent with the choice of interview language as 59% of
respondents chose to be interviewed in Swahili while 41% opted for English language.
4.5.6
Age of Respondents
The study sought information on the age of respondents. The results are presented in
Table 4.7.
Table 4.7: Age of Respondent
Age in Years
Frequency
Percentage (%)
18-24
119
14.8
25-34
319
39.7
35-44
230
28.6
45+
135
16.8
Total
803
100.0
Source: Primary Data
90
The results indicate that most of the respondents (85.2%) were over 25 years of age
while 14.8% were between 18 and 24 years. This is consistent with other studies on the
age groups of those using financial services (CGAP, 2009).
91
4.5.7
Gender of Respondents
The study sought information on gender. The results presented in Table 4.8 reveal that
majority of respondents were male (56%). This is consistent with previous studies
which indicate that mobile money customers are mostly urban and male users
representing the typical early adopters (GSMA, 2014). In many households, male
members are the first to buy a phone and may even determine whether a woman uses a
phone. Additionally, women in low income groups are less likely to own phones than
men.
Table 4.8: Respondents' Gender
Gender
Frequency
Percentage (%)
Female
353
44
Male
456
56
Total
803
100.0
Source: Primary Data
Some studies have also shown that lack of mobile phone access hinders women's
adoption of mobile money while it has been demonstrated that women are more likely
to receive than send mobile money while senders tend to be mostly men. Mobile money
operators have tended to focus on the `active' side of the transaction (the senders), and
less on the `passive' side of the transaction, the recipients (GSMA, 2014). Furthermore,
senders of mobile money are more likely to report a service failure and seek redress
than intended receivers.
4.5.8
Respondents' Occupation
The study sought information on occupation of respondents. The results are presented in
table 4.9.
92
Table 4.9: Occupation of Respondents
Frequency Percentage (%)
Senior Management / Head of Department
40
5
Middle Level Manager
58
7
Professional e.g. Doctor, Lawyer, Engineer, Surveyor etc.
72
9
Skilled workers e.g. Driver, Mechanic, Tailor, Carpenter,
Mason etc.
157
20
Unskilled workers eg Cleaner, Watchman, Gardener,
Househelp
210
26
Clerical Levels
97
12
Unemployed
94
12
Others
75
9
Total
803
100
Source: Primary Data
The majority of respondents were in three occupational groups namely skilled,
unskilled and clerical level. The results are consistent with the level of education of
respondents.
4.6
Assessment of Perceived Justice
The study set out to establish the level of perceived justice among subscribers of mobile
money transfer services in Kenya. Before assessing its interaction with other variables
it was important to establish respondents' rating of perceived justice in MMTS service
recovery encounters in Kenya. The research adopted scales from the literature with
relevant modification to fit the context. Fifteen items were used to measure perceived
justice dimension. The respondents were asked to rate the statements on the perceived
justice scale. These items were measured on a Likerttype scale ranging from 1 = "not
at all" to 5 ="very large extent". A standard deviation of 1 or less was interpreted to
mean that respondents had consensus on the rating while that of greater than 1 was
93
taken to mean that respondents differed in their perception of the issue. The coefficient
of variation (CV) is a measure of dispersion which is useful for comparing the degree
of variation from one data set to another even when the measures differ drastically.
The respondents were asked to indicate the extent to which they perceived various
aspects representing the dimensions of procedural justice, interactional justice and
distributive justice based on the manner in which the service provider handled the
service recovery. The pertinent results are presented in Table 4.10.
Table 4.10: Descriptive Statistics for Perceived Justice Dimensions
Dimension of Perceived Justice
N
Mean Score Std. Deviation CV (%)
Procedural Justice
803
3.57
1.379
38.7
Interactional Justice
803
3.98
1.180
29.6
Distributive Justice
803
3.69
1.347
36.5
Perceived Justice
803
3.75
1.319
35.2
Source: Primary Data
The overall of mean score of 3.75 for perceived justice as presented in Table 4.10
indicates that respondents associated the service recovery with perceived justice to a
large extent. Justice perception is related to fairness. This means that mobile money
transfer service subscribers consider the approach followed by service providers in
service recovery to be appropriate and fair. With reference to the dimensions of
perceived justice, Table 4.10 reveals that interactional justice was rated highest with a
mean score of 3.98 followed by distributive justice with a mean score of 3.69 and
procedural justice with a mean score 3.57.
Table 4.10 indicates that the coefficient of variation for perceived justice was 35.2%
implying that the perceptions of respondents varied by 35.2%. The highest coefficient
94
of variation was for procedural justice (38.7%) and the lowest was for interactional
justice (29.6%) indicating greater consensus among respondents with regard to
interactions than with procedures used for service recovery.
4.6.1
Perception of Procedural Justice
Procedural justice concerns the fairness evaluation of processes and procedures
followed by the service provider to implement service recovery. It considers the
practices adopted by the company for facilitating service recovery including awareness
of how to access service recovery, ease of access, waiting time, queuing system and
speed of response. The results are presented in Table 4.10. The mean score of 3.57 for
procedural justice indicates that MMTS subscribers consider the procedures followed
in handling service recovery to be reasonable to a large extent.
The highest ratings under procedural justice dimension were for awareness of where to
lodge complaint (mean score = 3.90) and company practices for handling customer
problems (mean score = 3.90). Notably the fairness of queuing system and waiting
time were rated lowest with mean scores of 2.89 and 3.38 respectively. Previous studies
have noted subscriber concern regarding the issue of waiting for long before being
connected to the customer service line.
4.6.2
Perception of Interactional Justice
Interactional justice was measured by assessing respondent's perception of the nature
of interaction with service providers' customer care staff during a service recovery.
This component considers the perception of respect, concern, courtesy and
attentiveness. Interactional justice focuses on the nature of interaction with service
personnel once a subscriber gets through to the designated customer care line, separate
95
from the prior experience with reference to accessibility and queuing time. The results
are presented in Table 4.11.
96
Table 4.11: Descriptive Statistics for the Indicators of Perceived Justice
Perceived Justice
N Mean
Std.
Deviation
CV
(%)
Procedural Justice
It was easy to know where to lodge your complaint
803
3.90
1.399
35.9
You waited for a short time to before connecting
with the customer care staff
803
3.38
1.566
46.3
The company employee responded quickly to your
problem
803
3.78
1.372
36.3
The queuing system is fair
803
2.89
1.349
46.7
The company practices for addressing customer
problems are reasonable
803
3.90
1.212
31.1
Sub-total (average)
3.57
1.379
38.7
Interactional Justice
You were treated with respect
803
4.35
.928
21.3
The employee listened attentively to your
complaint
803
4.34
.932
21.5
The employee apologized for the occurrence
803
3.08
1.694
55.0
The employee provided a useful information about
the problem
803
4.02
1.226
30.1
The employee showed concern about the problem
803
4.12
1.119
27.2
Sub-total (average)
803
3.98
1.180
29.6
Distributive Justice
The way the problem was solved resulted in a
positive outcome for you
803
4.02
1.275
31.7
The solution provided was worth the effort you
made to seek correction
803
3.95
1.293
32.7
The company offered you what you deserved
803
3.98
1.252
31.4
The compensation for your problem was fair
803
2.51
1.693
67.5
The transaction was carried out in a fair manner
803
4.02
1.223
30.4
Sub-total (average)
803
3.69
1.347
36.5
Overall (Grand) Mean Score
803
3.75
1.319
35.2
Source: Primary Data
97
The mean score of 3.98 for interactional justice means that MMTS subscribers consider
the quality of interaction with service providers' frontline personnel during service
recovery to be appropriate and fair to a large extent. The highest ratings under
interactional justice dimension were on respectful treatment, employee attentiveness
and employee sensitivity with mean scores of 4.35, 4.34 and 4.12 respectively. Further,
the aspect of informational justice, that is, the usefulness of the information provided
during service recovery was also rated as fair to a large extent with a mean score of
4.02. The lowest rating under interactional justice was for employee apologized for
the occurrence with a mean score of 3.08. An apology is considered important in
service recovery because it helps address the negative emotions most customers
experience following a service failure (Smith Bolton, 2002).
4.6.3
Perception of Distributive Justice
Distributive justice was evaluated by assessing the perceived fairness of the solution
provided, compensation and overall outcome of the service recovery encounter. The
component focuses on redress and compensation for the service failure experienced
and considers the fairness of the solution provided, how the problem was handled, the
compensation offered and fairness of the outcome based on the customer's effort to
seek service recovery. The results are presented in Table 4.11.
The study established that MMTS subscribers considered distributive justice as fair to
a large extent based on a mean score of 3.69. The highest ratings under distributive
justice dimension were for positive outcome and the fairness in the manner the
transaction was carried out both with mean scores of 4.02. The respondents found the
solution provided to have been worth the effort to seek corrective action to a great
extent supported by a mean score of 3.95. Notably fairness of compensation offered
98
was rated the lowest with a mean score of 2.51. Compensation is considered important
in redressing service failure since complaining customers have invested time, effort
and emotions to seek recovery (Smith et al., 1999).
4.7
Assessment of Service Failure Attribution
Service failure attribution was measured by assessing the respondent's perception of
locus of attribution, stability and controllability. It focused on customer perceptions of
who was responsible for the service failure occurrence, the frequency of the occurrence
and service providers' ability to prevent the problem. The results are presented in Table
4.12.
The overall mean score of 2.61 indicates a moderate rating. Locus of attribution relating
to self and company obtained moderate rating with mean scores of 2.78 and 2.77
respectively while that relating to the agent had a low rating of 1.81 implying that
agents were not considered responsible for most service failures reported by respondent
to the service provider. This low mean score for agents is different from previous
studies where service failure caused by agents was viewed as a major concern (CGAP,
2009; Intermedia, 2014). Table 4.12 presents the detailed findings.
99
Table 4.12: Descriptive Statistics for the Indicators of Service Failure
Attribution
Item
N
Mean
Std.
Deviation
CV %
Service Failure Attribution
The company was to blame for the problem
803
2.77
1.751
63.2
You were to blame for the problem
803
2.78
1.707
61.4
This problem occurs frequently
803
2.35
1.484
63.1
The company could have prevented the
problem
803
3.32
1.607
48.4
The problem was caused by the agent
803
1.81
1.321
73.0
Overall
803
2.61
1.574
60.3
Source: Primary Data
Stability attribution measured in terms of frequency of service failure occurrence had
a moderate rating with a mean score of 2.35. Controllability attribution, which relates
to the subscriber's perception of the service provider's ability to prevent the problem,
had a rating with a mean score of 3.32. Table 4.12 indicates that the coefficient of
variation for service failure attribution was 60.3% implying that there was wide
variation in perception of respondents with regard to the indicators of service failure
attribution.
The highest coefficient of variation was for locus of attribution with reference to agent
(73%) while the lowest was for controllability attribution (48.4%) indicating that there
was greater consensus among respondents with reference to controllability attribution
than locus of attribution. When customers perceive that the service provider is able to
prevent a problem but has done nothing to avert its recurrence, they feel a sense of
anger and disappointment (Swanson Hsu, 2011).
100
4.8
Assessment of Recovery Disconfirmation
The study sought to establish the extent of recovery disconfirmation among subscribers
of MMTS in Kenya. Recovery disconfirmation was measured by evaluating the
perception of subscribers of the extent to which the solution provided and the manner
in which the service recovery was handled met with their expectations. The results are
presented in Table 4.13.
Table 4.13: Statistics for Indicators of Recovery Disconfirmation
Items
N Mean Std. Deviation CV %
Recovery disconfirmation
The problem was handled the way you expected
803
3.86
1.269
32.9
The solution met your expectations
803
3.91
1.308 33.5
The company met your expectations
803
3.90
1.303 33.4
Overall
803
3.89
1.293 33.2
Source: Primary Data
The overall mean score of 3.89 indicates that the handling of the service recovery met
customer expectations to a large extent. There was not much difference between the
ratings for the various items in this component with mean scores ranging from 3.86 to
3.90. The standard deviation for all items was greater than 1 indicating a lack of
agreement about the issues among respondents. Table 4.13 indicates that the
coefficient of variation for recovery disconfirmation was 33.2% implying that the
respondents' perception of recovery disconfirmation indicators varied by 33.2%. The
coefficient of variation for various measures of recovery disconfirmation ranged from
32.9% to 33.5%.
101
4.9
Assessment of Recovery Satisfaction
The study sought to establish the extent to which MMTS subscribers were satisfied
with the outcome of service recovery. The measures of recovery satisfaction included
satisfaction with service provided, corrective action taken and overall service. The
study also evaluated overall fulfillment, repurchase intention and willingness to
recommend as part of the recovery satisfaction judgment. The results are presented in
Table 4.14.
Table 4.14: Statistics for the Indicators of Recovery Satisfaction
Item
N Mean Std. Deviation CV %
Recovery Satisfaction
How satisfied were you with the service provided 803
3.83
1.233
32.2
How satisfied were you with the corrective action
taken
803
3.86
1.266
32.8
How satisfied were you with the overall service
803
3.90
1.209
31.0
How satisfied were you overall
803
3.90
1.217
31.2
To what extent did you feel confident to continue
using the services of the company?
803
4.10
1.062
25.9
How likely are you to recommend your mobile
money transfer service to a friend
803
3.95
1.361
34.5
Overall
803
3.92
1.224
31.2
Source: Primary Data
The overall mean score of 3.92 indicates that subscribers were satisfied with service
recovery implementation by the service providers to a large extent. Repurchase
intention and willingness to recommend were rated highest with mean scores of 4.10
and 3.95 respectively. This was followed by satisfaction with the overall service with
a mean score of 3.90. Respondents also expressed satisfaction with corrective action
and the service provided.
102
Table 4.14 indicates that the coefficient of variation for recovery satisfaction was
31.2% suggesting that the respondents' perception of recovery satisfaction varied of
31.2%. The highest coefficient of variation was for likelihood to recommend (34.5%)
while the lowest was for re-purchase intention (25.9%) implying that there was greater
consensus among respondents with regard to re-purchase intention than the likelihood
to recommend service provider to others.
4.10 Test of Normality, Linearity, Collinearity and Homoscedasticity
Since parametric techniques such as correlation and regression analyses were to be used
in the data analysis, it was necessary to test the sample data for linearity, normality of
distribution, homoscedasticity and independence of residuals. These tests were
performed to ensure no violation of the critical assumptions for parametric tests existed.
4.10.1 Tests for the Pertinent Assumptions
Before further analysis it was essential to test for linearity, normality, outliers, linearity
and homoscedasticity, independence of residuals. This was essential to ensure that
predictions and confidence intervals produced by the regression model were effective.
These assumptions were checked by inspecting the normal probability plot (P-P) of the
regression standardized residual and the scatterplot. As seen in the normal P-P plot
figure (Appendix 6), the points lie in a reasonably straight diagonal line from bottom
left to the right an indication that there are no major deviations from normality. In the
scatterplot of standardized residuals (Appendix 7) residuals are distributed in a fairly
rectangular manner with most scores concentrated in the center, an indication that
assumptions are met.
103
4.10.2 Collinearity Statistics
Collinearity statistics showed variance inflation factor (VIF) of less than 5 and
Tolerance greater than 0.2 in all cases, an indication that the variables were not highly
correlated, hence no existence of multicollinearity. This is an indication of the
suitability of the variables for multiple regression analysis.
4.11 Correlation Analysis
The study sought to establish whether there were significant associations between
recovery satisfaction, service failure attribution, perceived justice and recovery
disconfirmation. In this study, Pearson product moment correlation was used to explore
relationships between the variables, specifically to assess both the direction and
strength of the relationship between the variables. Furthermore, the correlation matrix
helped to determine whether multicollinearity existed between the independent
variables before carrying further analysis using multiple regressions. Multicollinearity
exists when independent variables are highly correlated (r=0.9 and above) and leads to
a poor regression model.
Results of correlation analysis between recovery satisfaction, service failure
attribution, perceived justice and recovery disconfirmation are shown in table 4.15.
Table 4.15: Correlation analysis for the study variables
Variable
1
2
3
4
1. Recovery Satisfaction
1
-.043
.633
**
.702
**
2. Service Failure Attribution
1
.030
-.072
*
3. Perceived Justice
1
.737
**
4. Recovery Disconfirmation
1
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
104
There was a strong and positive correlation between recovery satisfaction and
perceived justice which was statistically significant, (r =.633, p0.01). In addition,
there was a strong positive correlation between recovery satisfaction and recovery
disconfirmation which was statistically significant (r=.702, p0.01). The relationship
between recovery satisfaction and service failure attribution was weak, negative and
statistically non-significant, (r = -.043). Since the correlation between service failure
attribution, perceived justice and recovery disconfirmation was not very high (r0.9),
the variables were considered suitable for further analysis using multiple regression.
4.12 Regression Analysis and Hypotheses Testing
The study was based on the premise that there is a relationship between perceived
justice and recovery satisfaction among subscribers of mobile money transfer services
in Kenya and that this relationship is mediated by recovery disconfirmation. The study
also predicted that service failure attribution will have a moderating effect on the
relationship between perceived justice and recovery satisfaction. To test the
hypothesized relationships, regression analysis was conducted and relevant statistics
tested at 95% confidence level.
4.12.1 Perceived Justice and Recovery Satisfaction
The first objective of the study was to determine the influence of perceived justice on
recovery satisfaction among subscribers of mobile money transfer services (MMTS) in
Kenya. Perceived justice was operationalized into procedural justice, interactional
justice and distributive justice. Recovery satisfaction measures included satisfaction
with the service provided, corrective action taken and the overall service. The
respondents had been asked to indicate the extent to which they perceived fairness in
the manner service recovery was implemented. To assess perceived justice and
105
recovery satisfaction among MMTS subscribers in Kenya, the following hypothesis
was tested:
Hypothesis One (H1): There is a statistically significant relationship between
Perceived Justice and recovery satisfaction.
Perceived justice was regressed on recovery satisfaction in order to determine their
relationship. The pertinent results are summarized in Table 4.16.
106
Table 4.16: Regression Results for Perceived Justice and Recovery Satisfaction
(a): Model Summary
Model
R
R Square
Adjusted R Square
Std. Error of the Estimate
1
.633
a
.401
.400
.77466
(b): ANOVA for the Effect of Perceived Justice on Recovery Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
321.924
1
321.924
536.457
.000
b
Residual
480.674
801
.600
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Std. Error
Beta
1
(Constant)
.749
.140
5.358
.000
Perceived Justice
.846
.037
.633
23.162
.000
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Perceived Justice
The results in Table 4.16 show that perceived justice had a statistically significant
influence on recovery satisfaction. As shown in Table 4.16 (a) the R
2
is 0.401 implying
that perceived justice accounts for 40.1% of the explained variation of recovery
satisfaction among subscribers of mobile money transfer services in Kenya. This
implies that other factors not included in the model explain 59.9% of the variation in
the recovery satisfaction scores.
The standardized regression coefficient () value of the scores for perceived justice was
.633 with a t-test of 5.358 and a significance level of (p-value=.000). The results
indicate that as predicted, perceived justice influences recovery satisfaction. The study
used standardized regression coefficient because it is free from original units of the
107
predictor and outcome variables allowing for ease of comparison. Standardized
regression coefficients have been used in previous studies including Ndungu (2013);
Njeru, (2013) and Waithaka (2014).
The p-value for the regression coefficient of perceived justice is statistically significant
(p0.05) meaning that perceived justice is a significant predictor of recovery
satisfaction. This conclusion is consistent with previous studies on the perceived
justice and recovery satisfaction relationship. Tan (2014) in a study on the effects of
the justice theory in service recovery in the Philippines concluded that there was a
positive relationship between perceived justice and recovery satisfaction. Ellyawati et
al. (2012) concluded that perceived justice affects recovery satisfaction in a study on
emotions and recovery satisfaction among retailers in Indonesia.
The hypothesis that there is a statistically significant relationship between perceived
justice and recovery satisfaction among MMTS subscribers in Kenya is supported by
the current study. Recalling the prediction equation Y = + 1X1 + e, where; =
Constant; Y= Recovery satisfaction, 1X1 = perceived justice and e = Error term, the
values for the regression weights are as follows:
Y= 0.749+0.633 X1 + e, which can be rewritten as follows:
RS = 0.749+0.633 PJ + e
Where: RS = Recovery Satisfaction
PJ = Perceived Justice
4.12.2 Dimesions of Perceived Justice and Recovery Satisfaction
To further examine the effect of perceived justice on recovery satisfaction, multiple
regression analysis using the subscales of perceived justice namely, procedural justice,
108
interactional justice and distributive justice was performed. The objective was to
determine the relationship between recovery satisfaction (dependent variable) and
procedural justice, interactional justice and distributive justice (independent variables).
Collinearity statistics indicated that VIF5 and Tolerance0.2, an indication that the
variables were not highly correlated; hence no noticeable multicollinearity. This is an
indication of the suitability of variables for multiple regression analysis. The pertinent
results are presented in Table 4. 17.
Table 4.17: Regression Results for Dimensions of Perceived Justice and Recovery
Satisfaction
(a): Model Summary
Model
R
R Square
Adjusted R Square
Std. Error
1
.642
a
.412
.410
.76870
(b): ANOVA for the Effect of Distributive Justice, Procedural Justice and
Interactional Justice on Recovery Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sign.
1
Regression
330.465
3
110.155
186.417
.000
b
Residual
472.133
799
.591
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Collinearity
Statistics
Std.
Error
Beta
Tolerance
VIF
1
(Constant)
.907
.145
6.248
.000
Procedural Justice
.200
.041
.157
4.841
.000
.701
1.426
Interactional Justice
.190
.042
.169
4.574
.000
.538
1.857
Distributive Justice
.418
.038
.416 10.998
.000
.514
1.945
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Distributive Justice, Procedural Justice, Interactional Justice
109
The regression model produced R² = .412, p0.05 which suggests that the joint effect
of distributive justice, procedural justice and interactional justice accounts for 41.2%
of the variance of recovery satisfaction among subscribers of mobile money transfer
services in Kenya. The results indicated that there was a positive and statistically
significant relationship between procedural justice and recovery satisfaction ( = .157,
t = 4.841, p .05). Similarly, there was a positive and statistically significant
relationship between recovery satisfaction and interactional justice ( = .169, t = 4.574,
p.05) as well as recovery satisfaction and distributive justice ( = .416, t = 10.998,
p.05).
The fact that the regression coefficients are positive means that increase in procedural
justice, interactional justice or distributive justice corresponds to increase in recovery
satisfaction. This corroborates the findings of Rio-Lanza, et al. (2009) and Tan, (2014)
who found that the three justice components have a positive influence on recovery
satisfaction.
Further, the results revealed that as shown in Table 4.17 distributive justice has the
strongest relationship with recovery satisfaction among the three dimensions. This
finding is consistent with previous studies where distributive justice was found to have
the strongest relationship with recovery satisfaction (Tan, 2014; Nibkin, et al., 2010).
In addition, the findings corroborate the findings of Chebat and Slusarczyk (2005) who
concluded that distributive justice is the most effective means of improving customer
satisfaction in situations involving recovery. It also agrees with the study by
Chepkwony et al. (2012) which found positive effects of distributive justice on
satisfaction. The finding disagrees with a study by Rio-Lanza et al. (2009) who
concluded that procedural justice has the strongest influence on recovery satisfaction.
110
Recalling the regression equation:
Y= + 11X1 + 12X2 + 13X3
+ e, where Y represents Dependent Variable (RS), = constant (y intercept); 11, 12,
13 represents regression coefficients and X1, X2, X3 = independent variable
dimensions (PJ) and e is the error term the regression equation can be rewritten from
the results of multiple regression.
Thus based on these results, the regression equation that predicts recovery satisfaction
based on the linear combination of distributive justice, procedural justice and
interactional Justice is as follows.
Y (Recovery Satisfaction) = 0.907 + 0.157X1(procedural Justice) + 0.169X2
(interactional justice) + 0.416X3(Distributive justice) + e
4.12.3 The mediating effect of recovery disconfirmation
The study set out to assess the intervening role of recovery disconfirmation on the
relationship between perceived justice and recovery satisfaction. The following
hypothesis was formulated:
Hypothesis two (H2): Recovery disconfirmation has a significant mediating
influence on the relationship between perceived justice and recovery
satisfaction
The mediating effect was computed using the four-step process proposed by Baron and
Kenny (1986).
111
Step 1
In the first of the mediation model, regression analysis was performed to assess the
association between recovery satisfaction (dependent variable) and perceived justice
(independent variable) ignoring the mediator (recovery disconfirmation). The
regression model produced R² = .401, F = 536.457, p0.05 which suggests that
perceived justice explains 40.1% of the variance of recovery satisfaction among
subscribers of mobile money transfer services in Kenya. Based on the results we
conclude that the relationship between recovery satisfaction and perceived justice is
statistically significant.
Step 2
In the second step of the mediation model, regression analysis was performed to assess
the association between recovery disconfirmation (intervening variable) and perceived
justice (independent variable) ignoring the dependent variable (recovery satisfaction).
The regression model produced R² = .543, F = 952.805, p0.05 which suggests that
perceived justice explains 54.3% of the variance of recovery disconfirmation among
subscribers of mobile money transfer services in Kenya. This implies that factors not
included in the model explain 45.7% of the variation in perceived justice scores.
Based on the results, we conclude that perceived justice had a positive and statistically
significant relationship with recovery satisfaction ( = .737, t = 30.868, p.05) as
shown in table 4.19 (c). The fact that the regression coefficient is positive means that
an increase in perceived justice corresponds to an increase in recovery disconfirmation.
When independent variable is zero, recovery disconfirmation has a negative value (=
-.761).
112
Step 3
In step 3 of the mediation model, regression analysis was performed to assess the
association between recovery disconfirmation (intervening variable) and recovery
satisfaction (dependent variable) ignoring the independent variable (perceived justice).
The regression model produced R² = .493, F = 777.970, p0.05 which suggests that
perceived justice explains 49.3% of the variance of recovery disconfirmation among
subscribers of mobile money transfer services in Kenya.
The correlation coefficient (R) 0.702 is an indication that there is a positive and strong
(0.5) relationship between recovery disconfirmation and recovery satisfaction. Based
on the results, we conclude recovery disconfirmation had a positive and statistically
significant relationship with recovery satisfaction ( = .702, t = 27.892, p.05). The
fact that the regression coefficient is positive means that an increase in recovery
disconfirmation corresponds to increase in recovery satisfaction.
Step 4
In the fourth step of the mediation model, regression analysis was performed to assess
the association between recovery satisfaction (dependent variable) and recovery
disconfirmation (intervening variable) and the independent variable (perceived justice).
The regression model produced R² = .522, F = 437.073, p0.05 which suggests that
perceived justice and recovery disconfirmation explains 52.2% of the variance of
recovery satisfaction among subscribers of mobile money transfer services in Kenya.
The correlation coefficient (R) is 0.723, an indication that there is a positive and strong
(0.5) relationship between recovery disconfirmation and recovery satisfaction. The
detailed results are summarized in appendix 8.
113
Based on these results, we conclude that both recovery disconfirmation and perceived
justice have a positive and statistically significant relationship with recovery
satisfaction. Table 4.18 presents the pertinent results.
Table 4.18: Regression Results for the Mediating Effect of Recovery
Disconfirmation
(a): Model Summary
Model
R
R Square
Adjusted R Square
Std. Error of the Estimate
1
.723
a
.522
.521
.69239
(b): The ANOVA for the effect of Recovery disconfirmation on Recovery Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
419.072
2
209.536
437.073
.000
b
Residual
383.526
800
.479
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Collinearity
Statistics
Std. Error
Beta
Tolerance
VIF
1
(Constant)
1.060
.127
8.359
.000
Perceived Justice
.339
.048
.254 7.020
.000
.457 2.190
Recovery
Disconfirmation
.409
.029
.515 14.235
.000
.457 2.190
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Recovery Disconfirmation, Perceived Justice
There was a positive and statistically significant relationship between perceived justice
and recovery satisfaction ( = .254, t = 7.020, p.05) as shown in table 4.18.Similarly,
there was a positive and statistically significant relationship between recovery
satisfaction and recovery disconfirmation ( = .515, t = 14.235, p.05). The fact that
the regression coefficients are positive means that increase in perceived justice or
recovery disconfirmation corresponds to increase in recovery satisfaction.
114
The regression equation is as follows:
Step 4: RS= 1.060 +0.254PJ+0 .515RD+ e, (Model is statistically significant, p0.05)
Table 4.19 presents the summary of the step by step mediation process.
Table 4.19: Summary of Mediating Effect of Recovery Disconfirmation
Analysis
R
R
2
R
2
Change
Significance
P-Value
Step 1: Recovery satisfaction on
perceived justice
.633
.401
.000
Step 2: Recovery disconfirmation
on perceived justice
.737
.543
.737
.000
Step 3:
Step 3a: Recovery satisfaction on
recovery disconfirmation
Step 3b: Recovery satisfaction on
perceived justice
.702
.723
.493
.522
.029
.702
5.15
.000
.000
Source: Primary Data
Hypothesis two (H2) explored the relationship between recovery satisfaction,
perceived justice and recovery disconfirmation by suggesting that recovery
disconfirmation has a significant mediating influence on the relationship between
perceived justice and recovery satisfaction. Perceived justice and recovery
disconfirmation are significant predictors of recovery satisfaction (step 1 and 3) and
their joint effect on recovery satisfaction is also significant.
Since the magnitude of the regression coefficient of perceived justice is reduced when
recovery disconfirmation is introduced into the regression equation (step 4), this
implies that recovery disconfirmation has a mediating role in the relationship between
perceived justice and recovery satisfaction. Hypothesis 2 was therefore accepted.
115
The path diagram for the mediation effect is depicted in Figure 4.1
Figure 4.1: Path Diagram for Mediation Effect of Recovery Disconfirmation
Source: Primary data
PJ: Perceived Justice; RD: Recovery Disconfirmation; RS: Recovery Satisfaction
The results in Figure 4.1 support the hypothesis that recovery disconfirmation has a
mediating effect on the relationship between perceived justice and recovery
satisfaction. The pertinent results show that R² increased from .493 to .522 when RD
was included (.493+.029=.522). The results imply that RD explains the additional 2.9%
of the variation in recovery satisfaction.
Previous studies have indicated that recovery disconfirmation has a significant
influence on recovery satisfaction. A study by Smith and Bolton (2002) concluded that
disconfirmation of service recovery has an intermediary effect in the relationship
between perceived justice and recovery satisfaction. Andreassen (2000) concluded that
both perceived justice and recovery disconfirmation are complementary drivers of
recovery satisfaction.
RD
Path b
R
2
=.401 =.633
P=.000
Path a
R
2
=.5.43 =.737
P=.000
PJ
RS
Path B: R
2
=.4.93 R
2
=.029 =.515 P=.000
116
4.12.4 The Moderating Influence of Service Failure Attribution
The study set out to assess the moderating effect of service failure attribution on the
relationship between perceived justice and recovery satisfaction among subscribers of
mobile money transfer services in Kenya. The following hypothesis was formulated:
Hypothesis three (H3): Service failure attribution has a significant
moderating influence on the relationship between perceived justice and
recovery satisfaction.
Multiple hierarchical regression analysis was used to explore the relationship. Prior to
conducting hierarchical multiple regression, the relevant assumptions of this statistical
analysis were tested. On testing for multicollinearity in the model the results indicated
that VIF was less than 5 and Tolerance was greater than 0.2 an indication that the
variables were not highly correlated, hence no noticeable multicollinearity.
The moderating effect was computed using the method proposed by Baron and Kenny
(1986). In order to test moderating effect, there was need to predict the outcome of
dependent variable (recovery satisfaction) from the predictor variables (perceived
justice and Service failure attribution). Secondly, the independent variable and the
moderator variable were centered and an interaction term was created by multiplying
the independent variable and the moderator. The interaction term was then entered in
the regression equation to assess whether the moderator variable alters the strength of
the causal relationship. To create an interaction term, perceived justice and service
failure attribution measures were first centered and a single item indicator representing
the product of the two measures calculated.
117
The results of hierarchical multiple regression analysis predicting recovery satisfaction
from perceived justice and service failure attribution are reported in Tables 4.20.
118
Table 4.20: Regression Results for Assessing the Moderating Influence of Service
Failure Attribution on the Relationship between Perceived Justice and Recovery
Satisfaction
(a): Model Summary
Model
R
R
Square
Adjusted
R Square
Std. Error
of the
Estimate
Change Statistics
R
Square
Change
F
Change
df1 df2
Sig. F
Change
1
.636
a
.405
.403
.77265
.405 272.202
2 800
.000
2
.649
b
.421
.419
.76250
.016
22.444
1 799
.000
(b): The ANOVA for the effect of Perceived Justice and Service Failure Attribution on
Recovery Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
325.005
2
162.502
272.202
.000
b
Residual
477.593
800
.597
Total
802.598
802
2
Regression
338.054
3
112.685
193.814
.000
c
Residual
464.544
799
.581
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Collinearity
Statistics
Std.
Error
Beta
Tolerance
VIF
1
(Constant)
.921
.159
5.804
.000
Perceived Justice
.849
.036
.635 23.279
.000
.999
1.001
Service Failure
Attribution
-.070
.031
-.062 -2.272
.023
.999
1.001
2
(Constant)
1.127
.163
6.933
.000
Perceived Justice
.818
.037
.612 22.360
.000
.967
1.034
Service Failure
Attribution
-.105
.031
-.094 -3.378
.001
.941
1.063
Interaction Term
(cPJ*cSFA)
.193
.041
.134 4.738
.000
.911
1.098
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Service Failure Attribution, Perceived Justice
119
The multiple regression model (model 1) produced R² = .405, F = 272.202, p .05.
The model (model 1) reveals a statistically significant relationship between recovery
satisfaction, perceived justice (independent variable) and service failure attribution
(moderator). In step 2 (model 2), the interaction between perceived justice (independent
variable) and service failure attribution (moderator) was entered into the regression
equation.
The change in variance accounted for (R
2
) was equal to .016, which was statistically
significant increase in variance accounted for over the step one model. Similarly, F
change was 22.444 and it was statistically significant (p0.05). Model 2 reveals a
statistically significant relationship between recovery satisfaction, perceived justice
and service failure attribution (moderator) and the interaction term. R² = .421, F =
193.814, p 0.05 as shown in table 4.22 (b). Model 2 accounted for 42.1% of the
variance in recovery satisfaction (R2 =.421).
As shown in Table 4.20 (c), before the inclusion of the interaction term (model 1), the
regression coefficient () value of perceived justice was .635 with a t-test of 23.279
(p0.05). The regression coefficient () value of service failure attributions was -.062
with a t-test of -2.272 (p0.05). After the inclusion of the interaction term, the beta
coefficient of perceived justice was .612 and it was statistically significant (p0.05).
The beta coefficient of service failure attribution was -.094 and it was statistically
significant (p0.05). The interaction term was also statistically significant (= 0.134,
t= 4.738, p=0.000) as shown in Table 4.20 (c).
Hypothesis three (H3) explored the relationship between recovery satisfaction,
perceived justice and service failure attribution among subscribers of mobile money
120
transfer services in Kenya by suggesting that service failure attribution has a significant
moderating influence on the relationship between perceived justice and recovery
satisfaction.
The value of R2 change was 0.016 and it was statistically significant. Similarly, the
interaction term was statistically significant (p0.05). This indicates that Service failure
attribution has a significant moderating influence on the relationship between perceived
justice and recovery satisfaction and therefore hypothesis three was accepted. The
regression equation is as follows:
Y= 1.127 + .818PJ -.105 SFA -.011 Interaction Term (cPJ*cSFA) + e,
Where: Y= recovery satisfaction, PJ=perceived justice, SFA= Service failure
attribution and = error term.
The results of the current study are consistent with previous studies. Nibkin et al. (2011)
established a moderating effect of failure attributions in a study on the impact of firm
reputation on customer responses to service failure in Malaysia.
4.12.5 The Moderation Effect of Dimensions of Service Failure Attribution
Further to assessing the moderating influence of service failure attribution, the study
sought to evaluate the role played by the various attribution components on the
relationship between perceived justice and recovery satisfaction among subscribers of
mobile money transfer services in Kenya. Multiple hierarchical regression analysis was
used to explore the relationship. Pertinent results relating to locus of attributions,
stability and controllability are presented in appendix 9, 10 and 11 respectively.
121
The study results revealed that locus of attribution had a significant moderating
influence on the relationship between perceived justice and recovery satisfaction
among subscribers of mobile money transfer services in Kenya. The relationship
between locus of attribution and recovery satisfaction is however negative and
statistically non-significant. Further, the findings revealed that that stability attribution
has a significant moderating influence on the relationship between perceived justice
and recovery satisfaction. The relationship between controllability attribution and
recovery satisfaction is however negative and statistically non-significant.
In addition the findings revealed that stability attribution has a significant moderating
influence on the relationship between perceived justice and recovery satisfaction.
However, the relationship between controllability attribution and recovery satisfaction
is however negative and statistically non-significant. The results of the current study
are consistent with Nibkin et al. (2011) who found that stability and controllability
attributions have a moderating influence on the relationship between perceived justice
and recovery satisfaction.
4.12.6 The Joint Effect of Perceived Justice, Disconfirmation and Service Failure
Attribution on Recovery Satisfaction
The study further sought to determine the joint effect of perceived justice,
disconfirmation and service failure attribution on recovery satisfaction. The following
hypothesis was formulated and tested:
Hypothesis four (H4): The joint effect of perceived justice, service failure
attribution and disconfirmation on recovery satisfaction is statistically
significant
122
Multiple regression analysis was used to explore the relationship. The pertinent results
are presented in Table 4.21
Table 4.21: Regression Results for Determining the Joint Effect of Perceived
Justice, Disconfirmation and Service Failure Attribution on Recovery Satisfaction
(a): Model Summary
Model
R
R Square
Adjusted R Square
Std. Error of the Estimate
1
.723
a
.522
.521
.69269
(b):
The ANOVA for the effect of Perceived Justice, Recovery Disconfirmation and Service
Failure Attribution on Recovery Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
419.224
3
139.741 291.239
.000
b
Residual
383.374
799
.480
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Collinearity
Statistics
Std.
Error
Beta
Tolerance
VIF
1
(Constant)
1.097
.143
7.680
.000
Perceived Justice
.343
.049
.256 7.033
.000
.450
2.223
Service Failure
Attribution
-.016
.028
-.014 -.563
.573
.980
1.021
Recovery
Disconfirmation
.407
.029
.512
14.01
3
.000
.448
2.233
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Recovery Disconfirmation, Service Failure Attribution,
Perceived Justice
The regression model produced R² = .522, F= 291.239, p0.05. R-square (R²) is .522
which suggests that jointly, Recovery Disconfirmation, Service Failure Attribution and
Perceived Justice account for 52.2% of the variance of recovery satisfaction among
subscribers of mobile money transfer services in Kenya. The correlation coefficient (R)
=0.723 an indication that there is a positive and strong (0.5) relationship between
123
recovery satisfaction, recovery disconfirmation, service failure attribution and
perceived justice.
Recalling the regression equation: Y= + 1X1+ 2X2 + 3X3 +e, where Y represents
Dependent Variable (RS), = constant (y intercept); 1, 2, 3 represents regression
coefficients and X1, X2, X3 = independent variable variables and e the error term, the
regression equation can be rewritten from the results of multiple regression. Thus based
on these results, the regression equation that predicts recovery satisfaction based on the
linear combination of recovery disconfirmation, service failure attribution, and
perceived justice is as follows:
Y (Recovery Satisfaction) = 1.097 + 0.256X1(perceived justice) - 0.014X2 (Service
failure attribution) + 0.512X3(recovery disconfirmation) + e
There was a positive and statistically significant relationship between perceived justice
and recovery satisfaction ( = .256, t = 7.033, p .05). Similarly, there was a positive
and statistically significant relationship between recovery satisfaction and recovery
disconfirmation ( = .512, t = 14.013, p .05). The relationship between recovery
satisfaction and service failure attribution was negative and not statistically significant
( = -.014, t = -.563, p.05). The fact that the regression coefficients of PJ and RD are
positive means that increase in perceived justice or recovery disconfirmation
corresponds to increase in recovery satisfaction. Based on these results, RD (0.512) has
the strongest relationship with recovery satisfaction compared to PJ (0.256) and SFA
(-0.014).
Hypothesis four (H4) explored the relationship between recovery satisfaction, recovery
disconfirmation, service failure attribution and perceived justice among subscribers of
124
mobile money transfer services in Kenya by suggesting that the joint effect of perceived
justice, service failure attribution and disconfirmation on recovery satisfaction is
statistically significant. The overall model was statistically significant (p0.05), an
indication that the joint effect of perceived justice, service failure attribution and
disconfirmation on recovery satisfaction is statistically significant. The hypothesis was
therefore accepted. Service failure attribution was however not a significant predictor
of recovery satisfaction.
4.13 Summary of Research Objectives, Hypotheses and Conclusions
The results show that all four hypotheses of the study were accepted. Table 4.22 shows
a summary of research findings and conclusions for each of the hypothesis in the study.
125
Table 4.22: Summary of Research Objectives, Hypotheses and Conclusions
Objective
Hypothesis
R
R
2
Significance
P-Value
Conclusion
Objective 1: Determine
the influence of perceived
justice
on
recovery
satisfaction
among
subscribers of mobile
money transfer services in
Kenya.
H1: There is a
significant
relationship
between
perceived justice
and
recovery
satisfaction.
.633 .401 .000
Hypothesis
one was
supported
Objective 2:
Assess the mediating
effect
of
recovery
disconfirmation on the
relationship
between
perceived justice and
recovery satisfaction.
H2:
Recovery
disconfirmation
has a significant
mediating
influence on the
relationship
between
perceived justice
and
recovery
satisfaction.
.702 .493 .000
Hypothesis
two was
supported
Objective 3:
Establish the moderating
influence of service failure
attribution
on
the
relationship
between
perceived justice and
recovery satisfaction.
H3:
Service
failure attribution
has a significant
moderating
influence on the
relationship
between
perceived justice
and
recovery
satisfaction.
.649 .421 .000
Hypothesis
three was
Supported
Objective 4:
Determine the joint effect
of
perceived
justice,
disconfirmation
and
service failure attribution
on recovery satisfaction.
H4: The joint
effect
of
perceived justice,
service
failure
attribution
and
disconfirmation
on
recovery
satisfaction
is
statistically
significant.
.723 .522 .000
Hypothesis
four was
Supported
Source: Primary Data
126
There was a significant and positive relation between perceived justice and recovery
satisfaction. The study results revealed that service failure attribution moderates the
relationship between perceived justice and recovery satisfaction. Further, the study
found that recovery disconfirmation partially mediates the relations between perceived
justice and recovery satisfaction. Finally, the results indicated that the joint effect of
perceived justice, service failure attribution and disconfirmation on recovery
satisfaction is significant. As predicted hypotheses one, two, three and four were
accepted.
The conceptual framework of the study hypothesized that first, there is a significant
relationship between perceived justice and recovery satisfaction; second, recovery
disconfirmation has a significant mediating influence on the relationship between
perceived justice and recovery satisfaction; third, service failure attribution has a
significant moderating influence on the relationship between perceived justice and
recovery satisfaction and finally, that the joint effect of perceived justice, service failure
attribution and disconfirmation on recovery satisfaction is statistically significant.
Based on the results of the study, Figure 4.2 presents the modified conceptual
framework, which shows that all four hypotheses tested in this study were statistically
significant and were therefore supported.
127
Figure 4.2: Modified Conceptual Model
Moderating Variable
Independent Variable
Dependent Variable
Mediating variable
The modified conceptual model in Figure 4.2 shows that first, there is a statistical
significance relationship between perceived justice and recovery satisfaction among
subscribers of MMTS in Kenya as evidenced by H1 (H1: R
2
=.401; =.633; P-
Value=.000). Secondly, the relationship between perceived justice and recovery
satisfaction is significantly mediated by recovery disconfirmation H2 (R
2
=.493; R
2
=.029; =.515; P-Value=.000)
Service Failure
Attribution
-Stability
-Controllability
-Locus
Perceived Justice
-Procedural Justice
-Interactional Justice
-Distributive Justice
Recovery
Satisfaction
-Service encounter
satisfaction
-Overall satisfaction
-Repurchase intention
Recovery
Disconfirmation
-Expectations
-Performance
H1: RS= 0.749+0.633X1(PJ)
=0 R².401; p0.05
H3: RS= 1.127 + 0.612PJ -
0.094 SFA -0.134;
R
2
=0.016 p0.05 R
2
=.421
H2: RS= 1.060 +0.254PJ
+0.515RD; R
2
=0.029
R²= 0.522, p0.05
H4: RS= 1.097 +
0.256PJ- 0.014SFA
+ 0.512RD;
R².522; p0.05
128
Thirdly, the relationship between perceived justice and recovery satisfaction is
moderated by service failure attribution H3 (H3:R
2
=.421; R
2
=.016; =.134; P-
Value=.000). Finally, the joint effect of perceived justice, recovery disconfirmation
and service failure attributions on recovery satisfaction is statistically significant H4
(H4: R
2
=.723; P-Value=.000). The section that follows discusses implications of the
study results.
4.14 Discussion of the Results
This section presents the discussion of results based on objectives and conceptual
hypotheses of the study. The researcher developed a conceptual framework based on
the literature on service failure and recovery satisfaction highlighting the hypothesized
relationships between study variables. These relationships were empirically tested.
4.14.1 The Relationship between Perceived Justice and Recovery Satisfaction
The first objective of the study was to establish the influence of perceived justice on
recovery satisfaction among mobile money transfer services in Kenya. The study
results show that perceived justice explained 40% of the variance in recovery
satisfaction and the regression model was statistically significant indicating that, as was
hypothesized, perceived justice has a significant influence on recovery satisfaction
among subscribers of MMTS. This is consistent with previous studies by Nibkin et al.
(2010) and Rio-Lanza et al (2009). These studies reported a positive relationship
between perceived justice and recovery satisfaction.
Upon further investigation of the dimensions of perceived justice, the results showed
that procedural justice, interactional justice and distributive justice are significant
predictors of recovery satisfaction among subscribers of MMTS in Kenya. An increase
129
in these dimensions leads to an increase in recovery satisfaction. Further, the results
show that distributive justice had the strongest influence on recovery satisfaction,
followed by interactional justice and finally procedural justice. The findings
corroborate the studies by Tan (2014) and Nibkin et al. (2010) which reported similar
results. The results contradict the study by Rio-Lanza et al (2009) which reported that
procedural justice had the strongest influence on recovery satisfaction compared to the
other two dimensions.
4.14.2 The Mediating effect of Recovery Disconfirmation
The second objective of the study was to assess the mediating effect of recovery
disconfirmation on the relationship between perceived justice and recovery satisfaction
among subscribers of mobile money transfer services in Kenya. The study
hypothesized that recovery disconfirmation had a significant mediating effect on
recovery satisfaction. This was assessed in a four step process as recommended by
Baron and Kenny (1986). The results indicate that both recovery disconfirmation and
perceived justice had a positive and significant relationship with recovery satisfaction.
The study revealed that the recovery disconfirmation has a mediating effect in the
relationship between perceived justice and recovery satisfaction. Previous studies have
indicated that recovery disconfirmation has significant influence on recovery
satisfaction. This study agrees with Smith and Bolton (2002) who concluded that both
perceived justice and disconfirmation of service recovery are intermediary factors in
the relationship between perceived justice and recovery satisfaction. Andreassen
(2000) noted that perceived justice and recovery disconfirmation are complementary
drivers of recovery satisfaction.
130
4.14.3 The Moderating effect of Service Failure Attribution
The third objective of this study sought to establish the influence of service failure
attribution in the relationship between perceived justice and recovery satisfaction
among subscribers of mobile money transfer services in Kenya. The results revealed a
statistically significant relationship between perceived justice, service failure
attribution and recovery satisfaction. After the introduction of the interaction term
between perceived justice and recovery satisfaction into the equation, there was a
statistically significant increase in variance supporting the prediction that service
failure attribution has a moderating influence on the relationship between perceived
justice and recovery satisfaction.
Further, the study indicated that each of the dimensions of service failure attribution
namely, locus, controllability and stability attributions have a moderating effect on the
relationship between perceived justice and recovery satisfaction. This agrees with the
findings by Nibkin et al. (2011) who found that stability and controllability attribution
factors have a moderating influence in the relationship between perceived justice and
recovery satisfaction. Swanson Hsu (2011) found that locus of attributions of service
failure impacts recovery satisfaction.
4.14.4 Perceived Justice, Service Failure Attribution, Recovery Disconfirmation
and Recovery Satisfaction
The final objective of the study was to determine the joint effect of perceived justice,
service failure attribution, recovery disconfirmation and recovery satisfaction among
subscribers of mobile money transfer services in Kenya. It was hypothesized that the
joint effect of perceived justice, service failure attribution, and recovery
disconfirmation on recovery satisfaction is statistically significant.
131
Multiple regression analysis was used to explore the relationship and the result
indicated that the joint effect of perceived justice, service failure attribution and
recovery disconfirmation on recovery satisfaction was statistically significant. Service
failure attribution was however not a significant predictor of recovery satisfaction. This
implies that recovery satisfaction of MMTS subscribers is jointly influenced by
perceived justice, recovery disconfirmation and service failure attribution.
4.15 Chapter Summary
The chapter has presented data analysis and results of the study objectives. Four
hypotheses were tested and based on objectives. The results revealed positive and
statistically significant results on the relationship between perceived justice and
recovery satisfaction; the moderating effect of service failure attribution; the mediating
effect of recovery disconfirmation and the joint effect of perceived justice, service
failure attribution and recovery disconfirmation on recovery satisfaction. The chapter
has also presented the discussion of results highlighting agreement or departure from
previous theoretical and empirical studies.
132
CHAPTER FIVE
SUMMARY, CONCLUSION, RECOMMENDATIONS AND IMPLICATIONS
5.1
Introduction
This study sought to establish the influence of perceived justice service failure
attribution, disconfirmation and recovery satisfaction among subscribers of mobile
money transfer services in Kenya. This chapter provides a summary of the main
findings of the study, implications, conclusions and recommendations centered on the
research objectives. The chapter further discusses the theoretical and managerial
implications as well as provides policy recommendations. In addition, it discusses the
limitations to the study and proposed areas for further research.
5.2
Summary
The introductory chapter of the study provided a detailed background of the study, the
research problem, study objectives and value of the study. In the second chapter on
literature review a theoretical and empirical grounding of the thesis is provided. In
addition, a discussion on the existing knowledge about the variables and their
relationships as well as gaps in knowledge is presented.
The third chapter on research methodology provided the philosophical orientation of
the study, the research design and population of interest. It also provided the sampling
method, data collection and analysis techniques used in the study. The fourth chapter
presented the results of the data analysis based on study objectives and hypotheses
while the fifth chapter presented the summary, conclusions and implications.
The study found that the majority of study respondents (66.4%) had used mobile money
transfer services for more than five years and were quite familiar with the providers'
133
designated line for reporting service failures. The respondents had experienced various
types of service failures emanating from their own mistakes, the service providers
system capabilities, service protocols and agents operations or behavior.
The leading problems for which customers sought recovery from the mobile network
operator included sending money to the wrong number, delay in notification upon
sending money and receiving fake mobile money transfer messages from strangers.
These problems constituted over half (55%) of all problems for which customers sought
service recovery. Other less common service problems that led the subscriber to seek
redress included non-responsive money transfer menu and sending the same transfer
twice unintentionally assuming the previous transaction had failed following a long
delay in confirming. Most of the respondents were male (56%), had secondary
education and above (79.5%) and were above twenty-four years of age (85.2%).
With regard to perceived justice with service recovery by MMTS providers in Kenya,
respondents had a positive rating meaning that they consider the approach followed by
service providers in service recovery to be appropriate and fair. With reference to
perceived justice dimensions interactional justice was rated highest followed by
distributive justice and procedural justice in that order. Respondent's opinion on
service failure attribution was moderate with controllability attribution having the
highest rating followed by locus of attribution and stability in that order.
Based on the study findings, it appears that respondents were inclined to distribute
responsibility for service failure to self and company equally, and that they considered
the service provider as only moderately capable of preventing most of the service
failures experienced. Recovery disconfirmation had the highest rating of all variables
134
meaning that subscribers were of the view that the solution provided and the manner in
which the service recovery was handled was largely consistent with their expectations.
Recovery satisfaction was rated positively indicating that subscribers were happy with
the action taken to rectify the service failure by the service provider. Respondents
further expressed strong repurchase intention and willingness to recommend the service
provider to others.
The study established significant correlations among the study variables. There was a
strong and positive correlation between perceived justice and recovery satisfaction, and
between recovery disconfirmation and recovery satisfaction and both were statistically
significant. The relationship between recovery satisfaction and service failure
attribution was weak, negative and statistically non-significant. Procedural justice,
interactional justice and distributive justice were not highly correlated. Service failure
attribution, perceived justice and recovery disconfirmation were not highly correlated.
Hypotheses were tested based on study objectives. In the first objective, regression
analysis was used to assess whether perceived justice significantly predicted recovery
satisfaction among MMTS subscribers in Kenya. The study established that perceived
justice has a positive and significant influence on recovery satisfaction among
subscribers of MMTS. Regarding the dimensions of perceived justice, the results
showed that procedural justice, interactional justice and distributive justice are
significant predictors of recovery satisfaction among subscribers of MMTS in Kenya.
Further, the results indicated that distributive justice had the strongest influence on
recovery satisfaction, followed by interactional justice and finally procedural justice.
This means that MMTS customers in Kenya view the procedures and practices adopted
135
by service providers to rectify service failure as fair and fitting. With regard to the role
of recovery disconfirmation in the relationship between perceived justice and recovery
satisfaction, the results revealed that both recovery disconfirmation and perceived
justice had a positive and significant relationship with recovery satisfaction. Further,
the results indicated that recovery disconfirmation had a significant mediating role in
the relationship between perceived justice and recovery satisfaction.
Concerning the influence of service failure attribution in the relationship between
perceived justice and recovery satisfaction, the results revealed a statistically
significant relationship between perceived justice, service failure attribution and
recovery satisfaction. Further, the results indicated that service failure attribution has a
significant moderating influence on the relationship between perceived justice and
recovery satisfaction.
With reference to the joint effect of perceived justice, service failure attribution,
recovery disconfirmation and recovery satisfaction the results indicated that the joint
effect was statistically significant. This implies that as predicted, recovery satisfaction
of MMTS subscribers is influenced by perceived justice, service failure attribution and
recovery disconfirmation.
3
Conclusions
The thrust of this study was to demonstrate the relationship between perceived justice
and recovery satisfaction as mediated by recovery disconfirmation and moderated by
service failure attribution. This study assessed the influence of perceived justice on
recovery satisfaction among subscribers of mobile money transfer services in Kenya
136
and concluded that perceived justice had a positive and statistically significant
relationship with recovery satisfaction.
Further, the results indicated that there was a positive and statistically significant
relationship between the dimensions of procedural justice, interactional justice and
distributive justice and recovery satisfaction. This implies that the processes, practices
and interactional methods used by mobile money transfer service providers to rectify
service failure are important in customers' recovery satisfaction judgment.
The results further revealed that distributive justice had the strongest relationship with
recovery satisfaction. This suggests that in evaluating service recovery, MMTS users
are particularly keen on a fair solution and compensation in order to make the effort to
seek corrective action worthwhile. By investing in effective interaction and distributive
approaches and effective service recovery procedures, MMTS providers should
experience enhanced recovery satisfaction.
The study examined the mediating effect of recovery disconfirmation on the
relationship between perceived justice and recovery satisfaction and concluded that
certainly, recovery disconfirmation mediates that relationship. This means that mobile
money subscribers are more likely to be satisfied when the service recovery outcome
meets their expectations. This implies that mobile money service providers should
invest in understanding the expectations of subscribers and developing effective
mechanisms for matching them in implementing service recovery.
The study investigated the moderating influence of service failure attribution on the
relationship between perceived justice and recovery satisfaction and concluded that
service failure attribution has a significant moderating influence in that relationship.
137
Further, the study concluded that controllability attribution had the greatest
contribution to that moderating effect. This means that MMTS subscribers are more
likely to perceive service providers' intervention through service recovery as fair and
therefore more satisfactory when the service failure experienced is not viewed as
preventable. In order to enhance recovery satisfaction, service providers should take
the measures necessary to address issues within their control so as to avoid service
failures that customers might perceive as preventable.
Finally, the study examined the joint effect of perceived justice, disconfirmation and
service failure attribution on recovery satisfaction. The study concluded that the joint
effect was statistically significant, meaning that jointly perceived justice, service failure
attribution and recovery disconfirmation are significant predictors of recovery
satisfaction. This means that perceived fairness and fulfillment of expectations in
handling service recovery impacts on customer satisfaction. To enhance recovery
satisfaction, providers of MMTS should consider these variables jointly.
5.4
Implications
The current research examined the relationship between perceived justice, recovery
disconfirmation, service failure attribution and recovery satisfaction. Specifically, the
study investigated the relationship between perceived justice and recovery satisfaction
and the intervening role of recovery disconfirmation in that relationship. Further, the
moderating role of service failure attribution in the relationship between perceived
justice and recovery satisfaction was examined. The joint effect of perceived justice,
recovery disconfirmation and service failure attribution on recovery satisfaction was
also investigated. The study results present theoretical, policy and managerial
implications.
138
5.4.1
Theoretical Implications
The study results provide support for the hypothesis that perceived justice predicts
recovery satisfaction. This is consistent with the commonly held view in the available
literature that there is a positive relationship between perceived justice and recovery
satisfaction and is in agreement with equity theory (Adams, 1965) which endeavors to
explain satisfaction based on perception of fairness in an exchange relationship
signifying that perception of fairness enhances satisfaction.
The study examined the mediating effect of recovery disconfirmation on the
relationship between perceived justice and recovery satisfaction and found it to be
positive therefore supporting the hypothesis. This finding adds to theory by introducing
recovery disconfirmation as a mediating variable in this relationship. The results
support the expectancy disconfirmation theory (Oliver, 1980) by endorsing its assertion
that satisfaction is enhanced when perceived performance matches expectations.
The research further established that service failure attribution has a significant
moderating influence on the relationship between perceived justice and recovery
satisfaction and that controllability attribution had the greatest contribution to that
moderating effect. This is consistent with attribution theory (Weiner, 2000), which
asserts that people perform attributional searches for causes of most negative events
such as service failure, along the dimensions of locus, stability and controllability.
Finally, the study confirmed that the joint effect of perceived justice, disconfirmation
and service failure attribution on recovery satisfaction was statistically significant. In
addition the findings support the cognitive dissonance theory which suggests that
139
human beings seek consistency in their beliefs and perception (Festinger, 1957).
Cognitive dissonance theory has been used in marketing to explain customer
dissatisfaction arising from the disparity between expectation and actual service
experience.
In summary, the results of this study add to theory by establishing a linkage between
perceived justice, recovery disconfirmation, service failure attribution and recovery
satisfaction. This has a theoretical implication as it integrates equity theory, attribution
theory, cognitive dissonance and expectancy disconfirmation model in the assessment
of customer satisfaction in service failure and recovery encounters. The results further
contribute to research interest on the role of these factors in service recovery and serve
to extend the prevailing discourse on the role of customer justice perceptions in
recovery satisfaction.
5.4.2
Policy Implications
The MMTS subsector is a critical component of the financial sector deepening and is
expected to contribute to financial inclusion in Kenya. The MMTS subsector arose out
of convergence of the financial and telecommunication sectors implying that regulators
for both sectors have a relevant policy role in the development of mobile money
transfer services. The Central Bank of Kenya (CBK) views mobile money with a prism
of financial sector deepening with the objective of enhancing financial access to the
unbanked and under-banked population in Kenya. The results of this study can be
useful to CBK when setting or reviewing policy on national payment systems and
particularly in developing guidelines for service providers regarding the handling of
customer complaints and redress systems for MMTS in Kenya.
140
The telecommunications sector regulator, Communication Authority of Kenya is likely
to apply the findings in developing regulations with reference to network performance
and customer protection. The study findings will also provide pertinent information to
support policy on service reliability, customer protection and service recovery
standards. As the current study demonstrates the influence of perceived justice, service
failure attribution and recovery disconfirmation on recovery satisfaction in the MMTS
sector, policy makers can use the results when revising service recovery guidelines.
This is important as recovery satisfaction is critical for the growth of communication
sector in Kenya.
The results of the study will support policy makers to develop and implement
appropriate policies to enforce standards of service delivery by providers. As Kenya is
the recognized leader in MMTS globally and other countries are seeking to replicate its
model, the study findings will be useful to both financial services and
telecommunications regulators in other countries in the region as they seek to develop
policies for entrenching mobile money services to promote financial inclusion.
5.4.3
Implications for Practitioners
Several noteworthy managerial implications arise from this study which provides
crucial insights into the key drivers of recovery satisfaction. The study established that
the perceived justice dimensions of procedural justice, interactional justice and
distributive justice do influence recovery satisfaction. The results further revealed that
distributive justice played a greater role in the recovery satisfaction judgment. This
implies that mobile money transfer service providers should seek to improve the
practices, procedures, distributive and interactional methods they use to rectify service
failure to achieve greater recovery satisfaction.
141
The results demonstrate the importance of ensuring fairness in the service recovery
process in MMTS sub-sector in Kenya. In addition, the study provides practitioners
with the relevant indicators for assessing recovery satisfaction for effective strategic
marketing decision-making. The findings suggest that service providers should
develop reasonable service recovery procedures and practices that address issues of
accessibility, fair queuing system and prompt response.
Service providers should also train the employees on how to manage interactions with
subscribers, provide relevant information, and demonstrate respect, courtesy, concern
and attentiveness during service recovery. Managers of service provider firms should
seek to enhance distributive fairness by developing clear guidelines on compensation
for service failure while training employees to quickly and properly react to various
service failure situations.
The finding that recovery disconfirmation mediates recovery satisfaction is important
for managers of mobile money service providers as they seek to enhance recovery
satisfaction. The managers need to develop effective systems for understanding
customer expectations and to design strategies to meet them as they implement service
recovery. They should train employees to ensure service recovery meets customer
expectations. The finding that service failure attribution moderates the relationship
between perceived justice and recovery satisfaction and that controllability attribution
contributes a greater role in this interaction has implications for practitioners.
To enhance recovery satisfaction managers should ensure that they address issues
within their control that are likely to cause service failure. This will help avoid the
142
perception by customers that the service failures experienced were preventable by the
organization. Overall, the study suggests that there is a need for service providers to
focus on perceived justice, service failure attribution and recovery disconfirmation in
order to improve recovery satisfaction.
5.5
Limitations of the Study
This study provided important insights into the perceived justice and recovery
satisfaction relationship among MMTS subscribers in Kenya. However, like all other
studies it has limitations that restrict the generalization of its findings. First the selection
of study variables that influence recovery satisfaction is not exhaustive. The factors
used in the conceptual framework to explain recovery satisfaction may not provide a
complete picture of the relationship between perceived justice and recovery
satisfaction. There are other factors that are relevant in understanding the recovery
satisfaction judgment which if added will provide additional insights on how to
augment recovery satisfaction.
A second limitation of this study relates to the use of a recall based survey in collecting
data to evaluate perceived justice, service failure attribution, recovery disconfirmation
and recovery satisfaction. Although customer satisfaction is normally assessed through
self-reports the recall approach requires customers to dig into their memory to recall
the experience. As such a study conducted immediately after the service recovery may
provide a more robust perspective on the perception of justice, service failure
attribution, recovery disconfirmation and recovery satisfaction.
Third, the study used cross-sectional survey design where respondents are interviewed
on the variables of the study at a single point in time. While cross sectional study is
143
commonly used in marketing studies and provides cost and time advantages while
allowing for generalization of findings, it does not support investigation of certain
aspects of customer perception as would be provided by data gathered at several points
over time.
Fourth, the study was based on customer responses about their experience with service
failure and recovery. Since service failure is associated with emotions of anger and
frustration (Rio-Lanza et al., 2009) it is likely that this may affect customer reporting
of service recovery. In addition, the choice of questions used in the study may not have
included all possible alternatives. Finally, the study only focused on one subsector of
financial services (mobile money transfer services) hence restricting the
generalizability of the findings to other financial services.
5.6
Suggestions for Further Research
The results of this study contribute to understanding the influence of perceived justice,
service failure attribution and recovery disconfirmation on recovery satisfaction. In
future research, the inclusion of other factors not covered in this study may improve
the robustness of the conceptual model and enhance understanding of the recovery
satisfaction judgment. Future research may consider testing the moderating effect of
additional variables such as customer personality and expectation of relationship
continuity.
Future research could also consider the collection of data immediately following
service failure and recovery to minimize the effect of time rag and the risk of forgetting.
In addition, the use of longitudinal data may provide a more complete picture in the
evaluation of the perceived justice and recovery satisfaction relationship over time.
144
Further, future research may also consider replication of this study in other financial
services or other contexts. This will help broaden the available knowledge on the
perceived justice and recovery satisfaction relationship. Finally, the replication of this
study in other countries in the region would help confirm the universality of the
influence of perceived justice, service failure attribution and recovery disconfirmation
on recovery satisfaction in general and in mobile money transfer services in particular.
145
REFERENCES
Aaker, D., Kumar, V. Day G. (2007). Marketing Research (9
th
ed.). New Delhi: John Wiley
Sons.
Adams, J.S. (1965). Inequity in social exchange. In Berkowitz, L. (Ed). Advances in
Experimental Social Psychology, Volume 2, New York: Academic Press.
AFI (2010). Enabling mobile money transfer: The CBK treatment of M-Pesa. Retrieved June
11,
2015
http://www.afi-
global.org/sites/default/files/publications/afi_casestudy_mpesa_en.pdf
Andreassen, T. W. (2000). Antecedents to Satisfaction with Service Recovery, European
Journal of Marketing, 4(1/2), 156-175.
Balasubramanian, K. Drake, D (2015) Mobile Money: The Effect of Service Quality and
Competition
on Demand. Harvard Business School Technology Operations Mgt.
Unit
Working
Paper
No.
15-059.
Retrieved
June
22,
2015
from,
http://dx.doi.org/10.2139/ssrn.2549171
Baron R. Kenny D. (1986). The moderator- mediator variable distinction in social
psychological Research: Conceptual and statistical considerations. Journal of
Personality and Social Psychology, 51 (6), 1173-1182.
Bassili, J. N. Fletcher, J. F. (1991) Response time measurement in survey research: a method
for CATI and a look at non-attitudes. Public Opinion Quarterly, 55(3) 331-346.
Brayman, A. (2012) Social Research Methods, (4th ed.). New York: Oxford University Press.
Blodgett, J. G., Hill, D. J., Tax, S. S. (1997). The effects of distributive, procedural, and
interactional justice on post complaint behavior. Journal of retailing, 73 (2),185-210.
Carmines, E. G. Zeller, R. A. (1979). Reliability and Validity Assessment, Beverly Hills,
CA: Sage.
Carrell, M. R. Dittrich, J. E. (1978). Equity theory: The recent Literature: M
ethodological
consideration and new directions. Academy of Management Review, 3, 202-210.
Cacioppo J.T, Larsen J.T Smith N.K Berntson G. G. (2004). The Affect System: What lurks
below the surface of feelings? In: Manstead A.S., Frijda N.H., Fischer, A.H. (Eds).
Feelings and Emotions. New York: Cambridge University Press.
Chang, H.H., Lai, M.K. Hsu, C.H. (2012). Recovery of online service: Perceived justice
and
transaction frequency, Computers in Human Behavior, 28 (6), pp. 2199-2208.
Collie, T. A., Sparks B. Graham B. (2000). Investing in Interactional Justice: A Study of
the
Fair Process Effects within a Hospitality Failure Context. Journal of Hospitality and
Tourism
Research,
24, 448-472.
Communications Commission of Kenya, Quarterly sector statistics Report, 2014
146
Communications Commission of Kenya, Quarterly sector statistics Report, 2015
Central Bank of Kenya (2014) Statistical Bulletin. Retrieved January 28, 2015, from
https://www.centralbank.go.ke/index.php/retail-payments-2/mobile-payments?yr=2014
CGAP (2009). Mobile payments in Kenya: Findings from a survey of M-Pesa users and agents.
Retrieved May 7, 2015, from
http://www.fsdkenya.org/pdf_documents/11-02-
14_Mobile_payments_in_Kenya.pdf
Chebat, J. C. Slusarczyk, W. (2005). How emotions mediate the effects of perceived justice
on loyalty in service recovery situations: an empirical study. Journal of Business
Research, 58(5), 664-73.
Chepkwony, J., Korir M., Lagat, C, Mumbo, H. Odera, O., (2012). Effects of Distributive
Justice Complaints Resolution Strategies on Customer Satisfaction in Kenya's
Banking Industry. European Journal of Business and Social Sciences, 1(6), 87-96.
Colquitt, J. (2001). On the dimensionality of organizational justice: A construct validation of a
measure. Journal of Applied Psychology, 86(3), 386400.
Cooper, D.R. Schindler, P.S. (2006). Business Research Methods, (9th ed.). New Delhi:
Tata McGraw- Hill.
Davidow, M. (2003). Organizational responses to customer complaints: what works and what
doesn't? Journal of Service Research, 5(3), 225-50.
Ellyawati J., Purwanto, B. M. Dharmmesta B. S. (2012). The Effect of Perceived Justice on
Customer Satisfaction in the Context of Service Recovery: Testing Mediating Variables.
Journal of Service Science, 5(2), 87-99.
Fairchild A. MacKinnon D. (2009). A General Model for Testing Mediation and Moderation
Effects. Prevention Science, 10, 87-99.
Fisk, R.P. Young, C.E. (1985). Disconfirmation of Equity expectations: effects on consumer
satisfaction with services. In, Hirschman, E.C. Holbrook, M. (Eds.), Advances in
Consumer Research, Provo: Association for Consumer Research.
Folkes, V.S. (1984). Consumer Reactions to Product Failure: An Attributional Approach.
Journal of Consumer Research, 10 (3), 398-409.
Ghalandari, K., Babaeinia, L. Ghaorbani, G. J. (2012). Investigation of perceived justice on
post-recovery overall satisfaction, post-recovery revisit intention, post-recovery word
of
mouth intention from Airline Industry in Iran: the role of corporate image. World
Applied
Science Journal, 18 (7), 957-970.
GoK, (2007) Kenya Vision 2030: the Popular Version. Retrieved 14th March, 2013
from,
http://www.vision2030.go.ke/cms/vds/Popular_Version.pdf
Greenberg, J. (1990). Looking Fair Versus Being Fair: Managing Impressions of
Organizational
Justice. In, Staw B.M. and Cummings, L.L. (Eds.), Research in
Organizational
Behavior, Greenwich, CT: JAI Press, 11157.
147
GSMA (2013). Global Mobile Money Adoption Survey Report. Retrieved 22
nd
January, 2014
http://www.gsma.com/mobilefordevelopment/wp-content/uploads/2013/03/MMU
GSMA (2014). Reaching half of the market: women and mobile money. Retrieved 2
nd
September, 2015
http://www.gsma.com/mobilefordevelopment/wpcontent/uploads/2014/09/2014_DI_Re
ach-half-
of-the-market-Women-and-mobile-money.pdf
GSMA (2015)
.
Enabling mobile money policies in Kenya: Fostering a Digital Financial
Revolution.
Retrieved
June
5,
2015
from,
http://www.gsma.com/mobilefordevelopment/wpcontent/uploads/2015/02/2015_MMU
Ha, J., Jang, S. (2009). Perceived justice in service recovery and behavioral intentions: The
role of relationship quality. International Journal of Hospitality Management, 28, 319
327.
Harris, K. E., Mohr, L. A. Bernhardt,
K. L. (2006). Online service failure, consumer
attributions and expectations. Journal of Services Marketing, 20 (7), 453 458.
Harris L.C. and Ogbonna E. (2006). Service Sabotage: A Study of Antecedents and
Consequences. Journal of the Academy of Marketing Science. 34 (4), 543-
558.
Hess, R. L., Ganesan, S. Klein, N. M. (2003). Service failure and recovery: The impact of
relationship factors on customer satisfaction. Journal of the Academy of Marketing
Science, 31(2), 127145.
Hocutt, M.A., Bowers, M.R. Donavan, D.T. (2006). The art of service recovery: fact or
fiction? Journal of Service Marketing, Vol.20, No.3, 199207
Hocutt M A, Chakraborty, G. Mowen J. C., (1997). The Impact of Perceived Justice on
Customer Satisfaction and Intention to Complain. Advances in Consumer Research,
2:457-463.
Hoffman, D.K Kelly, S.W (2000). Perceived Justice Needs a Recovery Evaluation: A
Contingency Approach. European Journal of Marketing, 34 (3) 418-429
Huseman, R.C., Hatfield, J. D. Miles, E.W. (1987). A New Perspective on Equity Theory:
The Equity Sensitivity Construct. The Academy of Management Review. 12 (2), 222-234.
IEA (2011). Institute of Economic Affairs. The State of Competition Report: Mobile Money
Transfer, Agricultural Bulk Storage and Milling, and the Media Sectors in Kenya,
Research Paper Series No.
1(1).
Retrieved
February
26,
2013
from,
http://www.ieakenya.or.ke
IFRC (2012). The International Federation of the Red Cross and Crescent Societies. Mobile
phone-based surveys can have it all. Retrieved April 15, 2015 from,
https://www.ifrc.org/PageFiles/96432/1229700-RAMP%20advocacy%20report-
EN_LR.
148
Intermedia (2014). Digital Pathways to Financial Inclusion: Findings from the First FII Tracker
Survey in Kenya. Retrieved June 2, 2015 from
http://finclusion.org/wp-
content/uploads/2014/04/FII-
Kenya-Wave-One-Wave-Report.pdf
Kanousi, A. (2005). An empirical investigation of the role of culture on service recovery
expectations. Managing Service Quality, 15 (1), 57-69.
Kau, A. K. Loh, E. W. (2006). The effects of service recovery on consumer satisfaction: a
comparison between complainants and non-complainants. Journal of Services
Marketing, 20(2), 101-11.
Keaveney, S.M. (1995). Customer Switching Behavior In Service Industry: An Exploratory
Study. Journal of Marketing. 59(2), 71-82.
Kelley, S.W., Douglas H. and Mark A. D. (1993). A Typology of Retail Failures and
Recoveries. Journal of Retailing. 69(4), 429-452.
Kim, T., Kim, W. G., Kim, H. B. (2009). The effects of perceived justice on recovery
satisfaction, trust, word-of-mouth, and revisit intention in upscale hotels. Journal of
Tourism Management, 30, 5162.
Komunda M. B.
Osarenkhoe, A. (2012) Remedy or cure for service failure? Effects of
service recovery on customer satisfaction and loyalty, Business Process Management
Journal, 18 (1), 82-103.
Kotler, P. Keller, K.
(2012)
A Framework for Marketing Management, Fourth Edition,
Prentice Hall. Pearson Education.
Krejcie, R. V. Morgan D. W. (1970) Determining sample size for Research Activities.
Educational and Psychological Measurement, 30, 607-610.
Kinoti, M.W. (2012), Green marketing practices, corporate image, organizational
characteristics and performance of ISO 9000 and 14000 certified organizations in
Kenya. Unpublished PhD Thesis, University of Nairobi, Kenya.
Krishna A., Dangayach, G.S. Jain R. 2011. Critical factors of service failure and
corresponding recovery system, Advances in Management, 4(7), 3944.
Levesque, T., McDougall, G. (2000). Service problems and recovery strategies: An
experiment. Canadian Journal of Administrative Sciences. 1, 20-37.
Magnini, V. P, John B., Edward, P Earl D, 2007), The service recovery paradox: Justifiable
theory or smoldering myth? Journal of Science Marketing; 2(21), 213-225.
Mattila, A.S. (2001), The effectiveness of service recovery in a multi-industry setting, Journal
of Services Marketing, 15(7), 583-96.
Maxham, J.G. Netemeyer, R. (2002).
Modeling customer perceptions of complaint handling
over time: the effects of perceived justice on satisfaction and intent.
Journal of Retailing
78(2)239-252.
149
Maxham, J.G. (2001), Service recovery's influence on consumer satisfaction, positive word-
of-mouth and purchase intentions, Journal of Business Research, 54, 11-24.
Mayombo, M. B. (2014). The influence of customer complaint behavior, firm responses and
servicequality on customer loyalty of mobile telephone subscribers in Uganda.
Unpublished PhD Thesis, School of Business, University of Nairobi.
McColl-Kennedy, J. R. Sparks, B. A. (2003). Application of fairness theory to service
failures and service recovery. Journal of Service Research, 5, 251-67.
McCollough, M.A., Berry, L.L. Yadav, M.S. (2000). An Empirical Investigation of
Customer Satisfaction after Service Failure and Recovery, Journal of Service Research,
3(2), 121137.
Michel, S. (2001). Analyzing service failures and recoveries: a process approach. International
Journal of Service Industry Management. 12(1), 20 33
Munyoki J. M. (2007). The effect of technology transfer on the organizational performance: A
study of medium and large firms in Kenya. Unpublished PhD Thesis, School of Business,
University of Nairobi.
Namkung Y. and Jang S. (2009). The Effects of International Fairness on Satisfaction and
Behavioral Intentions: Mature versus Non-Mature Customers. International Journal
of
Hospitality Management. 28, 397-405.
Ndungu, K. (2013). Quality drivers, managerial focus, customer perception and satisfaction in
large maize flour mills in Nairobi, Kenya. Unpublished PhD Thesis, School of Business,
University of Nairobi.
Nibkin D., Ismail, I. Abu-Jarad, I. Y. (2011). The impact of firm reputation on customer's
responses to service failure: the role of failure attributions. Business Strategy Series,
12(1), 19 -29.
Nibkin, D., Ishmail, I., Mairimuthu, M. Jalakamali, M., (2010). Perceived Justice in Service
Recovery and Recovery Satisfaction: The Moderating Role of Corporate Image,
International Journal of Marketing Studies, 2(2) 47-56.
Nikbin, D., Ismail, I., Marimuthu, M. Slarzehi, H. (2012). The relationship of service
failure
attributions, service recovery justice and recovery satisfaction in the context of
airlines.
Scandinavian Journal of Hospitality and Tourism, 12 (3), 232-254.
Njeru, W.C. (2013). Market orientation, marketing practices, firm characteristics, external
environment and performance of tour firms in Kenya.
Unpublished PhD Thesis, School
of Business, University of Nairobi.
Nunnally, J. C. Bernstein, I. H. (1994). Psychometric Theory. New York: McGraw- Hill.
Oliver, R. L. (1980). A cognitive model of the antecedents and consequences of satisfaction
decisions, Journal of Marketing Research, 16(June), 3954.
150
Oliver, R. L. (1997). Satisfaction: A Behavioral Perspective on the Consumer. New York:
McGraw Hill.
Omwansa, T.K. Sullivan N. P. (2012). Money Real Quick: The story of M-pesa. London:
Guardian Books.
Orsingher, C., Valentini, S. De Angelis, M. (2010). A meta-analysis of satisfaction with
complaint handling in services, Journal of the Academy of Marketing Science, 38, 169
186.
Ostrom , A. Iacobucci D. (1995). Consumer trade-offs and the evaluation of services.
Journal
of Marketing, 59 (1), 17-28.
Rio-Lanza, A. B., Vazquez-Casielles, R. Diaz-Martin A.M. (2009). Satisfaction with service
recovery: Perceived justice and emotional responses, Journal of Business Research,
62(8), 775-781.
Rashid, M. Ahmad, F. (2014).
The Role of Recovery Satisfaction on the Relationship
between Service Recovery and Brand Evangelism: A Conceptual Framework,
International Journal of Innovation, Management and Technology, 5(5), 401-405.
Schiffman L.G. Kanuk L.L. (2007). Consumer Behavior 9th Ed. Prentice Hall India.
Seawright K.K, DeTienne K.B, Bernhisel, M.P. Larson C.L.H (2008) An empirical
examination of service recovery design, Marketing Intelligence and Planning, 26(3)
253-74.
Sekaran, U. (2003). Research Methods for Business: A Skill Building Approach. New York,
John Wiley and Sons.
Serenko, A. Stach, A. (2009). The impact of expectation disconfirmation on customer loyalty
and recommendation behavior: investigating online travel and tourism services. Journal
of Information Technology Management, 20(30), 26-41)
Smith, A. K. Bolton, R. N. (2002). The effect of customers' emotional responses to service
failures on their recovery effort evaluations and satisfaction judgments. Journal of the
Academy of Marketing Science, 30(1), 523.
Smith, A.K., Bolton, R.N. Wagner, J. (1999). A Model of Customer Satisfaction with Service
Encounters Involving Failure and Recovery. Journal of Marketing Research, 36 (8) 356
372.
Sparks, B. Fredline, L. (2007). Providing an explanation for service failure: context,
content
and customer responses, 31 (2), 241-260.
Spreng, R., MacKenzie A., Olshavsky, R. (1996). A reexamination of the determinants of
consumer satisfaction. Journal of Marketing, 60, 15-32.
Stephens, R. E., Wremn, B. Sherwod, P. K. (2006). The Marketing Research Guide, 2nd
Edition. New York, The Haworth Press.
Spector, 2008. Spector, P.E. (2008). Industrial and Organizational Behavior (5th ed.). Wiley:
Hoboken, NJ.
151
Susskind, A. M. (2005). A content analysis of consumer complaints, remedies, and re-
patronage intentions regarding dissatisfying service experience, Journal of Hospitality
and Tourism Research, 29(2), 150169.
Swanson, S. Hsu, M. K. (2011). The Effect of Recovery Locus Attributions and Service
Failure Severity on Word-of-Mouth and Repurchase Behaviors in the Hospitality
Industry. Journal of Hospitality Tourism Research, 35(4), 511-529.
Swanson, S. Kelley, S. W. (2001). Service Recovery Attributions and Word-of-Mouth
Intentions, European Journal of Marketing, 35(1/2), 194-211.
Sweeney, J.C.; Hausknecht D. Soutar G.N. (2000) Cognitive dissonance after purchase: A
multidimensional scale. Psychology Marketing 17 (5), 369385.
Tan, A.G.N. (2014) Effects of Justice Theory on Service Recovery Satisfaction on Metro
Manila Dine-in Experiences. Philippine Management Review 21 (1), 25-38.
Tax, S.S., Brown, S.W., Chandrashekaran, M. (1998). Customer evaluation of service
complaint experience: Implications for relationship marketing. Journal of Marketing,
62(4), 60-77.
Trope, Y. Gaunt, R. (2003). Attribution and person perception. In M. A. Hogg J. Cooper
(Eds), Handbook of Social Psychology. New York: Sage Publications.
USAID (2011). Better than cash: mobile money market assessment. Retrieved 5
th
June, 2015
from: http://nethope.org/assets/uploads/Kenya-Mobile-Money-Assessment.pdf
Wang, C. Mattila, A.S. 2011. `A cross-cultural comparison of perceived informational
fairness with service failure explanations', Journal of Services Marketing, 75(6): 429
439.
Weber, K. Sparks, B. (2009). The Effect of Pre-consumption Mood and Service Recovery
Measures on Customer Evaluations and Behavior in a Strategic Alliance Setting, Journal
of Hospitality Tourism Research, 33, 106-125.
Weber, K. Sparks, B. (2010).
Service Failure and Recovery in a Strategic Airline Alliance
Context: Interplay of Locus of Service Failure and Social Identity. Journal of Travel
Tourism Marketing, 27(6), 547-564.
Webster, C. Sundaram, D.S. (1998). Service consumption criticality in failure recovery.
Journal of Business Research, 41, 153-159.
Weiner B. (2000). Intrapersonal and Interpersonal Theories of Motivation from an
Attributional Perspective, Educational Psychology Review, 12(1), 101-114.
Weun, S., Beatty, S., Jones, M. (2004). The impact of service failure severity on service
recovery evaluations and post-recovery relationships. Journal of Services Marketing,
18,
133-146.
Wirtz, J. Mattila, A. (2001). Exploring the role of alternative performance measures and
needs-congruency in the consumer satisfaction process, Journal of Consumer
Psychology, 11(3), 181- 192.
152
Wirtz, J. Mattila, A. S. (2004). Consumer responses to compensation, speed of recovery
and apology after a service failure. International Journal of Service Industry
Management, 15(2) 150-167.
World Bank (2014). Rethinking the Role of the State in Finance. Global Financial
Development Report. 9http://www.worldbank.org/financialdevelopment Retrieved 19
th
January 2015
Yim, C., Gu, F., Chan, K., Tse, D. (2003). Justice-based service recovery expectations:
Measurement and antecedents. Journal of Consumer Satisfaction, Dissatisfaction and
Complaining Behavior, 16, 36-52.
Zeithaml, V. A., Berry, L.L. Parasuraman A. (1993). The Nature and Determinants of
Customer Expectations of Service. Journal of the Academy of Marketing Science. 21
(1),
112.
153
APPENDICES
APPENDIX 1A
MOBILE MONEY TRANSFER SERVICES QUESTIONNAIRE
SECTION I: SCREENING
Q1.
Which mobile phone service provider do you use most frequently?
Q1
Safaricom
1
Airtel
2
Orange
3
YU
4
Q2.
Which mobile money transfer service do you use frequently and for how long have you
subscribed to that Mobile Money transfer service?
MMTS used
frequently
Tick
()
Period of
Subscription
Tick ()
M-Pesa
1
Up to 6 months
Airtel
Money
2
7-12 months
Orange
Money
3
1-5 Years
YU Cash
4
5 years
None
Close interview
154
SECTION II: SERVICE FAILURE/ RECOVERY ENCOUNTER
Q3.
Have you experienced any of the following service related problems with your
mobile money transfer service during the past 6 months?
i. Sending money to the wrong number
Yes
No
ii. Sending the same transfer twice unintentionally
Yes
No
iii. No notification message after transferring money to another
number or after buying airtime or bundles
Yes
No
iv. No response when trying to use/ access the money transfer
menu
Yes
No
v. Transaction taking long to confirm/delay in processing
transaction
Yes
No
vi. Problem activating banking services e.g. Mshwari account
Yes
No
vii. Agent dishonesty e.g. giving less money; charging more
than recommended
Yes
No
viii. Bill payment menu not responding
Yes
No
ix. Delay in confirming bill payment for KPLC, DSTV, Water
etc.
Yes
No
x. Unable to change personal identification number (PIN)
xi. Receipt of small amounts of money from con men
Yes
No
xii. Receipt of mobile money transfers from strangers
Yes
No
xiii. Agent lack of cash/float
Yes
No
xiv. Mobile bank agent service breakdown
Yes
No
xv. Lack of accessibility to service agent
Yes
No
xvi. Delayed reversal of funds (after wrong transaction)
Yes
No
xvii. Unavailability of mobile network service
Yes
No
Others, please specify
Yes
No
If no to all
close
interview
155
Q4a.
Did you report the issue you have just mentioned to the service provider with the
expectation of getting a solution?
Yes...1 CONTINUE No...........2 close interview
Q4b. Please state the service related problem which you reported to the service provider for
correction?
Service related problem
Tick
Sending money to the wrong number
Sending the same transfer twice unintentionally
No notification message after transferring money to another number or after
buying airtime or bundles
No response when trying to use/ access the money transfer menu
Transaction taking long to confirm/delay in processing transaction
Problem activating banking services e.g. Mshwari account
Agent dishonesty e.g. giving less money; charging more than recommended
Bill payment menu not responding
Delay in confirming bill payment for KPLC, DSTV, Water etc.
Unable to change personal identification number (PIN)
Receipt of small amounts of money from con men
Receipt of mobile money transfers from strangers
Agent lack of cash/float
Mobile bank agent service breakdown
Lack of accessibility to service agent
Delayed reversal of funds (after wrong transaction)
Unavailability of mobile network service
Others, Please specify
156
SECTION III: SERVICE FAILURE ATTRIBUTION
Q5.
Relating to the complaint/problem you experienced, to what extent would you say
that
Not
at all
To a
small
extent
To a
moderate
extent
To a
large
extent
To a very
large extent
a. The Company was to blame for
the problem(s)
1
2
3
4
5
b. You were to blame for the
problem
1
2
3
4
5
c. This problem occurs frequently 1
2
3
4
5
d. The company could have
prevented the problem
1
2
3
4
5
e. The problem was caused by the
agent.
1
2
3
4
5
Other (Please specify)
157
SECTION IV: PERCEIVED JUSTICE
Q6.
When you contacted the service provider to solve the problem to what extent would
you say...:
Procedural Justice
Not
at all
To a
small
extent
To a
moderate
extent
To a
large
extent
To a very
large
extent
a.
It was easy to know
where to lodge your
complaint
1
2
3
4
5
b.
You waited for a short
time before connecting
with the customer care
staff
1
2
3
4
5
c.
The company employee
responded quickly to
your problem
1
2
3
4
5
d.
The queuing system is
fair
1
2
3
4
5
f.
The company practices
for addressing customer
problems are reasonable
1
2
3
4
5
Interactional Justice
a.
You were treated with
respect
1
2
3
4
5
b.
The employee listened
attentively to your
complaint
1
2
3
4
5
c.
The employee apologized
for the occurrence
1
2
3
4
5
d.
The employee provided a
useful information about
the problem
1
2
3
4
5
f.
The employee showed
concern about the
problem
1
2
3
4
5
158
Distributive Justice
a.
The way the problem was
solved resulted in a
positive outcome for you
1
2
3
4
5
b.
The solution provided
was worth the effort you
made to seek correction
1
2
3
4
5
c.
The company offered you
what you deserved
1
2
3
4
5
d.
The compensation for
your problem was fair
1
2
3
4
5
e.
The transaction was
carried out in a fair
manner
1
2
3
4
5
Other, Specify
1
2
3
4
5
159
SECTION V: RECOVERY DISCONFIRMATION
Q7.
To what extent would you say that ...:
Not
at
all
To a
small
extent
To a
moderate
extent
To a
large
extent
To a very
large extent
a.
The problem was handled
the way you expected
1
2
3
4
5
b.
The solution met your
expectations
1
2
3
4
5
c.
The company met your
expectations
1
2
3
4
5
SECTION VI: RECOVERY SATISFACTION
Q8. How satisfied were you with:
Not
satisfied
at all
Slightly
satisfied
Satisfied Very
satisfied
Very
much
satisfied
a. The service provided
1
2
3
4
5
b. The corrective action taken
1
2
3
4
5
c. The overall service
1
2
3
4
5
d.
Overall, how satisfied were
you?
1
2
3
4
5
e To what extent did you feel
confident to continue using
the services of the company?
1
2
3
4
5
f
How likely are you to
recommend your Mobile
Money Transfer Service to a
friend or colleague:
1
2
3
4
5
160
SECTION VII: RESPONDENT DEMOGRAPHICS
GENDER
AGE-
GROUP
County of
residence
MMTS
Education
Level
Male
1 18
1
M-
PESA
1 Primary
1
Female
2 18-24
2
Airtel
Money
2
Secondary
2
25-34
3 Occupation YU
Cash
3 College/diploma
3
35-44
4
Graduate +
4
45+
5
Q9. Is there anything else you would like to say on this subject of how the service provider
solves problems of mobile money transfer service failure?
CLOSE INTERVIEW AND THANK RESPONDENT
161
APPENDIX 1B
KISWAHILI QUESTIONNAIRE
HOJAJI YA HUDUMA ZA UHAMISHAJI WA FEDHA KWA NJIA YA SIMU
SEHEMU YA I: UKAGUZI
Q1.
Unatumia huduma zipi za simu za mkononi mara nyingi?
Q1
Safaricom
1
Airtel
2
Orange
3
YU
4
Q2.
Je, ni huduma gani ya uhamishaji wa fedha kwa simu unayotumia mara nyingi na
umejiandikisha kwa huduma hii kwa muda mgani?
HUDUMA ZA
UHAMISHAJI
WA FEDHA
INAYOTUMIWA
MARA NYINGI
Weka
alama
()
Muda aliyotumia
huduma hii
Weka alama ()
M-Pesa
1
Kwa muda wa hadi
Miezi 6
Airtel
Money
2
Miezi 7-12
Orange
Money
3
Miaka 1-5
YU Cash
4
Zaidi ya miaka 5
Hakuna
Sitisha
mahojiano.
162
SEHEMU YA II: KUKOSEKANA KWA HUDUMA/HALI YA KUREJESHEWA
PESA.
Q3.
Umewahi kukumbana na matatizo yafuatayo kuhusiana na huduma ya kuhamisha
fedha kwa muda wa miezi sita iliyopita?
i.
Kutuma pesa kwa mtu mwingine usiyemjua
kimakosa?
Ndio La
ii.
Kutuma pesa mara mbili bila ya kukusudia
Ndio La
iii.
Kutopata ujumbe wa kukufahamisha baada ya
kutuma fedha au kuongezea muda wako wa
maoengezi au kuweka data
Ndio La
iv.
Kutopata majibu unapojaribu kutumia au kupata
menyu ya huduma za uhamishaji wa fedha
Ndio La
v.
Kuchukua muda mrefu kudhibitisha huduma au
kuchelewesha mchakato wa kufanya uhamishaji au
utoaji wa fedha
Ndio La
vi.
Shida za kuanzisha huduma za kubenki kwa mfano:
akaunti ya Mshwari
Ndio La
vii.
Kutoaminika kwa ajenti- kwa mfano: kukupa pesa
china ya kiwango kinacho faa, kulipisha Zaidi ya
kiwango kilicho pendekezwa
Ndio La
viii.
Kutopatikana kwa menyu ya malipo ya bili
Ndio La
ix.
Ucheleweshaji wa kudhibitisha malipo ya bili ya
stima, DSTV, maji n.k
Ndio La
x.
Kushindwa kubadilisha nambari yako ya siri (PIN)
Ndio La
xi.
Kupokea pesa cha kiasi kidogo kutoka kwa
walangusi
Ndio La
xii.
Kupokea pesa kutoka kwa watu usiojua
Ndio La
xiii.
Mtandao wa simu kukosekana/agent kukosa
pesa/float
Ndio La
xiv.
Kipindi cha simu ya benki kukosa kufanya
Ndio La
xv.
Agent kukosekana
Ndio La
xvi.
Urudishaji was pesa kuchelewa
Ndio La
xvii.
Mtandao wa simu kukosekana
Ndio La
xviii.
Nyingine, tafadhali elezea
ikiwa jawabu ni
La kwa maswali
163
yote, sitisha
mahojiano.
Q4.a
Uliripoti shida uliyotaja hapo juu kwa mtoa huduma hizo kwa matarajio ya kupata suluhu?
Ndio ...1 Endelea
La...2 Sitisha mahojiano.
Q4 b. Tafadhali elezea matatizo gani uliripoti
kwa mtoa huduma hizo kwa matarajio ya kupata
suluhu?
Matatizo
Weka
alama ()
Kutuma pesa kwa mtu mwingine usiyemjua kimakosa?
Kutuma pesa mara mbili bila ya kukusudia
Kutopata ujumbe wa kukufahamisha baada ya kutuma fedha au kuongezea muda
wako wa maoengezi au kuweka data
Kutopata majibu unapojaribu kutumia au kupata menyu ya huduma za uhamishaji
wa fedha
Kuchukua muda mrefu kudhibitisha huduma au kuchelewesha mchakato wa
kufanya uhamishaji au utoaji wa fedha
Shida za kuanzisha huduma za kubenki kwa mfano: akaunti ya Mshwari
Kutoaminika kwa Ajenti- kwa mfano: kukupa pesa china ya kiwango kinacho faa,
kulipisha Zaidi ya kiwango kilicho pendekezwa
Kutopatikana kwa menyu ya malipo ya bili
Ucheleweshaji wa kudhibitisha malipo ya bili ya stima, DSTV, maji n.k
Kushindwa kubadilisha nambari yako ya siri (PIN)
Kupokea pesa cha kiasi kidogo kutoka kwa walangusi
Kupokea pesa kutoka kwa watu usiojua
Mtandao wa simu kukosekana/Agent kukosa pesa/float
Kipindi cha simu ya benki kukosa kufanya
Agent kukosekana
Urudishaji wa pesa kuchelewa
Mtandao wa simu kukosekana
Nyingine, tafadhali elezea
164
SEHEMU YA III: UHUSISHO WA KUTOFANIKIWA KWA HUDUMA
Q5
Kuhusiana na shida/malalamishi uliokumbana nayo, ni kwa kiwango kipi unaweza
kusema kuwa
Hakuna
hata
kidogo
Kwa
kiasi
kidogo
Kwa kiasi cha
wastani
Kwa kiasi
kikubwa
Kwa kiasi
kikubwa
Zaidi
a.
Kampuni inafaa kulaumiwa kwa
shida hiyo
1
2
3
4
5
b.
Unafaa kujilauma kwa shida
uliokumbana nayo
1
2
3
4
5
c.
Shida hii hutokea mara kwa
mara
1
2
3
4
5
d.
Kampuni ingeweza kuzuia shida
hii
1
2
3
4
5
e
Shida ilisababishwa na Agenti
nyingine, elezea
1
2
3
4
5
SEHEMU YA IV: MAONI KUHUSU HAKI
Q6.
Ulipowasiliana na Mtoa huduma hii kuhusu kusuluhisha shida iliyotokea, ni kwa
kiasi kigani unaweza kusema ...:
Utaratibu wa kupata
haki
Hakuna
hata
kidogo
Kwa
kiasi
kidogo
Kwa kiasi
cha
wastani
Kwa
kiasi
kikubwa
Kwa kiasi
kikubwa
a.
Ilikuwa ni rahisi sana
kujua mahali pa kuelezea
malamishi uliokuwa nayo
1
2
3
4
5
b. Ulingojea kwa muda
mfupi mno kabla ya
kupata mhudumu katika
huduma za wateja
1
2
3
4
5
c.
Mhudumu wa kampuni
aliharakisha
kushughulikia shida
uliyokuwa nayo
1
2
3
4
5
165
d. Utaratibu wa kupanga
laini ulikuwa mzuri
1
2
3
4
5
f.
Jinsi kampuni
inavyoshughulikia shida
za wateja ni vyema
1
2
3
4
5
Haki ya kimahusiano
a.
Ulihudumiwa kwa
heshima
1
2
3
4
5
b. Mhudumu alisikiliza kwa
makini shida uliokuwa
nayo
1
2
3
4
5
c.
Mhudumu aliomba
msamaha kwa sababu ya
yaliyojiri
1
2
3
4
5
d. Mhuduma alikupa
ujumbe wa kusaidia sana
kuhusiana na shida yako
1
2
3
4
5
f.
Mhudumu alionyesha
kujali kuhusiana na shida
yako
1
2
3
4
5
Haki ya kiusuluhisho
a.
Jinsi shida
ilivyoshughulikiwa
ulipata matokea mema
1
2
3
4
5
b. Suluhu uliopata ilikua la
dhamana kwa jitihada
ulizofanya katika
kutafuta marekebisho
1
2
3
4
5
c.
Kampuni ilikupa
ulichohitaji
1
2
3
4
5
d. Suluhisho kwa shida
yako ilikuwa ya haki
1
2
3
4
5
e.
Shughuli yote ilifanywa
kwa njia ya haki
1
2
3
4
5
Nyingine, elezea
1
2
3
4
5
166
SEHEMU YA V: KUTODHIBITISHA UREJESHI
Q7.
Ni kwa kiasi kigani unaweza kusema kuwa ...:
Hakuna
hata
kidogo
Kwa
kiasi
kidogo
Kwa kiasi
cha
wastani
Kwa
kiasi
kikubwa
Kwa
kiasi
kikubwa
a.
Shida uliyokuwa nayo
ilishughulikiwa kulingana
na matarajio yako
1
2
3
4
5
b.
Suluhisho uliopata ilikutana
na matarajio yako
1
2
3
4
5
c.
Kampuni iliweza kukutana
na matarajio yako
1
2
3
4
5
SEHEMU YA VI: KURIDHIKA NA UREJESHI
Q8. Uliridhika kiasi kuhusiana na:
Haukuridhika
hata kidogo
Uliridhika
kwa kiasi
uliridhika Uliridhika
kabisa
Uliridhika
kabisa
sana
a. Huduma iliyotolewa
1
2
3
4
5
b. Kuhusiana na
marekebisho
yalichukuliwa
1
2
3
4
5
c. Huduma yote kwa jumla 1
2
3
4
5
d.
Kwa jumla, uliridhika
kiasi gani?
1
2
3
4
5
e Nieleze ni kwa kiwango
gani ambapo ulipata
uaminifu wa kuendelea
kutumia huduma za
kampuni hii?
1
2
3
4
5
167
SEHEMU YA VII: PROFAILI YA MSAILIWA
JINSI
A
KIKUND
I CHA
UMRI
KAUNTI
UNAYOKA
A
HUDUMA
YA
UHAMISHAJ
I WA FEDHA
Kiwang
o cha
masomo
Mume
1 18
1
M-
PESA
1
Shule ya
msingi
1
Mke
2 18-24
2
Airtel
2
Shule ya upili
2
25-34
3 Kazi
YU
Cash
3
Taasisi/diplom
a
3
35-44
4
Shahada +
4
45+
5
9. Kunalo jambo lingine unaweza kusema kuhusiana na mada hii jinsi watoa huduma
wanavyo suluhisha shida ya kutofanikiwa kwa huduma za kuhamisha fedha kwa njia ya
simu?
FUNGA MAHOJIANO NA UMSHUKURU MSAILIWA
168
APPENDIX 2
LIST OF MOBILE MONEY TRANSFER SERVICE PROVIDERS USING
MNO-LED MODEL
Safaricom - M-PESA
Airtel Airtel Money
Telkom Kenya Orange Money
Source: Communications Authority of Kenya (2014)
169
APPENDIX 2(A)
CBK MONTH ON MONTH MOBILE MONEY TRANSFER STATISTICS
2007 - 2016
Year 2007
Month Agents
Customer (Millions) Transaction (Millions) Value (KShs.
Billions)
Mar 307
0.020992
0.021714
0.0643905
Apr
362
0.054944
0.07
0.220896
May 447
0.107733
0.15
0.483709
Jun
527
0.175652
0.233661
0.720102
Jul
681
0.268499
0.354298
1.06537
Aug 819
0.432555
0.516239
1.57991
Sep
960
0.635761
0.669689
2.06969
Oct
1196
0.875962
0.958908
2.82955
Nov 1379
1.1332
1.22174
3.51495
Dec
1582
1.34527
1.2741
3.77027
TOTAL for 2007
5.470349
16.3188375
Source: Central Bank of Kenya: Payment System statistics for mobile payments
(2016)
170
Year: 2008
Month Agents Customer (Millions) Transaction (Millions) Value (KShs. Billions)
Jan
1812
1.5891
1.34683
4.05904
Feb
2067
1.82153
1.7399
5.21979
Mar 2329
2.07553
2.3975
6.74745
Apr
2606
2.37346
3.07289
8.38964
May 2770
2.71813
4.02127
10.9042
Jun
3011
3.03852
4.20144
10.9172
Jul
3378
3.36719
5.39108
14.0171
Aug 3761
3.72618
6.34241
16.7563
Sep
4230
4.14304
7.15191
19.2699
Oct
4781
4.42028
8.30365
21.6007
Nov 5399
4.75139
8.56681
21.7
Dec
6104
5.08247
10.2051
26.99
TOTAL for 2008
62.74079
166.57132
Source: Central Bank of Kenya: payment System statistics for mobile payments
(2016)
171
Year: 2009
Month Agents
Customer (Millions) Transaction (Millions) Value (KShs.
Billions)
Jan
7304
5.47828
10.1906
27.0749
Feb
7512
5.81602
11.0793
28.6863
Mar 13358
6.28952
13.5541
33.8202
Apr
14790
6.53192
13.7796
34.0201
May 16029
6.8427
15.0488
36.8062
Jun
16641
7.19062
15.9846
38.1756
Jul
18504
7.42641
16.8986
40.3374
Aug 18780
7.7141
17.0104
40.6787
Sep
19803
8.01624
18.3703
45.3683
Oct
20631
8.36803
19.92
48.6365
Nov 22476
8.61529
19.975
47.4656
Dec
23012
8.88258
21.6891
52.3417
TOTAL for 2009
193.5004
473.4115
Source: Central Bank of Kenya: Payment System statistics for mobile payments
(2016)
172
Year: 2010
Month Agents
Customer (Millions) Transaction (Millions) Value (KShs.
Billions)
Jan
24850
9.4767
20.0767
48.4625
Feb
25394
9.67495
20.8087
49.9055
Mar 27622
9.97211
24.0758
56.1167
Apr
29570
10.2026
22.6933
51.8136
May 31036
10.4928
24.6984
58.0795
Jun
31902
10.9147
25.0338
58.0993
Jul
32974
13.4701
26.915
61.7728
Aug 33864
14.5893
26.8233
61.531
Sep
35373
15.2239
29.4457
68.5062
Oct
37009
15.7346
31.3186
71.7947
Nov 38201
16.075
30.0386
70.2727
Dec
39449
16.4463
29.1183
75.8654
TOTAL for 2010
311.0462
732.2199
Source: Central Bank of Kenya: Payment System statistics for mobile payments
(2016)
173
Year: 2011
Month Agents
Customer (Millions) Transaction (Millions) Value (KShs.
Billions)
Jan
33968
16.6901
28.2047
75.4328
Feb
34572
16.8928
28.5462
76.3366
Mar 36198
17.4653
32.7301
88.9966
Apr
37309
17.7573
32.4254
86.0877
May 38485
17.9239
35.3457
94.3724
Jun
42840
18.1469
35.8222
92.6437
Jul
43577
18.3082
37.9763
99.7104
Aug 44762
18.6128
39.2993
107.424
Sep
46234
18.8916
39.2139
108.615
Oct
47874
19.2097
40.55
109.119
Nov 49091
19.46
41.1769
112.332
Dec
50471
19.191
41.7075
118.08
TOTAL for 2011
432.9982
1169.1502
Source: Central Bank of Kenya: Payment System statistics for mobile payments
(2016)
174
Year: 2012
Month Agents
Customer (Millions) Transaction (Millions) Value (KShs.
Billions)
Jan
52315
18.834
40.2449
114.06
Feb
53685
18.7921
41.7805
116.691
Mar 55726
19.2393
45.757
126.093
Apr
56717
19.53
44.35
117.36
May 59057
19.6943
47.9655
128.403
Jun
61313
19.7956
47.8763
124.02
Jul
63165
19.58
49.35
129.28
Aug 64439
19.38
49.7
131.38
Sep
67301
19.71
48.94
130.69
Oct
70972
20.02
51.89
137.68
Nov 75226
20.25
53.56
138.99
Dec
76912
21.06
55.96
150.16
TOTAL for 2012
577.3742
1544.807
Source: Central Bank of Kenya: Payment System statistics for mobile payments
(2016)
175
Year: 2013
Month Agents
Customer (Millions) Transaction (Millions) Value (KShs.
Billions)
Jan
85548
21.4181
53.4068
142.653
Feb
88393
21.8024
53.4683
141.126
Mar 93211
22.3292
52.3949
134.446
Apr
96319
23.0185
55.9993
142.609
May 100584
23.47
60.34
158.77
Jun
103165
23.75
60.03
152.5
Jul
105669
24.27
62.71
162.76
Aug 108559
23.87
64.71
168.1
Sep
110432
23.97
63.43
165.59
Oct
111697
24.43
68.27
175.29
Nov 112947
24.9
68.7
175.22
Dec
113130
25.3263
69.1378
182.495
TOTAL for 2013
732.5971
1901.559
Source: Central Bank of Kenya: Payment System statistics for mobile payments
(2016)
176
Year: 2014
Month Agents
Customer (Millions) Transaction (Millions) Value (KShs.
Billions)
Jan
114107
25.7568
67.0519
178.478
Feb
115015
26.1164
65.5934
172.797
Mar 116196
26.208
73.9817
192.695
Apr
116581
26.1399
72.0955
186.664
May 117807
25.8152
74.5472
198.131
Jun
120781
25.9284
74.0288
189.911
Jul
122462
26.2265
77.4651
200.992
Aug 124708
26.333
78.8987
206.72
Sep
124179
26.2995
78.1748
206.341
Oct
128706
25.996
82.8925
210.277
Nov 121419
24.9465
80.9984
203.239
Dec
123703
25.2492
85.6071
225.549
TOTAL for 2014
911.3351
2371.794
Source: Central Bank of Kenya: Payment System statistics for mobile payments
(2016)
177
Year: 2015
Month Agents
Customer (Millions) Transaction (Millions) Value (KShs.
Billions)
Jan
125826
25.3972
81.6534
210.54
Feb
127187
25.4556
80.7405
208.132
Mar 128591
25.6902
90.3477
231.836
Apr
129218
26.1392
84.9056
213.746
May 129735
26.4645
89.9024
230.152
Jun
131761
26.5028
90.6686
227.921
Jul
133989
26.7382
93.9985
238.864
Aug 136042
27.0497
94.12
248.154
Sep
138131
27.312
96.32
247.506
Oct
140612
28.5354
102.75
255.808
Nov 142386
30.0615
101.33
236.372
Dec
143946
31.6424
107.44
267.068
TOTAL for 2015
1114.1767
2816.099
Source: Central Bank of Kenya: payment System statistics for mobile payments
(2016)
178
Year: 2016
Month Agents Customer (Millions) Transaction (Millions) Value (KShs. Billions)
Jan
146710
33.09
108.13
243.37
Feb
148982
34.486
114.136
257.185
Mar 150987
36.693
121.716
273.585
TOTAL for 2016
343.982
774.14
Source: Central Bank of Kenya: payment System statistics for mobile payments
(2016)
179
APPENDIX 3
SAMPLE SIZE TABLE
Source: Krejcie, Robert V. and Morgan, D. W. "Determining Sample Size for Research Activities."
Educational and Psychological Measurement 30 (1970): 607-610
.
180
APPENDIX 4
Principal Component Analysis of Perceived Justice Scale (Rotated Component
Matrix)
Scale
Component
1
2
3
It was easy to know where to lodge your complaint
.242 .462 -.046
You waited for a short time before connecting with the customer care
staff
.010 .348 -.056
The company employee responded quickly to your problem
.140 .613
.067
The queuing system is fair
.254 .642 -.117
The company practices for addressing customer problems are reasonable
.193 .664
.212
You were treated with respect
.756 .004
.432
The employee listened attentively to your complaint
.726 .001
.471
The employee apologized for the occurrence
.289 .244
.492
The employee provided a useful information about the problem
.846 .052
.122
The employee showed concern about the problem
.833 .109
.241
The way the problem was solved resulted in a positive outcome for you
-.125 .149
.879
The solution provided was worth the effort you made to seek corrections
.146 .128
.859
The company offered you what you deserved
-.205 .175
.834
The compensation offered was fair
-.109 .147
.710
The transaction was carried out in a fair manner
-.085 .182
.811
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 4 iterations.
181
APPENDIX 5
Distribution of respondents by County
County
Frequency Percentage
(%)
County
Frequency Percentage
(%)
Baringo county
9
1.1
Marsabit county
4
.5
Bomet county
10
1.2
Meru county
31
3.9
Bungoma county
28
3.5
Migori county
15
1.9
Busia county
7
.9
Mombasa county
62
7.7
Elgeyo-Marakwet
13
1.6
Murang'a county
7
.9
Embu county
5
.6
Nairobi county
118
14.7
Garissa county
9
1.1
Nakuru county
45
5.6
Homabay county
10
1.2
Nandi county
13
1.6
Isiolo county
12
1.5
Narok county
18
2.2
Kajiado county
19
2.4
Nyamira county
10
1.2
Kakamega county
14
1.7
Nyandarua county
7
.9
Kericho county
14
1.7
Nyeri county
23
2.9
Kiambu county
40
5.0
Siaya county
14
1.7
Kilifi county
19
2.4
Taita-Taveta
23
2.9
Kirinyaga county
18
2.2
Tana River county
8
1.0
Kisii county
9
1.1
Tharaka-Nithi
4
.5
kisumu county
39
4.9
Trans Nzoia
11
1.4
Kitui county
9
1.1
Turkana county
5
.6
Kwale county
8
1.0
Uasin Gishu county
30
3.7
Laikipia county
3
.4
Vihiga county
8
1.0
Lamu county
6
.7
Wajir county
5
.6
Machakos county
16
2.0
Westpokot county
6
.7
Makueni county
19
2.4
SUB TOTAL
336
41.7
467
58.1
Total
803
100.0
182
APPENDIX 6
183
APPENDIX 7
184
APPENDIX 8
Regression Results for the Mediating Effect of Recovery Disconfirmation (Steps
1-3)
Step 1: Effect of Perceived Justice on Recovery Satisfaction
(a): Model Summary
Model
R
R Square
Adjusted R Square
Std. Error
1
.633
a
.401
.400
.77466
(b): The ANOVA for the Effect of Perceived justice on Recovery satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
321.924
1
321.924 536.457 .000
b
Residual
480.674
801
.600
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Std. Error
Beta
1
(Constant)
.749
.140
5.358
.000
Perceived Justice .846
.037
.633 23.162
.000
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Perceived Justice
185
Step 2: Effects of Perceived Justice on Recovery Disconfirmation
(a): Model Summary
Model
R
R Square
Adjusted R Square
Std. Error of the Estimate
1
.737
a
.543
.543
.85129
(b): The ANOVA for the effect of Perceived justice on Recovery Disconfirmation
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
690.500
1
690.500 952.805
.000
b
Residual
580.487
801
.725
Total
1270.987
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Std. Error
Beta
1
(Constant)
-.761
.154
-4.954
.000
Perceived Justice
1.239
.040
.737
30.868
.000
a. Dependent Variable: Recovery Disconfirmation
b. Predictors: (Constant), Perceived Justice
186
Step 3: Effect of Recovery Disconfirmation on Recovery Satisfaction
(a): Model Summary
Model
R
R Square
Adjusted R Square
Std. Error of the Estimate
1
.702
a
.493
.492
.71295
(b): The ANOVA for the effect of Recovery disconfirmation on Recovery Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
395.446
1
395.446 777.970
.000
b
Residual
407.152
801
.508
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Std. Error
Beta
1
(Constant)
1.754
.082
21.465
.000
Recovery
Disconfirmation
.558
.020
.702
27.892
.000
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Recovery Disconfirmation
187
APPENDIX 9
Moderating Influence of Locus of Attribution on the Relationship between
Perceived Justice and Recovery Satisfaction
(a): Model Summary
R
R
Square
Adjusted
R Square
Std. Error of
the Estimate
Change Statistics
R
Square
Change
F Change
df1
df2
Sig. F
Change
1
.635
a
.403
.402
.77364
.403
270.485
2
800
.000
2
.644
b
.415
.413
.76673
.011
15.497
1
799
.000
(b): The ANOVA for the effect of Perceived Justice and Locus of Attribution on Recovery
Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
323.781
2
161.891
270.485
.000
b
Residual
478.816
800
.599
Total
802.598
802
2
Regression
332.891
3
110.964
188.756
.000
c
Residual
469.706
799
.588
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig. Collinearity Statistics
Std. Error
Beta
Tolerance
VIF
1
(Constant)
.872
.156
5.585 .000
Perceived Justice
.850
.037
.636 23.255 .000
.997
1.003
Locus of
Attribution
-.056
.032
-.048 -1.762 .078
.997
1.003
2
(Constant)
1.026
.160
6.426 .000
Perceived Justice
.830
.037
.621 22.685 .000
.977
1.023
Locus of
Attribution
-.091
.033
-.078 -2.763 .006
.926
1.080
Interaction
Term (cPJ*cLA)
.165
.042
.112
3.937 .000
.908
1.101
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Locus of Attribution, Perceived Justice
c. Predictors: (Constant), Locus of Attribution, Perceived Justice, Interaction Term
(cPJ*cLA)
188
APPENDIX 10
The moderating influence of stability attribution on the relationship between
perceived justice and recovery satisfaction
Table 4.24 (a)
: Model Summary
(a): Model Summary
Model
R
R Square Adjusted R
Square
Std. Error of
the Estimate
Change Statistics
R Square
Change
F
Change
df1
df2
Sig. F
Change
1
.633
a
.401
.400
.77507
.401
268.020
2
800
.000
2
.640
b
.409
.407
.77023
.008
11.083
1
799
.001
(b): The ANOVA for the effect of Perceived Justice and Stability Attribution on
Recovery Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
322.015
2
161.007
268.020
.000
b
Residual
480.583
800
.601
Total
802.598
802
2
Regression
328.590
3
109.530
184.626
.000
c
Residual
474.008
799
.593
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Collinearity
Statistics
Std. Error
Beta
Tolerance
VIF
1
(Constant)
.763
.145
5.272
.000
Perceived Justice
.847
.037
.634
23.145
.000
.998
1.002
Stability
Attribution
-.007
.018
-.011
-.389
.697
.998
1.002
2
(Constant)
.830
.145
5.712
.000
Perceived Justice
.834
.037
.624
22.807
.000
.987
1.013
Stability
Attribution
-.016
.019
-.024
-.888
.375
.976
1.025
Interaction Term
(cPJ*cSA)
.082
.025
.092
3.329
.001
.965
1.036
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Stability Attribution, Perceived Justice
c. Predictors: (Constant), Stability Attribution, Perceived Justice, Interaction Term (cPJ*cSA)
189
APPENDIX 11
Regression Results for assessing the moderating influence of controllability
attribution on the relationship between perceived justice and recovery satisfaction
(a): Model Summary
Model
R
R Square Adjusted
R Square
Std. Error of
the Estimate
Change Statistics
R
Square
Change
F Change df1
df2
Sig. F
Change
1
.639
a
.408
.407
.77049
.408
275.982
2
800
.000
2
.646
b
.418
.415
.76490
.009
12.723
1
799
.000
(b): The ANOVA for the effect of Perceived Justice and Controllability Attribution on Recovery
Satisfaction
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
327.675
2
163.838 275.982
.000
b
Residual
474.922
800
.594
Total
802.598
802
2
Regression
335.119
3
111.706 190.925
.000
c
Residual
467.478
799
.585
Total
802.598
802
(c): Regression Coefficients
Model
Unstandardized
Coefficients
Standardized
Coefficients
t
Sig.
Collinearity
Statistics
Std. Error
Beta
Tolerance
VIF
1
(Constant)
.945
.153
6.192
.000
Perceived Justice
.841
.036
.629
23.099
.000
.997
1.003
Controllability
Attribution
-.053
.017
-.085
-3.113
.002
.997
1.003
2
(Constant)
1.078
.156
6.908
.000
Perceived Justice
.815
.037
.610
22.116
.000
.959
1.043
Controllability
Attribution
-.062
.017
-.100
-3.647
.000
.974
1.027
Interaction Term
(cPJ*cCA)
.087
.024
.099
3.567
.000
.942
1.061
a. Dependent Variable: Recovery Satisfaction
b. Predictors: (Constant), Controllability Attribution, Perceived Justice
c. Predictors: (Constant), Controllability Attribution, Perceived Justice, Interaction Term (cPJ*cCA)
190
GLOSSARY OF TERMS
Attribution - the perceived cause of an occurrence
Disconfirmation - the discrepancy between expectation and performance
Perceived Justice - the perception of fairness
Recovery Disconfirmation - the discrepancy between expectations and performance
in the context of service recovery
Recovery Satisfaction level of customer satisfaction with the corrective action taken
after a service failure
Service Failure Attribution - the perceived cause of a service failure
Service Recovery the corrective action taken after a service failure
New literature
The Effect of Recovery Locus Attributions and Service Failure Severity on
Word-of-Mouth and Repurchase Behaviors in the Hospitality Industry
Scott R. Swanson
University of WisconsinEau Claire
Maxwell K. Hsu
University of WisconsinWhitewater
Abstract
Based on a survey of 377 American hospitality customers, this study examines
the effect of recovery locus attributions and service failure severity on customer
word-of-mouth and repurchase behaviors. Findings indicate that for satisfactory
recoveries attributed to a hospitality firm, relative to employee or customer
attributions, the customer is more likely to discuss the encounter, share
information with a wider social network, and both convince others to use the
service provider and to have repatronized the firm. The results also suggest that
191
the more severe the initial failure, the greater the likelihood that a critical
incident had been discussed with a wider social network and the greater the
likelihood of warning and convincing others to not patronize the hospitality
organization. For unsuccessful (i.e., dissatisfactory) hospitality-based recovery
attempts, the recovery locus attribution was not significantly associated with the
word-of-mouth and repurchase behaviors investigated in this study
The normal random numbers were stored in the variable Y1, the double
exponential random numbers were stored in the variable Y2, the t random
numbers were stored in the variable Y3, and the lognormal random numbers
were stored in the variable Y4.
H0: the data are normally distributed
Ha: the data are not normally distributed
Y1 test statistic: D = 0.0241492
Y2 test statistic: D = 0.0514086
Y3 test statistic: D = 0.0611935
Y4 test statistic: D = 0.5354889
Significance level: = 0.05
Critical value: 0.04301
Critical region: Reject H0 if D 0.04301
As expected, the null hypothesis is not rejected for the normally distributed data,
but is rejected for the remaining three data sets that are not normally
distributed.
Frequently asked questions about the Mobile Money Transfer Services Language Preview
What is this document?
This document is a language preview including the title, table of contents, objectives and key themes, chapter summaries, and key words of a doctoral thesis.
What is the title of the doctoral thesis?
The title of the doctoral thesis is: "Perceived Justice, Service Failure Attribution, Disconfirmation and Recovery Satisfaction among Subscribers of Mobile Money Transfer Services in Kenya"
What does the table of contents include?
The table of contents lists the following: Acknowledgment, List of Tables, List of Figures, Abbreviations and Acronyms, Abstract, Chapters (Introduction, Literature Review, Research Methodology, Data Analysis, Interpretation and Discussion, Summary, Conclusion, Recommendations and Implications), References, and Appendices.
What is included in the Appendices?
The appendices include: Mobile Money Transfer Services Questionnaire, Hojaaji ya Huduma za Uhamishaji wa Fedha kwa Njia ya Simu, List of Mobile Money Transfer Service Providers Using MNO-LED Model, CBK Month on Month Mobile Money Transfer Statistics 2007 - 2016, Sample Size Table, Principal Component Analysis of Perceived Justice Scale (Rotated Component Matrix), Distribution of respondents by County, and Regression Results for the Mediating Effect of Recovery Disconfirmation (Steps 1-3), the Moderating Influence of Locus of Attribution on the Relationship between Perceived Justice and Recovery Satisfaction, APPENDIX 10 ... 188 The moderating influence of stability attribution on the relationship between perceived justice and recovery satisfaction ... 188, APPENDIX 11 ... 189 Regression Results for assessing the moderating influence of controllability attribution on the relationship between perceived justice and recovery satisfaction ... 189
What are the study's objectives?
While not explicitly stated in the provided snippets, the study objectives, based on the table of contents and abstract, likely include investigating the relationships between perceived justice, service failure attribution, recovery disconfirmation, and recovery satisfaction in the context of mobile money transfer services in Kenya.
What theories underpin this research?
The theoretical foundation includes Equity Theory, Attribution Theory, Expectancy Disconfirmation Theory, and Cognitive Dissonance Theory.
What are the key words associated with this research?
Key words: Perceived Justice; Service Failure attribution; Recovery Disconfirmation; Recovery satisfaction; Mobile money transfer
Excerpt out of 204 pages
- scroll top
- Quote paper
- Catherine Ngahu (Author), 2016, Perceived Justice, Service Failure Attribution, Disconfirmation and Recovery Satisfaction among Subscribers of Mobile Money Transfer Services in Kenya, Munich, GRIN Verlag, https://www.grin.com/document/375025
Look inside the ebook
