MscBA Marketing Management
Theo IJdema
August 2008
MASTER THESIS
MscBA Marketing Management
How to perform service recovery that leads to reduced negative consumer perceptions of the company after a controllable service failure?
RijksUniversiteit Groningen Arriva Trein
Name: Theo IJdema Student number: 1677829 Department:Marketing Qualification: master thesis E-mail:theoydema@hotmail.com Phone:06-42010629
Supervisor:Liane Voerman
Unprompted actions H2 Pre-recovery perceptions * Severity * Satisfaction Customer situation Service recovery efforts H3 H4ab
Figure 1 - Conceptual model master thesis Service failure type H5 Post-recovery perceptions * Severity * Satisfaction H1
Management summary
Where services are performed, failures can occur. This thesis studied service failures according to the following problem statement:
How to perform service recovery that leads to reduced negative consumer perceptions of the company after a controllable service failure?
To gain more insights, this problem statement was analyzed using literature to answer the research questions. Answers to these questions were:
§ Service failures can occur in different forms, the most important distinction is that of
core related service failures and non-core related service failures (Bitner et al. 1990).
§ Responses of customers to these failures are then seen in the light of a customer’s
personal situation, which can be restrained by time or non-restrained by time (Liao, 2007; Casado Diaz and Más Ruiz, 2002).
§ Recovery strategies towards service failures can be reactive after the service failure or proactive to mitigate the damage of possible upcoming failures (Hess Jr. et al. 2003). Proactive strategies are aimed at building a relationship with customers and reactive strategies are performed after a failure and can be a form of compensation (Wirtz and Mattila, 2004).
§ Theories to address customer perceptions after a service failure are the cognitive
dissonance theory (Festinger, 1957) and the disconfirmation of expectations paradigm (Oliver, 1980).
Figure 1 displays the conceptual model of the research that was conducted after analysis of the above mentioned literature.
H1 If a core service failure is faced, respondents perceive this as more severe than when a
non-core service failure is faced.
H2 The situation in which a customer finds him/herself in, influences perceived severity.
If a respondent is in a time restrained situation the perceived severity is higher than when respondents are in a non-time restrained situation.
H3 Unprompted or proactive employee actions did not have a effect upon customers
perceptions after a service failure.
H4a For the high recovery effort group, perceived post-recovery severity is of a lower level
and the change in severity before and after recovery is larger than for the low recovery effort group.
H4b For the high recovery effort group, post-recovery customer satisfaction is of a higher
level and the change in satisfaction before and after recovery is larger than for the low recovery effort group.
H5 Unprompted or proactive employee actions do not have a reinforcing effect with
service recovery effort, no relationship upon post-recovery perceptions was found. Furthermore, a relation between age and post satisfaction was found; the higher the age of respondents the higher the post-recovery satisfaction score is. Moreover, for less frequent travelers that faced a high recovery effort, post-recovery retention levels are higher than for frequent travelers. Together with the researched hypotheses, the conclusions lead to answers to the problem statement that will be given in recommendations:
- A high level of service recovery leads to higher levels of customer perceptions and
lower levels of perceived severity. Due to the fact that core service failures are found more severe this indicates that, after a controllable core service failure, high service recovery efforts should be implemented. Because of the costs involved this should be tested before the upcoming Reizigers Tevredenheids onderzoek 2008 after which the results can be analyzed and further decisions can be taken.
- Because growth opportunities are present during off-peak hours and less frequent
travellers show higher levels of retention after a high recovery effort, set up a specific recovery procedure during these periods. This will give the opportunity to bond less frequent travellers and make them travel more often – moreover this initiative will employ the ageing of society (could also be tested before the upcoming customers satisfaction research).
- A customer’s time situation is the dominant influencing variable upon the perceptions
after a service failure. When a customer is in a time-restrained situation an employee has to take personal interest (try to relate with customers) and offer a high level of recovery effort to reduce negative perceptions.
Preface
Writing a master thesis can be compared in a way with the passion of the researcher; playing football. Writing this thesis I found myself in situations that were quite similar as playing a football match. Let me explain. Without the right coaches a football player is lost because of the absence of a good strategy in a match just as I would be lost without the advice of my three coaches; Liane Voerman, Jenny van Doorn and Yvonne Dubben. I would like to thank them for giving me renewing views on different aspects of the game. Furthermore, without the help of its team players a football player can never win a match, just like the researcher would never finish the thesis the right way without the help of a colleague student like Hinke Komrij and colleagues at Arriva Train as André Cnossen, Dagomar Deelstra, Bart Uilenberg, Jan Zijlstra and Esther van der Mei. Moreover, very important for football players are the supporters, whose cheering leads them to reach a higher level of performance. Just like the cheering of family, friends and my lovely girlfriend did for me, resulting in reaching a higher level of performance for this thesis. I would like to thank everybody who was involved in this “match”, it was a great win! I hope to play many more games with you to eventually become an unbeatable champion.
Theo IJdema
Customers don’t expect you to be perfect; they do expect you to fix things when they go wrong
Table of content
1. Introduction ____________________________________________________ 12
1.1 Theoretical background_____________________________________________________ 12 1.1.1 Service recovery ______________________________________________________ 13 1.1.2 Consumer perceptions __________________________________________________ 13 1.1.3 Controllable and uncontrollable service failures ______________________________ 13 1.2 Problem statement _________________________________________________________ 14 1.3 Research questions ________________________________________________________ 14 1.4 Relevance and uniqueness of the thesis ________________________________________ 15 1.5 Company overview _________________________________________________________ 15 1.6 Outline __________________________________________________________________ 16
2. Theoretical Framework ___________________________________________ 17
2.1 Types of service failures ____________________________________________________ 17 2.1.1 Core and non-core service failures ________________________________________ 17 2.1.2 Severe and less-severe service failures _____________________________________ 18 2.1.3 Controllable and uncontrollable service failures ______________________________ 19 2.1.4 Integration of service failure classification types _____________________________ 20 2.2 Customer perceptions of service failures _______________________________________ 21 2.2.1 Perceptions of core and non-core service failures _____________________________ 21 2.2.2 Perceptions of severe and less-severe service failures _________________________ 21 2.2.3 Perceptions of controllable and uncontrollable failures ________________________ 22 2.2.4 Perceptions of a first or repeated failure ____________________________________ 22 2.2.5 Integration of customer perceptions of service failures ________________________ 22 2.3 Types of service recovery strategies ____________________________________________ 24 2.3.1 Proactive service recovery strategy ________________________________________ 24 2.3.2 Reactive service recovery strategy ________________________________________ 25 2.3.3 Recovery efforts for controllable failures ___________________________________ 26 2.3.4 Integration of service recovery strategies ___________________________________ 27 2.4 Conceptual Framework _____________________________________________________ 29
3. Research Design _________________________________________________ 30
4. Results _________________________________________________________ 41
4.1 Descriptive statistics ________________________________________________________ 41 4.2 Hypotheses testing _________________________________________________________ 42 4.2.1 Hypothesis 1: failure type & severity ______________________________________ 42 4.2.2 Hypothesis 2: time restraints & severity ____________________________________ 43 4.2.3 Hypothesis 3: employee actions & severity _________________________________ 43 4.2.4 Hypothesis 4a: recovery effort & severity __________________________________ 44 4.2.5 Hypothesis 4b: recovery effort & satisfaction ________________________________ 45 4.2.6 Hypothesis 5: employee actions & recovery effort upon perceptions ______________ 45 4.2.7 Revised conceptual framework ___________________________________________ 46 4.3 Additional analyses ________________________________________________________ 46 4.3.1 Justice and recovery effort ______________________________________________ 46 4.3.2 Retention probability __________________________________________________ 47 4.3.3 Scenarios ____________________________________________________________ 48
5. Conclusions and Recommendations _________________________________ 49
5.1 Conclusions ______________________________________________________________ 49 5.1.1 Service failures and time restraints ________________________________________ 49 5.1.2 Employee actions _____________________________________________________ 49 5.1.3 Recovery effort _______________________________________________________ 50 5.1.4 Impact of customer descriptions __________________________________________ 51 5.1.5 Impact of additional analyses ____________________________________________ 52 5.1.6 Retrospective view on the problem statement ________________________________ 52 5.2 Recommendations _________________________________________________________ 53 5.2.1 Recovery effort and failure type __________________________________________ 53 5.2.2 Recovery effort and time restraints ________________________________________ 53 5.2.3 Recovery effort and travel frequency ______________________________________ 54 5.2.4 Recovery effort and retention ____________________________________________ 55 5.2.5 Recommendations summary _____________________________________________ 55 5.3 Limitations and further research _____________________________________________ 56
6. References ______________________________________________________ 57 7. Appendices _____________________________________________________ 60
1. Introduction
1.1 Theoretical background
Due to the fact that the western economies have moved from product orientation towards service orientation, research is increasingly directed towards understanding consumer expectations, perceptions and evaluations of service delivery (Worsfold et al. 2007). When service delivery goes well this will lead to customer satisfaction, problems will arise when things go wrong and a service failure occurs. If a customer faces a service failure, a part of the damage done towards the customers’ perception can be made undone by service recovery. Service recovery is an umbrella term for systematic efforts by a firm to correct a problem following a service failure and to retain a customer’s goodwill. Efforts of service recovery play a crucial role in achieving (or restoring) customer satisfaction (Lovelock and Wirtz 2007). Moreover, service failures and (failed) recoveries are the leading cause of customer switching behaviour in service organizations (DeWitt and Brady 2003). Past service failure and service recovery works have focused on:
• Failures at the service encounter (Bitner et al. 1990; Hoffman et al. 2003)
• Employee efforts when facing a service failure or recovery (Hoffman et al. 1995;
Liao 2007; Bitner 1990)
• Consumer responses towards service failure and recovery (Wirtz and Mattila
2004; Keaveney 1995; Verma 2003);
• Types of service recovery strategies (DeWitt and Brady 2003; Worsfold et al.
2007)
1.1.1 Service recovery
Since service production and service consumption often occur simultaneously, the performance of service delivery depends on various factors such as the attitude of front-line employees, the capacity of the service system, and the behaviour of the customer concerned as well as that of other customers (Michel 2001). The presence of front-line employees during the service delivery implies that when a service fails these employees are at the scene of failure and can therefore act fast upon this failure. In a study by Hoffman et al. (1995) this statement is confirmed, their findings amplified the importance of service facilities and employee behaviours. Moreover, several other studies (Bitner 1990; Chung-Herrera et al. 2004; Liao 2007) underline the role of the employee as being a crucial influence factor during the service recovery process.
1.1.2 Consumer perceptions
Although service failures can have clear negative implications, research suggests that effective service recovery techniques may enable a company to maintain customer loyalty (Hoffman et al. 1995). It is even stated that a good recovery turns angry, frustrated customers into loyal ones, creating more goodwill than if things had gone smoothly in the first place (Hart et al. 1990). Andreassen (2001) nevertheless found that although excellent recovery efforts aid in restoring company intent and image, it not raised satisfaction to levels at or above pre-failure levels. If this service recovery paradox holds or not is for this research not of the largest interest, for it needs no argumentation that a good service recovery will always lead to higher consumer satisfaction levels up against no recovery or a disappointing one. 1.1.3 Controllable and uncontrollable service failures
A distinction can be made between failure types based on controllability:
• Controllable service failures
These are failures that are within the controllability of a company
• Uncontrollable service failures
These are failures that cannot be controlled within a company.
Jr. et al. 2003). Customers that attribute failures to controllable factors are less forgiving in satisfaction evaluations (Magnini et al. 2007). Supplementing results are that if the cause of a delay is perceived to be under the control of the firm, the customer’s anger escalates, the perceived wait length increases and satisfaction declines (Taylor 1994). These findings may very well lead to different types of recovery strategies when facing one failure or the other. This is affirmed by studies by Hess Jr. et al. (2003) and Choi and Matilla (2008). Results from both studies show that customers assume lower service provider inputs for controllable compared with uncontrollable failures, and therefore react more negative to this type of failure and expect higher recovery efforts.
1.2 Problem statement
As said, many studies investigated service failure and service recovery issues. Examining these studies led to deeper insights in both themes leading to the following problem statement:
How to perform service recovery that leads to reduced negative consumer perceptions of the company, after a controllable service failure?
This problem statement is relevant for both the company and the service recovery literature and will lead to more insights in service recovery scripts for controllable failures in service companies.
1.3 Research questions
To enhance further insights into the problem statement, the following research questions were formulated:
1. What types of service failures can occur?
2. How do consumers perceive and respond to service failures?
3. What are the possible service recovery strategies in response to these failures?
4. Which types of recovery efforts are best suited for a controllable service failure? These research questions will form the outline for the search for relevant literature
1.4 Relevance and uniqueness of the thesis
After a service failure process the concerned company is given the unique opportunity to recover this failure and restore the customers’ faith. Looking at this process, several issues are already widely addressed and studied. This thesis nevertheless covers areas that are not yet thoroughly addressed in service recovery literature. There are studies that conclude that customers react more negative to controllable service failures. However, the best way to recover these failures is not yet thoroughly examined. Furthermore, the studies conducted cover many different types of service markets [e.g. video/DVD rental and restaurants (Worsfold et al. 2007); restaurants and hotels (Hoffman et al. 2003); banks (Michel 2001) and an international airline (Bamford and Xystouri 2005)]. The public transport market has never been examined. Moreover, a study regarding employee actions towards customers after a service failure that is controllable has not been conducted before in depth. Such a renewing research in the market of services makes the issue relevant and unique for a master thesis and will contribute to the existing literature.
1.5 Company overview
Arriva Netherlands is part of Arriva PLC which was founded in the United Kingdom. Arriva PLC is, besides the Netherlands and the United Kingdom, present in other European countries (e.g. Denmark, Italy, Portugal, Spain and others). This makes the company one of the largest commercial companies in the European public transport industry. Arriva Netherlands exists out of four business units (BU’s) of which public transport is the largest followed by the BU Touring. The BU Public Transport consists out of the areas Arriva Bus and Arriva Train. These BU’s have a strong mutual consistency and supplement each other where necessary (Annual report Arriva PLC 2007).
receive from facing a service failure. When service companies have the right scripts that anticipate service failures this can be a valuable asset to retain customers after a negative experience. For this reason it is important for Arriva Train to gain insights in these possible scripts and how to make them work to restore customer satisfaction.
1.6 Outline
2. Theoretical Framework
Within this part of the master thesis, literature will be examined to gain deeper insights into the problem statement and the research questions. The research questions will be dealt with in different paragraphs leading to different hypotheses. These hypotheses will than serve as the input for the conceptual framework, which is the guideline for the gathering of empirical data.
2.1 Types of service failures
Several studies address different categorizations of services failures (e.g. Bitner et al. 1990; Craighead et al. 2004; Choi and Mattila 2008). Three different categorization types of services failures will be emphasized in this paragraph; core and non-core – severe and non-severe – controllable and non-controllable service failures. After discussing these failure types, an integrative view will be given on the differences and the impact of the categorization types of failures regarding the master thesis.
2.1.1 Core and non-core service failures
Many distinctions can be made between different types of service failures. An important, often cited study on different types of servicescape positive and negative incidents is that of Bitner et al. (1990). These authors performed research in three different industries: hotel, restaurant and airline. Their findings resulted in the incident classification system; as shown in figure 2.1 three groups of employee behaviours that can account for all satisfactory and dissatisfactory incidents.
Is there a service delivery system failure? Group 1 Is there an implicit/explicit request for accommodation? Total sample Yes No Yes No Yes No Group 2 Group 3 Is there an unprompted unsolicited action by employee?
Customer Loyalty Perception of quality Severity of failure Service guarantee Antecedents of service recovery Service failure encounters Failure type creation and validation Type 1 Loyal/severe Type 2 Loyal/non-severe Type 3 Non-loyal/severe Types of service failure encounters
Figure 2.2 - Part of framework Craighead et al. (2004)
These three categories of employee behaviours are:
• Employee responses to service delivery system failures (group 1) • Employee responses to customer needs and requests (group 2) • Unprompted and unsolicited employee actions (group 3)
This incident classification system makes a distinction between service delivery system failures which are core service failures (group one) and non-service delivery system failures, the non-core service failures (groups two and three). Findings of Bitner et al. (1990) show that the largest proportion of dissatisfactory encounters is related to employees’ inability or unwillingness to respond to core service failures (group one). These results are supplemented by a study of Hoffman et al. (1995), who identified and classified failure incidents and recovery strategies in the restaurant industry. This study shows that service delivery system failures and the involved employee responses to these failures (group one) accounts for the largest part (44.4 percent) of the total number of failures. This indicates that failures that occur most often are failures related to the core of the provided service.
2.1.2 Severe and less-severe service failures
Service failures are also distinct regarding the severity of the service failure (Craighead et al. 2004). The severity of a service failure refers to a customer’s perceived intensity of a service problem (Weun et al. 2004). The severity of a failure
non-core related, service features, as outlined in the former paragraph. Craighead et al. (2004) measured severity in their study by merging three variables (inconvenience, time and cost) into one, using factor analysis. Figure 2.2 describes that this measured severity then was seen as one out of four antecedents of service recovery to identify the type of service failure encounter. The eventual distinction in types of failure was made in a combination between loyalty of customers and severity of the failures.
The described distinction nevertheless needs a note, for Liao (2007) concludes that customers define the severity of a failure and the justice of the recovery that they are facing themselves by their individual, subjective evaluation of the service failure. This indicates that a classification into core and non-core service failures cannot create a clear distinction between service failures because an individual customer can perceive a non-core service failure as a severe failure.
2.1.3 Controllable and uncontrollable service failures
A third distinction between failure types is the difference between a controllable and an uncontrollable failure (Choi and Matilla 2008). Controllable failures are failures that are situated within the influence of the company; it is often perceived that the company could have prevented these failures. Non-controllable failures are the failures that the firm cannot be accounted for, it could not have prevented or foresee such a failure (Magnini et al. 2007). Recent studies (e.g. Magnini et al. 2007 and
Choi & Matilla 2008) show that controllability has a significant impact on the perceived service failure and service recovery process. Choi and Matilla (2008) compared customer reactions to failures and satisfaction of recovery between perceptions of controllability, failures that are contributed to: the service firm, partial self-blame or to an unknown variable (figure 2.3). Outcomes of their research
Perceived Controllability over the cause for the failure:
• Service firm • Partial self-blame • Unknown Service expectation: • High • low Service failure Attribution Encounter satisfaction Overall reaction
“The severity of a failure and the loyalty of a customer are both subjective and therefore unstable variables”
Craighead et al. 2004
will be addressed in subparagraph 2.2.3 Perceptions of controllable and uncontrollable failures.
Although it is acknowledged that controllability is an influencing variable in the service recovery process, Hess Jr. et al. (2003) did not find a significant relation between controllability and service recovery outcomes. Nevertheless, the researchers explained this result as a possible study bias because respondents might have judged the failures they were confronted with as highly controllable in general.
2.1.4 Integration of service failure classification types
When studying customer reactions towards service failures, it is important to make a clear distinction between these failures. Bitner et al. (1990) categorize employee behaviours that could account for all satisfactory and dissatisfactory incidents. The different categories in this process nevertheless do not take severity or controllability fully into account. But, it could be stated that the different groups vary in severity and controllability themselves. It is very likely that employees should react in different ways when conducting recovery after a controllable/non-controllable or severe/non-severe failure.
Such a view based on severity is acknowledged by Craighead et al. (2004), who distinguish between service failures based on the
loyalty of consumers and the severity of a failure. This is interesting due to the fact that it could be important to know how customers react towards
failures that differ in severity. One note should be made in this categorization because the severity of a failure and the loyalty of a customer are both subjective and therefore unstable variables.
2.2 Customer perceptions of service failures
Now different types of service failures are uncovered, it is important to analyze how customers perceive these failures. The perceptions and responses of customers towards the addressed types of service failures will be discussed in the upcoming subparagraphs. Furthermore, a subparagraph will take failures into account that are within the categorization and will analyse them in the light of repetition.
2.2.1 Perceptions of core and non-core service failures
Customer reactions differ per service failure type that is faced. The study of Bitner et al. (1990) makes a distinction between employee actions towards core service failures and non-core service failures. The authors find that customers are more likely to have negative perceptions and are more dissatisfied if a failure was more related to the core service and its recovery. In their research the authors find that it is not only the initial core service failure that causes dissatisfaction, but also the employee response to this failure. In all of the reported core service failure incidents in group one, the employee failed to handle the situation in a way that could have satisfied the customer. Moreover, the first and the foremost cause of customer outrage was core service failure (Verma 2003). This possibly explains why Keaveney (1995) found that severe core service failures are the largest reason for customers to switch service.
2.2.2 Perceptions of severe and less-severe service failures
“When customers perceive that the service firm could have prevented the incident, their
reactions were highly
negative”
Choi and Mattila 2008
importance, the perceived anger increases (this is also acknowledged by Craigheid et al. (2004) who used time as a variable of severity).
2.2.3 Perceptions of controllable and uncontrollable failures
Strong differences in consumer perceptions are found between controllable and uncontrollable failures. If a service failure occurs, a
customer can attribute the blame in the following ways: the firm can be blamed, customers do not know who to blame or customers become aware that they
are partly to blame themselves (Laufer et al. 2005). Controllability has a large impact on responses, for controllability attributions reflect the customer’s belief that the service firm could have prevented the failure (Weiner, 2000 and Hess Jr. et al. 2003). The inability of a firm to prevent these controllable failures is perceived as poor company performance and as a lack of ability to prevent mistakes (Poon et al., 2004). Recent study affirms that when customers perceive that the service firm could have prevented the incident, their perceptions are highly negative. Focal encounter satisfaction, overall satisfaction, return intentions and positive word-of-mouth will be significantly lower among these customers than among their counterparts in the self-blame or control unknown condition (Choi and Mattila 2008).
2.2.4 Perceptions of a first or repeated failure
Within the categorization of service failures mentioned in the previous paragraphs another factor may affect consumer reactions towards service failures, being the previous experiences with a specific company. Maxham and Netemeyer’s (2002) findings are that when customers experience a second failure with the same firm customer satisfaction and repurchase intentions are lower than after the first failure, even though, after both failures, the recovery efforts are evaluated as acceptable. Liao (2007) goes further; his findings show that even after a much higher service recovery performance a company may not be able to recover customer satisfaction from repeated failures.
2.2.5 Integration of customer perceptions of service failures
Feature: Type of failure/customer: Consumer responses:
Core/non-core
Core Very negative; likely to respond negative and more dissatisfied
Non-core Also negative, depends on the customers situation and the severity of the failure
Severe/non-severe
Severe failure Depends on a customers’ situation has a main negative effect on customer trust, customer commitment, and likelihood of engaging in negative word-of-mouth Non-severe failure Depends on a customers’ situation - satisfaction after
recovery is created more easily
Controllable/non-controllable
Controllable failure Negative reaction; company performance is seen as poor and overall satisfaction, return intent and positive word-of-mouth are significantly lower
Non-controllable failure Failure uncontrollability or self-blame reduced the negativity effect of a service failure
Repetition
First failure Failure is seen as negative - it is likely that recovery can change this perception
Repeated failure More negative; lower satisfaction and repurchase intention, recovery can hardly change this
Table 2.1 – summarizing table of failure types and consumer responses
The conclusion after the first analysis is that the research will address controllability and failure severity. This choice was made based on the fact that controllable failures are seen as more severe and core/non-core service failures and repetition can both be studied in light of severity.1 Bitner (1990) shows that incidents that affect service delivery system failures, or core service failures, account for the largest part of the total number of negative incidents. Moreover, Keaveney (1995) found that severe core service failures are the largest reason for customers to switch service providers. These findings lead to the first hypothesis, which assumes that:
Liao (2007) indicates that a customer’s situation affects the customer's perceived severity of a service failure. Craighead et al. (2004) measured this severity according to three variables; inconvenience, time and cost. Time restraints are also acknowledged by Casado Diaz and Más Ruiz (2002), who show that punctuality
1
Higher controllability is found to be more severe in literature, differences between controllable and non-controllable failures will therefore not be addressed in this thesis it will only focus on controllable failures.
H1 In the event of a service failure, it is likely that the more “core service related”
“The basic idea of a proactive approach is that reactive actions are often initiated too late to recover all of the perceived damage”
Gwinner et al. 1998
importance of a customer influences the perceived anger of the service failure. These findings lead to the second hypothesis, which assumes that: 2
2.3 Types of service recovery strategies
Several authors studied the influences of proactive and reactive service recovery upon customer severity and satisfaction (e.g. DeWitt and Brady 2003; Boshoff 1996 and Hoffman et al. 1995). Roughly recovery strategies aimed at mitigating the effects of service failures can be divided into two classes: proactive and reactive (Worsfold et al. 2007). The first subparagraph of this chapter will address proactive service recovery; the second subparagraph will address reactive service recovery. The third subparagraph outlines the recovery possibilities specifically for controllable failures and the final subparagraph than gives an integrative view upon the types of recovery and will address the associated hypothesis.
2.3.1 Proactive service recovery strategy
It might seem strange to recover a failure before it has occurred however, it follows the rule of the phrase: it is better to prevent than to cure. A study of Hess Jr. et al. (2003) shows that loyal customers react less negative to service failures than non-loyal customers. This is an important conclusion
because in the case of proactive service recovery, rapport plays a significant role. The type of rapport is in this case: a customer’s perception of having an
enjoyable interaction with a service provider employee, characterized by a personal connection between the two interactants (DeWitt and Brady 2003).
The basic idea of a proactive approach is that reactive actions are often initiated too late to recover all of the perceived damage caused by the failure to the consumer. Therefore the suggestion is that if customers receive relational benefits, they may
2
A customers’ situation can be described in several ways. This thesis will make use of a situation that is either restrained by time or is not restrained by time. These limitations will also include inconvenience (a time-restrained situation is more inconvenient than a non time restrained situation). Costs will not be involved because public transport is known to be the least expensive way of transport up against its substitutes; going by car or cab.
H2 In the event of a service failure, the more time-restraint a customer situation
“Explanations given by personnel are likely to reduce the customer's perception of injustice and moral outrage, leading to a reduction in dissatisfaction”
Dunning et al. 2004
remain in a relationship even if they perceive the core service attributes to be less than superior (Gwinner et al. 1998). In four studies De Witt and Brady (2003) confirms this suggestion. Their findings show that rapport has an inoculation effect on service failures wherein goodwill that is developed from a positive relationship with an employee is later applied to the interpretation of a failed service encounter. They nevertheless also found a negative implication. The foundation of a relationship causes that customers are less likely to complain because they do not want to damage their relationship.
The positive implications of a customer service provider relationship are supplemented by Worsfold et al. (2007). Their research shows that good processes occurring prior to service failures can buffer the damaging effects of poor outcomes. Moreover, they suggest that in highly competitive service industries, it may often be too late to positively influence customer evaluations after a service failure has occurred. Their finding of main effects associated with rapport suggests that this proactive approach may effectively control damage associated with subsequent service failure.
2.3.2 Reactive service recovery strategy
The employee ability to deal effectively with customer problems closely relates to employee satisfaction and loyalty (Bamford and
Xystouri 2005). This employee ability should be dynamic, for there are different options to react upon a service failure. Detailed research into this issue is conducted by Wirtz and Mattila (2004).
“People are more forgiving if they feel that the failure was not reasonably foreseeable to the service provider”
Magnini et al. 2007
process. Another important conclusion is that the speed of recovery significantly relates to the efficiency of the service provider; a delayed response implies inefficiency and thus induced consumers to think that the same problem is likely to occur in the future.
In a study conducted by Liao (2007) these findings on the role of service recovery performance after service failures are supplemented. Results of this study are that frontline service employee’s behaviours of prompt handling of customer complaint, effective problem solving, making an apology and being courteous play a pivotal role in shaping customers’ justice perceptions of service recovery. These findings are supplemented by other studies (e.g. Verma 2003 and Dunning et al. 2004) which show that when providers take personal interest, relate with customers emotionally and make an earnest attempt to recover a problem, they feel delighted. Furthermore, findings from this study show that adequate explanations given by sales personnel for customer grievances are likely to reduce the customer's perception of injustice and moral outrage, leading to a reduction in dissatisfaction. 2.3.3 Recovery efforts for controllable failures
The above mentioned strategies are applicable for controllable and uncontrollable failures. Nevertheless, for controllable failures there are differences in the way they should be recovered. Research of Magnini et al. (2007) shows, that people are more forgiving if they feel that a failure was not reasonably
foreseeable by the service provider. This indicates that uncontrollable failures are perceived as less severe than controllable failures and therefore,
“The ability of service employees to identify the severity of service failures indicates a greater need for empowering front line employees”
Weun et al. 2004 • Information providing - it is important that when a failure is outside a firm’s
control, it is crucial to let customers know what circumstances led to the failure (reactive).
• Failure prevention - let the customers know that although the failure was under
control, the firm took action to prevent such kinds of failures in the future (reactive).
• Enhance expectations - enhance customer expectations to reduce the potential
negative impact of inevitable service failures on business outcomes (proactive
Despite the positive influence of a strong recovery on satisfaction, Weun et al. (2004) found that the severity of the service failure has a main effect on customer trust, customer commitment, and likelihood of engaging in negative word-of-mouth after the service failure. Their findings indicate that customers may still be upset, engaging in negative word-of-mouth, and be less likely to develop trust and commitment, even with strong recovery, if the original problem is severe. These results indicate that there are negative consequences for (severe) controllable service failures, regardless of the successfulness of the service recovery.
2.3.4 Integration of service recovery strategies
When examining proactive and reactive service recovery it is important to acknowledge that studies that address both types state that the two rather enforce than weaken each other (DeWitt and Brady 2003 and Worsfold et al. 2007). Furthermore, main conclusions of these studies
acknowledged the important role of the frontline employees. Weun et al. (2004) show that the ability of service employees to identify the severity of service failures indicate a greater need for empowering front
leads to dissonance, there was a positive or neutral image of the company/service but now this is affected by negative perceptions. According to the cognitive dissonance theory (Festinger, 1957) this experienced dissonance is psychologically inconsistent and the customer will try to reduce this dissonance. In combination with the articles that emphasize the role of frontline employees, this indicates that positive employee behaviours might lead to reduced negative perceptions after a service failure. These findings lead to the third hypothesis:
In a similar vein, the cognitive dissonance theory of Festinger (1957) can also be used to analyze the effect of service recovery on the reduction of dissonance. Hess Jr. et al. (2003) show, that when customers face a failure that is perceived as controllable, they expect greater recovery efforts by the service provider in order to restore equity to the exchange. These findings lead to hypothesis 4a that assumes that:
Other findings from literature suggest that customers react more negatively when they believe the service firm is responsible for the service failure (Choi and Mattila, 2008; Weun et al. 2004)). These reactions can be explained by the disconfirmation of expectations paradigm (Oliver, 1980). This theory suggests that consumers reach satisfaction decisions by comparing product or service performance with prior expectations about how the product or service would or should perform. These expectations are compared with actual perceptions of performance as the product/service is consumed. If expectations exceed performance, dissatisfaction results but, when expectations are met, or when the performance actually exceeds expectations, satisfaction results (Bitner, 1990). Regarding this theory it could be stated that a service failure causes expectations exceeding performance. With service recovery however, the perceived performance level can exceed expectations, leading to satisfaction. The preceding reasoning leads to hypothesis 4b:
H3 Unprompted or proactive employee actions will lead to a lower level of
perceived severity before recovery.
H4b Service provider recovery efforts will have a positive effect on customer
satisfaction.
H4a Service provider recovery efforts will have a positive effect on perceived
In their research Choi and Mattila (2008) show that even after severe, controllable failures customers can still have negative perceptions after a successful recovery. This suggests that a reactive recovery alone might not be sufficient. To analyze an assumed reinforcing relationship between proactive employee actions and reactive recovery, the previously stated theory suggests hypothesis five:
2.4 Conceptual Framework
The conceptual framework is based on findings, conclusions and theories found in the theoretical part of the Thesis. The literature outcomes led to the following framework and hypotheses:
H1 In the event of a service failure, it is likely that the more “core service related” the failure is, the more it will be perceived as severe (before recovery).
H2 In the event of a service failure, the more time-restrained a customer situation is, the more the failure will be perceived as severe (before recovery).
H3 Unprompted or proactive employee actions will lead to a lower level of perceived severity before recovery.
H4a Service recovery efforts will have a positive effect on perceived severity.
H4b Service recovery efforts will have a positive effect on customer satisfaction.
H5 Unprompted or proactive employee actions will reinforce service recovery efforts leading to a more positive level of post-recovery perceptions.
H5 Unprompted or proactive employee actions will reinforce service recovery efforts leading to a more positive level of post-recovery perceptions.
Unprompted actions H2 Pre-recovery perceptions * Severity * Satisfaction Customer situation Service recovery efforts H3 H4ab
3. Research Design
To find answers to the proposed hypotheses, empirical research will be conducted. This research exists out of two parts, a qualitative part and a quantitative part. In this chapter these types of research, the research designs and the analyses involved will be addressed. For both types, the research methods and the data collection and sampling plan will be handled. The exploratory research leads to results that will be discussed and used as input for the quantitative research. Furthermore, the process of the quantitative research will be addressed.
3.1 Qualitative research
First a qualitative research will be conducted. The goal for this research is to find input for the causal research and to gain deeper insights into the research problem. 3.1.1 Research method
The type of qualitative research that will be conducted is an exploratory research. The objective of an exploratory research is to explore or search through a problem or situation to provide insights and understanding (Malhotra 2007). Outcomes of the exploratory research will give input for the quantitative part of the research and will therefore make it more thorough. Table 3.1 shows that two types of exploratory research were used:
Type Classification Objective How
Secondary data Internal source – complaints received by the customer contact centre
Gain insights in the most mentioned complaints – rank them
Order and analyze the complaint data – search in internal complaint reports
External source – ministry of V&W - customer satisfaction research public transport
Gain insights into the satisfaction levels
Analyze results of the customer satisfaction research public transport
Qualitative research In-depth (expert) interviews Gain insights in severity, perceived failures and recovery efforts
Expert interviews with customers (5) and the account manager of Arriva Train for recovery possibilities (1)
3.1.2 Data collection and sampling plan
The type of data collection for the exploratory research will be semi-structured expert interviews. The choice for expert interviews is made due to the fact there will be no bias due to social pressure and the goals of the research are related to the probing of individuals (Malthotra, 2007). The sample size for customers will be five subjects, these are used to gain insights in recovery efforts, customer situation, justice and severity. The expert interviews with the travellers will be conducted in the train. During several train rides the researcher will sit next to customers and ask for permission for an interview. Furthermore, research in complaint and satisfaction reports from 2007 and an expert interview with the customer and market manager of Arriva Train were conducted to gain insights in recovery efforts, failure types and unprompted actions.
The semi-structured interviews will be conducted to reveal the respondents’ thoughts, opinions and feelings on the issues to be described (failure type, customer situation, unprompted employee actions and recovery efforts). These issues will be mentioned by the interviewer and the respondents will be asked to respond to these issues in the light of their experiences with Arriva Train. All interviews are summarized by topic and analyzed in the light of the addressed theoretical issues. These insights than will be used as input for the causal research (Malhotra, 2007). 3.2 Outcomes of the qualitative research
The exploratory research led to several findings that are more industry specific. Secondary data and qualitative research showed that the theory applied onto the public transport market which led to important input for the quantitative research scenarios. The outcomes of this research will be discussed using the same issues as the theoretical framework, being; failure types, customer situation, unprompted employee actions and recovery effort. After gaining insights in these issues a subparagraph with conclusions which will define variables that will be included in the quantitative research. Furthermore, the scenarios to be used will be addressed. 3.2.1 Service failure types
Technical problems - These types of failure are core service failures and are related to technical problems on the trains or the rails. The research will focus on the technical problems on the trains for these problems are controllable service failures. Technical problems causing a delay or unavailable service are seen as the most annoying service failures and most of the complaints received by the company describe this issue (complaint processing report 2007). This was supplemented by the customer interviews (appendix 1) where all respondents named a delay or unavailable service as unacceptable and severe.
Inconvenient situations - Situations that do not affect the core service directly can have a negative influence on the service delivery. These situations are summarized under ‘inconvenient situations’. These are situations that do not stop or delay the travelling nevertheless can be perceived as a service failure. Although often perceived as less severe, because they do not cause core service problems, these situations are seen as inconvenient and create negative perceptions of the company. Inconvenient situations that can occur when travelling per train are: no seating, a broken air-conditioning, a lack of travelling information and a dirty environment. These situations were all named during interviews with customers, no seating was named as being extremely annoying as was a lack of information (often in combination with a delay). Less severe were a broken air-conditioning and a dirty environment.3
Other failures – within the interviews no other failures were mentioned by the respondents however, in the complaint and satisfaction report 2007 ten percent of the respondents complained about the personnel. When respondents were confronted with this result they mainly concluded that these travellers possibly complained due to the fact that they got a fine and wanted to react their grievances upon the steward by complaining.
3.2.2 Time restraints
An issue that influences the level of response and recovery effort is the severity of a failure. This is closely linked to the customer situation which was affirmed by the respondents. Severe was a failure that occurred in a situation in which a failure caused a high level of inconvenience. Respondents named having an important meeting, an exam or a job interview as such important situation. This indicates that
3
when customers have an appointment that includes time restraints, a failure is seen as severe. Supplementing results were that when asked about the same failure in a different setting, as for going into town with a friend, the respondents all rated the failure as less severe.
Although the same failures were rated as more severe in different time situations, delays and unavailable services were always seen as the most severe failures. These findings were affirmed by Yvonne Dubben, manager customer and market of Arriva Train, who acknowledged that for every customer complaint and satisfaction research that has been done since the start of Arriva Train in the Netherlands, the same three issues come up in the same order, regardless the customer situation:
1. Delay or unavailable service 2. No seating
3. Lack of information
3.2.3 Unprompted or proactive actions
Unprompted or proactive actions are employee activities that are not standard procedures. These actions can be positive and negative and can influence a customer’s perception of the service provided. Positive actions can be seen in the light of proactive service recovery; they can positively influence the outcome of a recovery. This statement of positivism after an (positive) unprompted employee action was supported by the in-depth interviews. Respondents stated that the personnel gave them a safer feeling. One respondent was very positive about the help she received with her luggage. For the quantitative research this implies that positive unprompted actions are likely to reinforce the service recovery efforts provided.
3.2.4 Service recovery efforts
service failure.4 This implies that service recovery often constitutes arranging replaceable transport within half an hour, what in reality is often the next train. This agreement however, does not apply to the customers. Travellers can miss a transfer to another train if they are five minutes late or miss an appointment. Appendix 1 shows that this statement was confirmed by the respondents which were very clear about the service recovery efforts of Arriva: ‘Are there any?’ was the most mentioned answer.
When respondents were asked what kind of efforts they would appreciate, compensation was the most favourite choice. The respondents that did not mention this answer themselves, still found this option most favourable when it was named by the interviewer. Different kinds of compensation for different types of failures were seen as appropriate, for a delay or unavailable service a total refund seemed justified. The second best option was a coupon for a consumption of some sort.5
These findings were integrated in the interview with the manager ‘customer and market’ of Arriva Train. The option for full refund was a non-option from a company perspective: customers travel with different tickets and it is found impossible to check whether a traveller with a subscription was in a delayed train or not. The coupon for a consumption of some sort is a formula that is already used by a competitor (NS) in special situations, but is not a standard procedure. This option seems as a realistic way of recovering.6 This recovery effort is more specific and can be executed by employees; it is likely that this will enforce the contact and relationship with the customers.
3.2.5 Implications for the quantitative research
Table 3.2 shows that the secondary data and the qualitative research resulted in a choice for a specification of independent variables that are necessary for the quantitative research:
4
Dutch public transport companies have two types of customers; the governments that provide grants and the travellers that buy the actual train tickets. It needs no argumentation that the travellers are the ‘real’ consumers however; agreements are made with the government regarding service failures and the level of fines when they occur. This indicates that in some cases, according to the governmental agreements, a service failure has not occurred while travellers do face a service failure by their standards.
5
See appendix 1
6
Failure type
Core service
failure
Delay Delay is seen as the most important core service failure
Non-core service failure
Broken air-conditioning
Broken air-conditioning is seen as inconvenient where no seating and no information are seen as more severe
Customer situation
Time restrained Job interview All possible respondents can identify with a job interview were students might not identify with an important meeting and working people not with an exam
Non-Time restrained
Going out with friends
Is seen as a free day – with no stress and time limits – is easy to identify with
Unprompted actions
Present Photo employee
helping
Employee helping a customer by pointing out the right platform - might reduce the severity because of the positive memory of a helping employee
Non-present Photo employee
not helping
Normal behaviour – doing the job and nothing more
Recovery efforts
High Consumption
coupon
Is a realistic possibility and is seen as a high recovery effort
Low Apology Is a normal reaction and low profile effort - should be a standard procedure
Table 3.2 independent variables
3.3 Quantitative research
The second research method is quantitative research, which will be conducted to find influencers and relationships of and between the related variables (e.g. controllable service failure, severity, customer situation, unprompted actions, satisfaction and service recovery). This paragraph also addresses the process of the research and situations that were faced conducting the research. Respondents’ characteristics (age, travel frequency and travel goals) are described and response errors are highlighted.
3.3.1 Research method
phenomenon (Malhotra 2007). Within this research category a 2x2x2x2 between subjects factorial design was chosen to test the influence of the independent variables (type of failure, limitations of the consumer’s situation, the recovery efforts and the influence of unprompted actions) towards post-recovery satisfaction and severity perceptions. From the company’s perspective it is also interesting to measure the influence of the variables upon retention probability, which was based on Wirtz and Mattila (2003).
During the research respondents will be exposed to different scenarios describing different situations regarding service failure, customer situation and recovery effort. The use of a scenario based experiment has proven to be successful in service failure research (e.g. Bitner 1990 and Smith et al. 1999) and was chosen for this study for several reasons:
• The method avoids the problems of intentionally imposing service failures on
customers, as would be required in a field experiment (Wirtz and Mattila 2003).
• The method minimizes memory-bias, which is common in self-reports of service
failures in survey designs (Smith et al. 1999).
• Scenarios enhance the variability in customer responses to service recovery, as
they can be effectively manipulated (Smith and Bolton 1998).
• The scenario method enhances internal and statistical conclusion validity by
controlling extraneous and manipulated variables, and by reducing random noise in the dependent variables (Bitner 1990).
The scenarios will be tested using a scenario-based questionnaire of which a sample of the scenario is provided below (the alternative experimental conditions are cursive and provided in parentheses)7:
You are travelling with Arriva to a destination; you have a job interview (for a day out with friends). Just when the train was about to leave it had a technical problem: you are going to be late on your job interview (there was a broken air-conditioning). The train driver tells the passenger through the microphone that the technical (air-conditioning) problem was to blame to Arriva, he apologizes humbly on behalf of the company (apologizes and every traveller gets a consumption coupon).
Figure 3.1 shows the pictures that explain the unprompted actions (based on: Bitner, 1990):
The factorial design leads to sixteen scenarios. Using a fractional factorial design, this number of sixteen scenarios can be decreased to increase the number of subjects per scenario, to make the research more viable. For a statistically accountable decrease of scenarios an orthogonal design was used to find the right scenarios for the research. The orthogonal design procedure creates a reduced set of profiles that is small enough to include in a survey but large enough to assess the relative importance of each factor (SPSS 15.0). The orthogonal design led to a decrease in scenarios to a number of twelve.8
3.3.2 Data collection and sampling plan
The data that is necessary for the quantitative part of the research is collected via a mall-intercept personal interview type of approach (Malhotra 2007). The master thesis addresses service failures at the public transport company Arriva Train so
7
The original questionnaires (appendix 3) were made in Dutch for the convenience of the respondents, this is an edited English translation of the sample scenario.
8
See appendix 4
Unprompted actions present Unprompted actions not-present
mall-intercept was actually “train-intercept personal interviews”. A sample of 420 travellers of Arriva Train served as a subject pool.9 Respondents were asked to fill in questionnaires on different times of the day (during peak hours and off-peak hours) and on different days of the week (midweek and weekend). During the travel respondents have enough time to fill in the questionnaire. Moreover, these respondents could very well project themselves into particular imaginary research situations what is important for a scenario based research (Wirtz and Mattila 2003). 3.3.3 Plan of analysis
The involved variables within the causal research are:
All of the dependent variables were measured using a seven-point Likert scale (Choi and Mattila 2008). Perceptions were measured twice, before and after the recovery had taken place, to examine the expected differences resulting from the service recovery. Most of the items were labelled with “strongly agree/strongly disagree” while others were anchored with descriptors as “very severe/very unsevere” and “very likely/very unlikely” (Blodget et al. 1997). The dependent variables will be measured by using a separation criterion: recovery effort. Before (pre-) and after (post-) recovery, the variables will be measured to put more emphasis onto the differences in perception before and after recovery. The faced situation will than be:
The differences between these types of dependent variables can be measured by: Post-perception ⁿ – pre-perception ⁿ = positive / negative change ⁿ
9
The average of ‘subjects per scenario’ was measured based on five scenario based factorial design studies (Bitner 1990; Blodget et al. 1997; Choi & Mattila 2008; Smith et al. 1999 and Wirtz & Mattila 2003), the average of subjects per scenario for these studies was approximately 23 subjects. For the causal research 420 respondents were addressed for 12 scenarios what makes the number of subjects per scenario 35 and above average.
Independent variables:
• Service failure type (core/ non-core)
• Customer situation (Time restrained/ non-Time restrained) • Service recovery efforts (low/ high)
• Unprompted actions (present/ non-present)
Dependent variables: • Satisfaction • Severity • Retention Pre-perceptions (satisfaction/ severity/ retention) Failure type Customer situation Unprompted action
These differences can explain the power of the recovery effort upon change in perceptions and the effect of the other independent variables can also be measured. Using this method there is no interference of the recovery effort upon the recovery perceptions, what can determine specific influencers upon these pre-perceptions.
The results will be analyzed with the statistical program SPSS. To discover differences between means of two independent groups, independent t-tests were used. To compare multiple independent groups, ANOVA n-way analyses of variance were used (Malhotra 2007). Furthermore, to find relations between variables, correlation analyses and interaction analysis via ANOVA n-way analyses of variance were conducted (Huizingh, 2002). It should be acknowledged that for the different analyses the level of N differed substantially. For the scenario comparison 35 respondents were used per scenario, for the comparison of independent variable groups (two types), each of the groups contained an N of 210. Before conducting the eventual causal experiment the questionnaires were extensively pre-tested. Based on this pre-testing, items were slightly reworded. 3.3.4 Respondents’ characteristics
A convenience sample of 420 train travellers served as respondents. The surveys were taken on different days of the week10 during and off peak hour. Within the sample 56 percent of the respondents were female and the majority of the respondents were between the age of 21 to 40 years (table 3.3).
Table 3.4 (see next page) shows that the majority of the respondents has a frequent or
regular travel frequency (42,9 percent and 28,6 percent respectively). The distribution of travel
10
Monday, Wednesday, Thursday and Saturday, respectively Age Frequency Percentage under the age of 20 139 33,0 between 21 to 40 years 190 45,3 between 41 to 60 years 65 15,5 over the age of 61 years 22 5,2
Missing 4 1,0
Total 420 100,0
Travel frequency Frequency percentage frequent (4< per wk) 180 42,9 regularly (1-3 per wk) 120 28,6 sometimes (1-3 per mth) 81 19,3 rarely (1 or more per yr) 39 9,3
Total 420 100,0
Table 3.4 – Travel frequency distribution
goals (table 3.5) shows two goals that can be accounted for almost 75 percent of all travel goals mentioned being; work (31,2 percent and education 43,6 percent).11 For only two percent of the respondents something else was the travel goal, a hospital visit is an example of such a different travel goal.12
3.3.5 Response errors
Only few response errors occurred. Some respondents filled in multiple travel goals (although the question states clearly: “for which travel goal do you use the train the most?”). Furthermore, four respondents did not alter an age. Two respondents, that possibly had to leave the train because they arrived at their destination, did not fill in post-recovery severity and post-recovery retention. These cases were excluded when measuring these variables.
11
Other research results indicate that shopping is the third largest motive. This could be explained because surveys were also taken in a train towards Harlingen during the Oeral festival.
12
These general sample aspects are congruent with the Quality research report 2007
Travel goal Frequency percentage
work 131 31,2 education 183 43,6 shopping 33 7,9 visiting 38 9,0 day out 28 6,7 something else 7 1,7 Total 420 100,0
4. Results
This chapter will describe the results from the quantitative research into three parts. The first part will address descriptive statistics; the second part will address the hypotheses which will lead to either a rejection or acceptation of these hypotheses and will these results will be inserted in the conceptual framework; the third part will emphasize additional analyses.
4.1 Descriptive statistics
In this paragraph descriptive statistics will be handled. The effect of respondents characteristics upon satisfaction are analyzed to gain deeper insights in perceptions of different types of travellers. Furthermore, justice is measured to find out if there are differences in justice perceptions between a high and low recovery effort. Interesting to acknowledge are the pre- and post-perceptions of respondents and the characteristics of these respondents. Table 4.1 gives more insights on satisfaction and perceived severity of the different types of respondents.
N Satisfaction Severity Pre-recovery Post-recovery ∆ Pre- recovery Post- recovery ∆ Age > 20 years 139 5,6853 4,3636 1,3217 2,8252 3,5315 ,7063 21 - 40 years 190 5,5053 4,4579 1,0474 2,8263 3,6720 ,8624 41 - 60 years 65 5,3385 4,1692 1,1692 2,7231 3,9077 1,1846 < 61 years 22 4,2273 3,0909 1,1364 3,2273 4,2381 1,0000 F 6,954 4,443 1,045 ,531 1,470 1,976 Sig. ,000*** ,004*** ,373 ,661 ,222 ,117 Travel frequency Frequent 180 5,5333 4,3889 1,1444 2,9778 3,6222 ,6444 Regularly 120 5,6583 4,5583 1,1000 2,5500 3,5042 ,9832 Sometimes 81 5,3827 4,1235 1,2593 2,8519 3,9753 1,1235 Rarely 39 4,8205 3,5641 1,2564 2,9744 3,9737 1,0000 F 3,568 3,790 ,275 1,801 1,646 3,098 Sig. ,014** ,011** ,844 ,146 ,178 ,027** Gender Female 234 5,5085 4,2265 1,2821 2,9316 3,8798 ,9614 Male 186 5,4301 4,4140 1,0161 2,7043 3,4486 ,7459 F ,304 1,230 3,752 2,032 6,546 2,663 Sig. ,582 ,268 ,053* ,155 ,011** ,103
