Linking service quality and recommendation: a study in higher education

Linking service quality and recommendation: a

study in higher education

Baoxun Wang

July 2009

University of Groningen

Linking service quality and recommendation: a

study in higher education

Master Thesis

Author:

Student Number:

Date:

Faculty:

Specialization:

Supervisor 1:

Supervisor 2:

Author’s Address:

Phone:

Management Summary

This master thesis focus on the following objective: to test the linkage between service quality and word-of-mouth recommendation in order to investigate whether service quality evaluation directly influences (students‟) word-of-mouth recommendation or whether satisfaction acts as a direct mediator on the relationship between service quality and word-of-mouth recommendation, specifically in higher education sector.

In this study, an experiment was executed in which a survey was used to collect the necessary information. The research was conducted at the university of Groningen, the Netherlands. A combination of both qualitative and quantitative methods was used. Data was gathered during a 3-week period between April and May 2009. In the end, 177 out of 180 questionnaires were completed.

Most of all, the empirical results suggest that linking service quality to (student‟s) word-of-mouth recommendation via satisfaction is reasonable. The mediating role of satisfaction has been confirmed in the linkage. The findings from this study reinforce the need for servie providers to devise operations strategies that focus on the dimensions of service quality that enhance satisfaction, which in turn can result in positive word-of-mouth recommendation.

Moreover, regarding service quality dimensions, four factors were captured in the present study, namely, “non-academic aspects”, “academic-programme aspects”, “comfort” and “access”. It is interesting to note that “academic-programme aspects” derived from this study is a combination of “academic aspects” and “programme issues” of HEdPERF scale. In this case, a recommendation for higher education institutions is to generally focus their resources on those attributes important to customers (i.e. students).

Additionally, we found out that three of four service quality dimensions have significant impacts on (students‟) satisfaction, namely, “non-academic aspects”, “academic-programme aspects”, “access”. In this sense, efforts should be made to signal current and future customers (students) about the quality of these three dominant service factors.

Preface

How time flies, my one-year master study in the Netherlands almost comes to an end. Hereby, I would like to express my sincere gratitude to my first supervisor, Dr. Liane Voerman for her patience and unceasing guidance through the course of this research. Next to that, I would like to thank Dr. Erjen van Nierop, my second supervisor for his valuable comments to adjust the research at the final phase of this thesis.

My appreciation also goes to my friends, Dong, Lidija, Mike for their great help in the research survey. A special thanks goes to Adela, for sharing happiness and story during my entire study.

At the end this observation seems to be terribly sad. I hope to go through my later journey of life with gratitude, still to feel pain and joy, to enjoy every sunrise and sunset. A quote comes to a conclusion:

“Never settle down. Stay hungry, stay foolish”- Steve Jobs.

To my parents and sisters for their continuing love and support.

Baoxun Wang

Contents

1. Introduction ... 7

1.1 Background ... 7

1.2 Problem statement ... 9

1.3 Research questions ... 10

1.4 Structure of this thesis ... 10

2. Theoretical Framework ... 11

2.1 Introduction ... 11

2.2 Service quality ... 11

2.2.1 Measuring service quality ... 11

2.2.2 Measuring Service Quality in higher education ... 16

2.3 The relationship between service quality and satisfaction ... 18

2.3.1 Service quality and satisfaction ... 18

2.3.2 Service quality and satisfaction in higher education ... 20

2.4 The differences between international students and local students... 21

2.5 Satisfaction as mediator between service quality and word-of-mouth recommendation ... 21

2.5.1 The mediating role of satisfaction ... 22

2.5.2 The mediating role of satisfaction in higher education ... 23

2.6.Conceptual Model ... 24 3. Empirical research ... 26 3.1 Introduction ... 26 3.2 Questionnaire development ... 26 3.2.1 Scale development ... 26 3.2.2 Other variables ... 27 3.2.3 Pre-test ... 27 3.3 Sample ... 29 3.4. Plan of analysis... 30

4. Results and discussions ... 32

4.1 Introduction ... 32

4.3 Test of hypotheses ... 33

4.3.1 Dimensions of service quality ... 33

4.3.2 The relationship between service quality and satisfaction ... 37

4.3.3 The differences between international students and local students ... 39

4.3.4 The mediating role of satisfaction ... 39

4.3.5 Concluding remarks regarding the test of hypotheses ... 42

5. Conclusions and implications ... 44

5.1 Introduction ... 44

5.2 Service quality dimensions ... 44

5.3 Service quality and satisfaction ... 45

5.4 The differences between international students and local students... 46

5.5 The mediating role of satisfaction ... 46

5.6 Limitations and directions for future research ... 47

References ... 49

1. Introduction

1.1 Background

Service industries are playing an increasingly important role in the economy of many nations. In today‟s world of global competition, rendering good quality of service is a key for success, and many experts concur that the most powerful competitive trend currently shaping marketing and business strategy is service quality (Firdaus, 2005). Over the past two decades, the theory and practice of service quality has received considerable attention from academics and practitioners alike. The pursuit for service quality has become an imperative factor for all organisations that are driven by the need to survive and remain competitive. The introduction of service quality in many service firms was as an element designed to effect competitive advantage (Shemwell,1998).

To date, the study of service quality, satisfaction and customer‟s behavioral intentions issues have dominated the services literature. There is an increased interest in understanding such important constructs as “service quality”, “customer behavioral intentions” and “customer satisfaction”. A substantial number of discussions has been both operational and conceptual, with particular attention given to identify the relationships among and between these constructs of service quality, satisfaction and customer‟s behavioral intentions.

There are many areas of pursuit, these seem to begin with the study of service quality, then carry through to satisfaction research and behavioral intentions. Consistent with this direction, this research focuses on the evaluation of the particular consumer of higher education: student, and analyzes the functional linkage between service quality and behavioral intention as well as the role of satisfaction in the path relationship. Despite competing theories, the service quality → satisfaction → behavioral intentions link has received the strongest support in the studies. The majority of studies indicate that service quality influences behavioral intentions only through satisfaction (e.g. Cronin and Taylor, 1992; Anderson et al., 1994; Rust and Oliver, 1994)). It may thus be stated that there is general consensus in published works that perceived quality is understood as an antecedent of satisfaction, and satisfaction mediates the relationship between service quality and behavioral intentions.

quality, particularly in higher education. Moreover, higher education institutions are increasingly placing greater emphasis on meeting the needs of their participating customers, that is, the students. The education services provided by different institutions attract students from around the world, the gap in the quality of education presumably should be very narrow. While the global market becomes more homogeneous, students‟needs of their education should become increasingly similar (Li-Wei Mai, 2005). In this regard, intense competition in today‟s competitive educational market forces universities to adopt a market orientation strategy to differentiate their offerings from those of their competitors (W.Deshields Jr et al, 2005). Overall, reduced public funding, global competition in the education sector, the freedom of students to choose the best attainable education they can receive, and the speed of information exchange have contributed greatly to the awareness and implementation of education quality (Li-Wei Mai, 2005).

Hereby, as a form of customer behavioral intentions , word-of-mouth recommendation is selected as one factor of behavioral intentions in this study. As we know, although word-of-mouth communication can be very influential in any purchase decision, it is particularly important in a service context, because services are intangible and, thus difficult to evaluate before purchase (Mazzarol, et al, 2007). Whilst in the greater part of services, complaint behaviour, word-of-mouth, loyalty, repetitive purchasing behaviour and profit are shown as consequences of satisfaction, some of these consequences do not make sense in higher education (Alves and Raposo, 2007). Unlike commercial products and services, loyalty is less an issue as behavioral intentions in higher education in terms of repeat purchases, but personal recommendation is (Li-Wei Mai, 2005). In addition, facing the wide range of education providers available, students usually make decisions by referring to quality assurance systems such as word-of-mouth recommendations (Cuthbert, 1996). A long term relationship with students can provide an institution with a type of competitive advantage, particularly at a positive word of mouth level concerning potential, present and future students (Alves and Raposo, 2007). More importantly, with the growth and evolution of the internet, electronic peer-to-peer referrals have become an important phenomenon. E-mail referrals, online forums of users and newsgroups, as well as customer reviews encouraged by merchant websites allow consumers to share information far more easily than ever before (Bruyn and Lilien, 2008).

In addition to the relationships between service quality, satisfaction, word-of-mouth, the differences in the perception of service quality and the level of satisfaction between international students and local students will be investigated, since foreign student numbers have grown, universities have increasingly paid attention to cultural, individual and national differences between international students and their local counterparts.

1.2 Problem statement

There is no wonder that researchers have given considerable time and effort in assessing service quality and satisfaction and also in investigating the relationships which ultimately end in a form of behavioral intentions (word-of-mouth recommendation). In recent years, most of the published work on quality assessment has tended towards the study of external or market effects (e.g. Parasuraman et al., 1988; Gronin and Taylor, 1992; Firdaus Abdullah, 2005 ) . The emergence of different scales for measuring perceived service quality (namely SERVQUAL, SERPERF, etc) has facilitated the assessement of these external effects (Bou-Llusar et al, 2001). However, regarding the measures of service‟s quality dimensions, there is still little consensus as to which method is universally most suitable and it therefore remains in the judgement of the individual researcher to select the most appropriate framework for a given situation ( Angell et al, 2008). Furthermore, Jusoh et al (2004) hold that the development of the dimensions in service quality is keen on expanding because the nature of the higher learning institution itself is dynamic and unique. Thus, it is advisable to reassess the quality dimensions in a specific context, such as academic aspects and non-academic aspects.

Besides, it is meaningful to assess the differences in the perception of service quality and the level of satisfaction between international students and local students, since foreign student numbers have grown, universities have increasingly paid attention to cultural, individual and national differences between international students and their local counterparts. A good understanding of the differences between international students and local students will help education institutions improve their service performance in the light of increased competition with the development of global education markets.

Overall, this leads to the following problem statement:

How can a higher education institution link service quality to (student’s) word-of-mouth recommendation and what is the role of satisfaction in this linkage?

1.3 Research questions

To gain a deep insight into the problem statement, the following research questions are formulated:

1) What are the structural dimensions of service quality?

2) What are the relationships between the main dimensions of service quality and (student‟s) satisfaction?

3) What are the differences in the perception of service quality and the level of satisfaction between international students and their local counterparts?

4) How does satisfaction mediate the relationship between service quality and (student‟s) word-of-mouth recommendation?

1.4 Structure of this thesis

2. Theoretical Framework

2.1 Introduction

In this chapter, the theoretical framework will be explained. In this theoretical framework the following subjects will be discussed:

 Measuring service quality

 The relationships between the main dimensions of service quality and satisfaction  The differences between international students and local students

 The mediating role of satisfaction on the relationship between service quality and customer behavioral intentions (i.e. word-of-mouth recommendation).

In addition, various hypotheses will be established based on the theoretical findings of this framework, which serves as the foundation of a conceptual model, that will be tested in an empirical, higher-education setting.

2.2 Service quality

In this part service quality will be analysed in general. After this general analysis a more specific examination about service quality in higher education will be given.

2.2.1 Measuring service quality

2.2.1.1 SERVQUAL SCALE

In the last decade, the diverse instruments of measurement emerged, among them, the most widely used methods is the SERVQUAL instrument ( Parasuraman et al., 1988), which grounded in the gap model ( Parasuraman et al., 1985) that defined service quality construct as an attitude resulting from customers‟ comparision of their expectations with their perceptions of the service encounter. As indicated, SERVQUAL was purported to be a generic measure of service quality which would identify the difference in service encounter along five service quality dimensions – reliability, responsiveness, assurance, empathy and tangibles (Parasuraman et al., 1988). According to this measurement, service quality could be assessed by a 22-item scale - SERVQUAL. This scale was designed to measure the discrepancy between consumers‟ perceptions and expectations based on a seven-point likert scale. In addition, the revised version of SERVQUAL (Parasuraman et al., 1991) introduces a number of changes and Parasuraman et al., (1994) also develops three alternative questionnaire formats to address the difficulty of service quality measurement. Parasuraman et al. (1988, 1991) describe SERVQUAL as a concise multiple-item scale with good reliability and validity and offering a number of potential applications across a broad spectrum of services. Sureshchandar et al,. (2002) also claim that SERVQUAL instrument, a 22-item scale that measures service quality along five factors, forms the cornerstone on which all other works have been built.

2.2.1.2 Criticism of SERVQUAL

Several studies using SERVQUAL scale, however, have demonstrated the existence of difficulties resulting from the conceptual or theoretical component as much as from the empirical component. Despite the popularity of SERVQUAL, a number of criticisms was directed at this scale, aimed at both the conceptual and the operational level (Firdaus Abdullah, 2005).

Francis Buttle (2006) summarizes the major criticisms of SERVQUAL in two categories – theoretical and operational:

1) Theoretical:

cognitive content, satisfaction may be more heavily loaded with affect. According to Andersson (1992), SERVQUAL fails to draw on previous social research, particularly economic theory, statistics, and psychological theory.

b) Gaps model: there is little evidence that customers assess service quality in terms of P-E gaps (Babakus and Boller, 1992; Lacobucci et al, 1994; Babakus and Inhofe, 1991; Teas, 1993a, 1993b, 1994; Gronroos, 1993; Wotruba and Tyagi, 1991; Gronin and Taylor, 1992). Babakus and Boller (1992) claim that the difference scores do not provide any additional information beyond that already contained in the perceptions component of the SERVQUAL scale. Lacobucci et al (1994) also suggest that expectations might not exist or be formed clearly enough to serve as a standard for evaluation of a service experience. According to Babakus and Inhofe (1991), expectations may attract a social desirability response bias. Teas (1993a, 1993b, 1994) suspect the meaning of identified gaps, as well. A further criticism is that SERVQUAL fails to capture the dynamics of changing expectations (Gronroos, 1993). Wotruba and Tyagi (1991) agree that more work is needed on how expectations are formed and changed over time. c) Process orientation: SERVQUAL focuses on the process of service delivery,

not the outcomes of the service encounter ( Cronin and Taylor, 1992; Mangold and Babakus, 1991; Richard and Allaway, 1993). Critics have argued that outcome quality is missing from SERVQUAL scale‟s formulation of service quality.

d) Dimensionality: SERVQUAL‟s five dimensions are not universals (Garman‟s, 1990; Saleh and Ryan‟s, 1992; Gagliano and Hathcote‟s, 1994; Bouman and van der Wiele‟s, 1992; Babadus et al‟s, 1993b; Babadus and Boller, 1992). The SERVQUAL instrument has been employed in modified form, up to nine distinct dimensions of service quality have been revealed, the number varying according to the service sector under investigation. Nine factors accounted for 71 percent of SQ variance in Garman‟s (1990) hospital research. Five factors were retained in Saleh and Ryan‟s (1992) work in the hotel industry. Four factors were extracted in Gagliano and Hathcote‟s (1994) investigation of SQ in the retail clothing sector, whereas three factors were identified in Bouman and van der Wiele‟s (1992) research into car servicing and one factor was recognized in Babadus et al‟s (1993b) survey of 635 utility company customers. Overall, Babadus and Boller (1992) comment that the domain of service quality may be factorially complex in some industries and very simple and unidimensional in others.

2) Operational:

battery in the SERVQUAL instrument. In addition, Gronroos (1993) claims the bad-service paradox, which means that a customer may have low expectations based on previous experience with the service provider, if those expectations are met there is no gap and service quality is deemed satisfactory.

b) Item composition: four or five items can not capture the variability within each SQ dimension (Garman‟s, 1990; Bouman and van der wiele, 1992; Saleh and Ryan, 1992; Fort, 1993; Babakus and Mangold, 1992). Each factor in SERVQUAL scale is composed of four or five items. It has become clear that this is often inadequate to capture the variance within, or the context-specific meaning of each dimension. Garman‟s (1990) research of hospital services employed 40 items. Bouman and van der wiele (1992) uses 48 items in their car service study, whereas 33 items in Saleh and Ryan‟s (1992) hospitality industry research, Fort (1993) 31 items in his analysis of software house service quality and Babakus and Mangold (1992) 15 items in their hospital research.

c) Moment of truth (MOT): customers‟ assessments of SQ may vary from MOT to MOT (Garman, 1990). Many services are delivered over several moments of truth or encounters between service staff and customer.

d) Polarity: the reversed polarity of items in the scale causes respondent error (Bakakus and Boller, 1992; Babakus and Mangold, 1992). Of the 22 items in the SERVQUAL, 13 statement pairs are positively worded, and nine pairs are negatively worded. Bakakus and Boller (1992) criticize that item wording may be responsible for producing factors that are method artifacts rather than conceptually meaningful dimensions of service quality. Babakus and Mangold (1992), in their application of SERVQUAL to a hospital setting, therefore decide to employ only positively-worded statements.

e) Two administrations: two administrations of the instrument causes boredom and confusion (Bouman and van der Wiele, 1992; Garman, 1990; Gronroos, 1993; Lewis, 1993; Babakus and Boller, 1992; Babakus et al, 1993b; Glow and Vorhies, 1993). Respondents appear to be bored, and sometimes confused by the administration of E and P versions of SERVQUAL (Bouman and van der Wiele, 1992). Garman (1990) also comments on the timing of the two administrations.

measure.

2.2.1.3 SERVPERF SCALE

As claimed by many researchers (Babakus and Boller, 1992; Lacobucci et al, 1994; Babakus and Inhofe, 1991; Teas, 1993a, 1993b, 1994; Gronroos, 1993; Wotruba and Tyagi, 1991), there is little evidence that customers assess service quality in terms of P-E gaps. Cronin and Taylor (1992) also suspects the notion of the „expectation minus performance‟ gap as a basis for measuring service quality. They challenged the framework of SERVQUAL, developing their own „performance only‟ measure of service quality, called SERVPERF. In their empirical work, they found that the performance-based scale developed (SERVPERF) is efficient in comparison with the SERVQUAL scale. Cronin and Taylor (1994) advocate that the performance-based measures of service quality captured by the SERVPERF scale can provide a longitudinal index of the service quality perceptions of a service firm‟s constituencies. The SERVPERF scale can provide managers with a summed overall service quality score that can be plotted relative to time and specific consumer subgroups. As such, the SERVPERF scale provides a useful tool for measuring overall service quality attitudes by service managers (Cronin and Taylor, 1994). What‟s more, Brady et al (2002) replicate and extend Cronin and Taylor‟s (1992) research and further confirm the superiority of SERVPERF as a more appropriate method for measuring service quality.

In spite of the theoretical foundation and empirical support documented in the literature, the SERVPERF has only been operationalized as a summed index (derived by averaging the distinctive dimensions of service quality) with regard to its predictive value in relation to other outcome constructs such as satisfaction and behavioral intentions (Lianxi Zhou, 2004). Cronin and Taylor (1994) also suggest that although the aggregation of the SERVPERF dimensions is useful for the purpose of comparative analysis across alternative models and service industries, great care should be exercised by managers of service firms in attempts to derive more specific information from data derived using the SERVPERF scale for strategic decision-making.

2.2.1.4 Concluding remarks regarding the measurement scales

modification to fit the specific application situation and supplemental context-specific items. The general conclusion appears to be that industry-specific service quality measures may be a more viable research strategy to pursue ( Zeithaml et al, 1985; Cronin and Taylor, 1992). The concerns raised indicate that there is still scope for further research in the subject of service quality, which is, by nature, an abstract concept that is difficult to comprehend (Sureshchandar, et al. 2002).

2.2.2 Measuring Service Quality in higher education

Also, within the tertiary education sector service quality has attracted considerable attention (Firdaus, 2005). Firdaus (2005) reveals that previous studies have produced scales that bear a resemblance to the generic measures of service quality, which may not be totally adequate to assess the perceived quality in higher education, since higher education setting has its particular charateristics in which the replication of these generic instruments (such as SERVQUAL and SERVPERF) is still hazy.Thus, it would seem rational to develop a new measurement scale that incorporates not only the academic components, but also aspects of the total service environment as experienced by the student.

Following this argument, Angell et al (2008) develop a framework to measure service quality in postgraduate education. Using qualitative and quantitative analyses, the service attributes were finally reduced into four service quality factor: “academic”, “leisure”, “industry links”, and “cost”. It is worth noting that this study was conducted on a postgraduate-based research. Other universities should adapt this framework to their own needs when measuring service quality.

In another empirical study, Owlia and Aspinwall (1998) also develop a framework for quality measurement in higher education. Following a theoretical study of the dimensions of quality in education sector, the paper reports on the empirical research carried out to validate the framework through internal consistence testing, factor analysis and predictive validity. A customer-oriented strategy for measurement was selected on the basis of customer‟s perceptions on quality dimensions, the consisted of a list of 19 items grouped into four dimensions: academic resources, competence, attitude and content. However, due to the emphasis being on the teaching aspects of higher education, the scale measurement based on this research is not viable to apply to the generic measures of service quality in higher education.

test and expert validation (Firdaus, 2005). The proposed 41-item instrument has been empirically tested for unidimensionality, reliability and validity using both exploratory and confirmatory factor analysis. Research findings confirm that the six dimensions, i.e. non-academic aspects, academic aspects, reputation, access,

programme issues and understanding were distinct and conceptually clear. He claims

that it can be posited that student perceptions of service quality can be considered as a six-factor structure consisting of the identified six dimensions.

Another study by Firdaus (2006) attempts to empirically compare and contrast the HEdPERF scale against two alternatives (the SERVPERF and the merged HEdPERF- SERVPERF scales) in order to assess the relative strengths and weaknesses of each instrument and to determine which instrument has the superior measurement capability in terms of unidimensionality, reliability, validity and explained variance of service quality. The empirical analysis indicates that a modified five-factor structure with 38 items resulted in more reliable estimations, greater criterion and construct validity, greater explained variance, and consequently a better fit. The modified HEdPERF scale also had the advantage of being more specific in areas that are important in evaluating service quality within the higher education sector. So, service quality in higher education can be deemed as a five-factor structure with conceptually clear and distinct dimensions, namely, non-academic aspects, academic aspects,

reputation, access, programme issues. The five factors identified can be described as

Table 1 (The original items included in each dimension on HEdPERF are shown in appendix 1).

Table 1. Descriptions of five factors in HEdPERF

Dimensions Descriptions

Factor 1:non-academic

aspects

This factor consists of items that are essential to enable students fulfill their study obligations, and it relates to duties carried out by non-academic staff. Factor 2:academic aspects The items that describe this factor are solely the responsibilities of

academics.

Factor 3:reputation This factor is loaded with items that suggest the importance of higher learning institutions in projecting a professional image.

Factor 4:access This factor consists of items that relate to such issues as approachability, ease of contact, availability and convenience.

Factor 5:programmes issues

This factor emphasizes the importance of offering wide ranging and reputable academic programmes/specializations with flexible structure and syllabus.

“non-academic aspects”, “academic aspects” and “programme issues” are more related to what service is provided, involving in technical dimension. On the other hand, “reputation” and “access” are more considered as functional quality, standing for how the service is provided.

Based on the review of service quality literatures and the context of this study, I include all five dimensions of HEdPERF in this study. Therefore, the hypothesis is proposed below:

H 1: There are five distinct dimensions ( namely, non-academic aspects, academic aspects, reputation, access, programme issues), which clearly measure service quality in higher education sector.

2.3 The relationship between service quality(SQ) and satisfaction(SAC)

In this section the general relationship between service quality and satisfaction will be described. After this analysis the relationship between service quality and satisfaction in higher education will be discussed.

2.3.1 Service quality and satisfaction

Regarding the relationship between service quality and customer satisfaction, it is normally assumed that there is a positive association between service quality and customer satisfaction (Rust and Oliver, 1994).

In a study by Bou-Llusar et al. (2001), the results show that firm perceived quality exerts a definite and statistically significant influence on satisfaction. Similarly, an investigation in the hotel industry by Hu et al. (2009) identifies that service quality has positive impacts on customer satisfaction. In the context of internet banking, Rod et al. (2008) reveal three dimensions of internet banking service quality directly affecting customer perceptions of overall internet banking service quality which influence overall customer satisfaction with the bank. In their study, the strong positive association between internet banking service quality and customer satisfaction suggests that when internet banking service quality is perceived to be high, customers are more likely to be satisfied with their online service and consequently will be more satisfied with their bank.

To be precise, as satisfaction can result from any of the dimensions of service quality (Rust and Oliver, 1994), a dimension-specific analysis of the relationship between service quality and satisfaction is likely to provide more diagnostic value for improvement of service quality (Lianxi Zhou, 2004). In other words, some of the service dimensions seem to be more relevant to customer satisfaction. In the study of retail banking, Lianxi Zhou (2004) attempts to assess the relationship between these specific dimensions of service quality and consumer satisfaction. His study further asserts that specific dimensions of service quality are a source of consumer satisfaction (See Figure 1). The empirical evidence of his study, based on a convenience sample of retail bank consumers in China, depicts that the service quality dimension of reliability/assurance primarily drives the satisfaction of target markets, whereas the other dimensions remain insignificant. Therefore, it is intuitively viable to hold that there are positive relationships between specific dimensions of service quality and customer satisfaction.

Figure 1. Dimensions as determinants of overall satisfaction

2.3.2 Service quality and satisfaction in higher education

Similar to other service industries, a direct relationship between specific dimensions of service quality and satisfaction in higher education is strongly supported.

That is, a study in higher education by Li-Wei Mai (2005) confirms that satisfaction is a result of the perception of service quality. Specifically, the research found that the “overall impression of the school” and “overall impression of the quality of education” ( key dimensions of service quality) are two significant predictors for the “overall satisfaction of the education”, more than other service dimensions.

Furthermore, a research conducted by Alves and Raposo (2007) investigates the factors which influence student satisfation in higher education (See Figure 2). The findings show that the influence of quality perceived was seen that apart from being significant, this was more related with the technical quality of the education service. Figure 2. Conceptual model in Higher Education (Alves and Raposo, 2007)

As deemed, non-academic aspects, academic aspects and programme issues are involved in technical dimension, while reputation and access are considered as functional quality.

To sum up, based on the discussion previously, with regard to the relationship between service quality and satisfaction, the following hypotheses are proposed: H2a: There are positive associations between service quality dimensions (non-academic aspects, academic aspects, reputation, access, programme issues) and satisfaction.

2.4 The differences in the perception of education quality and the level of satisfaction between international students and local students

As foreign student numbers have grown, universities have increasingly paid attention to cultural, individual and national differences between international students and local students. In their striving for higher education, international students pursue values that differ from local students. This pursuit of different values may be that international students in a foreign country tend to achieve several objectives. These include improving their language skills, providing an opportunity for independence from parents, experiencing different cultures and generally gaining new life experiences (Gabbott et al, 2002). Therefore, it is normally assumed that there are differences in the perception of education quality and the level of satisfaction between international students and their local counterparts ( Li-Wei Mai, 2005).

A study by Li-Wei Mai (2005) confirms that there is indeed a difference in the perception of education quality between foreign students and local students. Her results also found that the foreign students expressed significantly lower levels of satisfaction compared with domestic students. She explained that as the home and foreign students received the same education when the service quality is constant for both groups, it is plausible to infer that foreign students have higher levels of expectation compared with home students when they enroll for a degree. Thus, international students have lower levels of overall satisfaction towards service received.

Based on the points raised above, the hypotheses are established below:

H3a: International students have lower mean scores for service quality than local students.

H3b: International students have lower levels of satisfaction compared with local students.

2.5 Satisfaction as mediator between service quality and word-of-mouth recommendation

general and in higher education sector will be explained separately.

2.5.1 The mediating role of satisfaction

The majority of studies indicate that service quality influences behavioral intentions only through satisfaction (e.g. Cronin and Taylor, 1992; Rust and Oliver, 1994). Despite competing theories (e.g. Bolton and Drew, 1991; Parasuraman et al., 1998), the service quality → satisfaction → behavioral intentions link has received the strongest support in the studies (e.g. Anderson et al., 1994; Cronin et al., 2000). As defined by Oliver (1981), satisfaction acts as the summary psychological state resulting when the emotion surrounding disconfiemed expectations is coupled with the customer‟s prior feelings about the consumption experience. Satisfaction is expected to directly affect behavioral intentions due to its more emotive nature (Oliver, 1997). Customer satisfaction is a much better predictor of behavioral intentions (Dabholkar et al., 2000).

Based on theoretical framework and empirical evidence, many researchers support the mediating role of satisfaction (e.g. Cronin and Taylor, 1992; Anderson et al., 1994; Rust and Oliver, 1994). Cronin et al (2000) also claim for consideration of the indirect effect that service quality has on consumers‟ behavioral intentions (i.e., word-of-mouth recommendation) through customer satisfaction, which is similar with the observations by Cronin and Taylor (1992) , Bou-Llusar et al (2001) and Brady and Robertson (2001). In the work of Cronin and Taylor (1992), the direct effect of service quality on behavioral intentions is not significant, instead there is an indirect effect through satisfaction. Similarly, Bou-Llusar et al (2001) identify that overall satisfaction acts as a mediating variable on the relationship between firm perceived quality and purchase intentions. Specifically, the results obtained showed that firm perceived quality exerts a definite and statistically significant influence on satisfaction, satisfaction has a positive and statistically significant effect on purchase intentions. As well, a study by Brady and Robertson (2001) employed a cross-cultural perspective to explore the antecedent role of service quality and satisfaction in the development of service customer‟s behavioral intentions. Their results showed that the effect of service quality on behavioral intentions is mediated by a consumer‟s level of satifaction and that this relationship is consistent across cultures.

service quality and behavioural intentions was not significant. The results nonetheless show that service quality has an indirect effect on behavioural intentions via service satisfaction.

Moreover, an investigation by Dabholkar et al. (2000) presented a mediator model of customer satisfaction (See Figure 3). According to their results, a strong mediating role was found. In other words, customer satisfaction strongly mediates the effect of service quality on behavioral intentions.

Figure 3. Mediator Model of Customer Satisfaction (Dabholkar et al, 2000)

2.5.2 The mediating role of satisfaction in higher education

The mediating role of satisfaction is obviously found in higher education sector, as well. In the context of higher education, a study conducted by Deshields et al (2005) investigated the role of satisfaction and intentions on retention by incorporating a number of factors that are assumed to have impact on satisfaction, which, in turn, would influence intentions. Their results showed that students‟s college experience influences their satisfaction, which in turn influences student intentions to stay (See Figure 4). Student college experience is related to student service experience, incorporating different aspects of service quality. In this sense, student experience of service quality would impact satisfaction, which in turn would influence student‟s behavioral intentions. In other words, there seems to exist a mediational effect of satisfaction on the relationship between service quality and student‟s behavioral intentions (i.e. word-of-mouth recommendation) in higher education sector.

Figure 4. Research model in higher education (Deshields et al, 2005)

Based on the discussion previously, the hypothesis is established below:

H4: Satisfaction mediates the relationship between service quality and word-of-mouth recommendation.

2.6.Conceptual Model

Based on the previous review of the literature, a model is developed in Figure 5 to conceptualize the theoretical framework of the study. The model shows the dimensions of the service quality in higher education from customer‟s (student‟s) perspective and a link between service quality and word-of-mouth recommendation, as well as the differences in the perception of education quality and the level of satisfaction between international students and local students.

3. Empirical research

3.1 Introduction

To verify the theoretical framework (fig.5) and to test the hypotheses, an experiment was executed in which a survey was used to collect the necessary information needed to analyze statistically if the proposed hypotheses could be accepted or should be rejected. The research was conducted at the university of Groningen, the Netherlands. A combination of both qualitative and quantitative methods was used. This required the application of two specific research methods including two focus group interviews and the execution of a quantitative survey instrument. During the focus group stage, participants were questioned on those service attributes deemed important to students when rating their perceived service quality, aiming at finding and solving potential problems in the quality measurement scale. This more qualitative stage of the research was conducted to support the development of the multi-item questionnaire used for further quantitative research. It is important to note that this survey adaptation is similar to the essence of Parasuraman et al.‟s (1994) approach. After that, a pre-test of the questionnaire was realised with 10 students. A full survey was then executed.

3.2 Questionnaire development

3.2.1 Scale development

In this scale development stage, participating students were first interviewed about the quality elements. Then, the items identified were reviewed by a second group of students.

First of all, six participants, all of whom were master students in Marketing Program, were interviewed; they would be a good position to judge the quality elements regarding the university and they would help in improving the research method in some way as they gained in-depth knowledge about marketing research. During the interview, they were asked to answer the perception version of the 38 original items and express any difficulty in understanding or answering questions, as well as elaborating on any additional factors. On the basis of suggestions, some changes were made. First, 9 of the original HEdPERF scale items were deleted due to their perceived irrelevance to the research setting ( e.g., „hotel facilities and equipment‟, „ideal campus locatin/layout‟ )1

. In addition, two items were combined as one, since

1

they were seen as the description of same issue („quality programmes‟ and „reputable academic programmes‟). Moreover, one item related to library was added since it was considered as an vital factor. The result was a 29-item framework of service quality which was used for further investigation in the second interview research.

Subsequently, the items were reviewed by the second focus group consisting of 10 students in a student house (Plutolaan). They were asked to explain any problem with the questions. Appropriate item wording was modified to better fit the research setting. For instance, “Institution” was replaced with “University”, and “Academic facilities” was replaced with “Technological facilities”. Furthermore, some examples were added to make the questions clearer. Moreover, the label of “Reputation” is changed as “comfort” representing those items that make students feel comfortable and confident of the university, because of the deemed confusion of the name “Reputation”. Agreement was reached, upon concluding this stage of the research, that the 29 items (See table 2: SQ1 to SQ29) included on the final measurement instrument were relevant to the domain of service quality evaluation in the context of higher education service.

3.2.2 Other variables

In addion to the measures of service quality, the other variables (See table 2) were included to make the questionnaire operational:

 Overall service quality (OSQ): this variable was included. Evaluating customer‟s opinions on the overall service quality can serve as a separate measure with which the detailed perception measures may be compared (Owlia and Aspinwall, 1998). A 7-point item was applied, with endpoints very poor and very good, “The

overall quality of the university's services is”.

 Overall satisfaction (Satisfac): The overall satisfaction was assessed by a single item. This item was modified from Firdaus Abdullah‟s (2006) Satisfaction variable. The seven-point item was used with a range of values from 1 (very dissatisfied) to 7 (very satisfied), “My feelings towards the university' services

can best be described as” .

 Word-of-mouth recommendation (Rec): Recommdation was captured with a 7-point, with endpoints very unlikely and definitely likely.

29 items of service quality identified in the previous step were assessed using Likert scales, with a range of values from 1 (strongly disagree) to 7 ( strongly agree).

3.2.3 Pre-test

your experience, how do you rate your general level of satisfaction with the university”, according to Oliver, R.L. (1997)‟ satisfaction scale. The results of the pre-test also demonstrated that the questionnaire was easily understood by the students and that it took approximately 5 minutes to fill in. (A complete quesionnaire is shown in Appendix 2).

Table 2. Variables

Variables Descriptions

Other variables:

1.OSQ Overall service quality

2. Satisfac Satisfaction

3. Rec word-of-mouth recommendation Non-academic aspects:

4. SQ1 Sincere interest in solving problem

5. SQ2 Caring and individualized attention

6. SQ3 Efficient/prompt dealing with complaints

7. SQ4 Convenient opening hours

8. SQ5 Positive attitude

9. SQ6 Good communication

10. SQ7 Knowledgeable of systems/procedures Academic aspects:

11. SQ8 Knowledgeable in course content

12. SQ9 Caring and courteous

13. SQ10 Sincere interest in solving problem

14. SQ11 Positive attitude

15. SQ12 Good communication

16. SQ13 Feedback on progress

17. SQ14 Educated and experienced academicians Comfort:

18. SQ15 Technological facilities

19. SQ16 Excellent quality programmes

20. SQ17 Recreational facilities

21. SQ18 Library

22. SQ19 Minimal class sizes

23. SQ20 Easily employable graduates Access

24. SQ21 Equal treatment and respect

25. SQ22 Fair amount of freedom

26. SQ23 Confidentiality of information

27. SQ24 Easily contacted by telephone/Email

28. SQ25 Student's Union

29. SQ26 Feedback for improvement

30. SQ27 Service delivery procedures Programme issues

32. SQ29 Flexible syllabus and structure

3.3 Sample

Due to the technical impossibility of researching all students at the university, a sample was selected with the objective of reflecting the opinions of both international students and local students (Dutch). A snowball sampling method was used in this study. According to Malhotra and Birks (2007), snowball sampling is a technique for developing a research sample where existing study subjects generate additional subjects. Specifically, initial subjects from various faculties were recruited to access to potential respondents. As the sample built up, questionnaires were delivered to the subjectes. Data was gathered during a 3-week period between April and May 2009. In the end, 177 out of 180 questionnaires were completed.

A list of the responding sample‟s characteristics is presented in Table 3. In the sample, 46.9% (83) of the repondents were female, and 53.1% (94) were male. The majority of subjects (88.1%) were between 21 and 35 years old. 49.2% (87) of the respondents were international students, and 50.8% (90) were Dutch students2. The subjects (63.3%) were mainly master students, while 23.7% (42) of bachelor students and 13% (23) of pre-master students. The subjects were mostly from the Faculty of Economics and Business (37.9%).

Table 3. Sample’s characteristics Variables n (%) Gender Female 83 (46.9) Male 94 (53.1) Age 20 and below 19 (10.7) 21-35 156 (88.1) 36-45 2 (1.1) Nationality International 87 (49.2) Local (Dutch) 90 (50.8) Level of study Bachelor 42 (23.7) Pre-master 23 (13.0) Master 112 (63.3) Faculty of study

Economics and Business 67 (37.9)

Behavioural and Social Sciences 27 (15.3)

2

Arts 26 (14.7)

Law 31 (17.5)

Medical Sciences 10 (5.6)

Philosophy 9 (5.1)

Mathematics and Natural Sciences 7 (4.0)

3.4. Plan of analysis

In order to test the established hypotheses, it is necessary to analyze and interpret the data. Thus, the analysis of data was processed using the Statistical Package for the Social Science (SPSS). The ways to test the hypotheses are:

1) To identify the dimensional structure of service quality within the higher education sector, factor analysis was used in this study, as factor analysis provides the tools for analyzing the structure of the interrelationships among a large number of variables by defining highly interrelated variables into sets of factors with reducing number (Hair et al., 2006). The first step involves assessing the appropriateness of factor analysis using KMO and Bartett‟s test of phericity. According to Hair et al. (2006), KMO and Bartett‟s test of phericity are run to justify the adequacy of correlations between items. In the second step, all the 29 items of service quality were subjected to a factor analysis utilizing the principal components procedure with a varimax rotation. As suggested by Hair et al. (2006), varimax rotation helps in simplifying the columns of the factor matrix and maximizes the sum of variances of required loadings of the factor matrix, in which manner gives a clearer separation of the factors. The decision to include a variable in a factor was based on factor loadings greater than 0.4 (Hair et al, 2006). After the retention of factors, the reliability of the composite score was examined.

2) To assess whether there is a significant association between two or more variables is done with the usage of the regression analysis (Malhotra and Birks, 2007). Regression analysis describes the relationship between a dependent variable and one or more independent variables. A regression analysis has been performed to find out which factors of service quality have significant influence on satisfaction. Additionally, the relative importance of technical and functional quality on satisfaction was examined.

3) The independent t-test is used to compare the statistical significance of a possible difference between the means of two groups on some independent variable (Malhotra and Birks, 2007). In this case, An independent t-test was used to examine whether there are any significant differences in the perception of education quality and the level of satisfaction between international students and local students.

occur when (1) model 1: the independent variable (service quality) significantly affects the mediator (satisfaction), (2) model 2: the independent variable (service quality) significantly affects the dependent variable (word-of-mouth recommendation) in the absence of the mediator (satisfaction), (3) the mediator (satisfaction) has a significant unique effect on the dependent variable (word-of-mouth recommendation), and (4) model 4: the effect of the independent variable (service quality) on the dependent variable (word-of-mouth recommendation) becomes insignificant upon the addition of the mediator (satisfaction) to the model.

Figure 6. Regression models

Satisfaction _{Word of mouth}

recommendation Model 3

Service

4. Results and discussions

4.1 Introduction

In this chapter, the description of data will first be presented. Subsequently, the hypotheses will be tested. This chapter ends up with concluding remarks as to the test of hyphtheses.

4.2 Descriptive analysis

Mean scores as well as standard deviations for “overall service quality”, “satisfaction”, “word-of-mouth-recommendation”, and 29 items of service quality are shown in Table 4. According to Table 4, the average responses to “satisfaction” and “word-of-mouth recommendation” were relatively high with above “5.5” on seven-point scale, while “overall service quality” also gained a mean score of nearly “5.0”. Regarding the mean scores of the 29 items of service quality, they all laid between “4.44” and “5.78”. Among these items, “Equal treatment and respect” had the highest score with “5.90”, while a lowest score for “good communication” with “4.44”. From these results, we can basically see that students rated somewhat high scores for all the variables.

Table 4. Descriptive analysis. Variables

Mean

Std. Deviation

Overall Service Quality 4.92 1.092

Satisfaction 5.52 0.983

Word-of-mouth Recommendation 5.78 1.051

Sincere interest in solving problem 5.25 0.997

Caring and individualized attention 5.56 0.852

Efficient/prompt dealing with complaints _{4.6 1.523}

Convenient opening hours 5.2 1.12

Positive attitude 4.47 1.581

Good communication 4.44 0.897

Knowledgeable of systems/procedures 5.32 1.068

Knowledgeable in course content 5.29 1.144

Caring and courteous 4.84 1.119

Positive attitude 4.88 1.291

Good communication 5.16 1.176

Feedback on progress 5.07 1.173

Educated and experienced academicians _{5.11 1.222}

Technological facilities 5.4 1.154

Excellent quality programmes 5.3 1.121

Recreational facilities 4.57 0.946

Library 5.21 1.136

Minimal class sizes 5.7 0.998

Easily employable graduates 4.95 1.311

Equal treatment and respect 5.9 0.84

Fair amount of freedom 5.55 0.797

Confidentiality of information 5.77 0.858

Easily contacted by telephone/Email 5.76 1.108

Student's Union 5.67 1.121

Feedback for improvement 5.59 1.115

Service delivery procedures 5.4 1.001

Variety of programmes/specializations 5.08 1.107

Flexible syllabus and structure 5.37 1.064

Note: Numbers measured on a seven-point scale

4.3 Test of hypotheses

4.3.1 Dimensions of service quality

4.3.1.1 Adequacy of factor analysis

4.3.1.2 Identification of SQ dimentions

In order to gain a better understanding of the factor structure, all 29 items of service quality were subjected to a factor analysis utilizing the principal components procedure with a varimax rotation.

Appendix 4 contains the information regarding the 29 possible factors and their relative explanatory power as expressed by their eigenvalues and common variance. By assessing the importance of each component, 8 factors are chosen since their eigenvalues are greater than 1 (Hair et al, 2006). But Hair et al. (2006) suggest that variables should generally have communalities of greater than .50 to be retained in the analysis. In this regard, factor 6, 7, 8 are unreasonable to be retained. In addition, 4 factors account for 51.58% of the total variance of 29 variables while 5 factors (56.59%), which are deemed to be sufficient. Meanwhile, the scree test (Appendix 5) indicates that 4 or 5 factors may be appropriate when considering the changes in eigenvalues, i.e. the “elbow” in the line. Combining all these criteria, 4- factor and 5-factor solutions are retained for further investigation.

Besides of the criteria above, factor loading (the contribution of a variable to a component) is also critical in determining the number of factors (Hair et al, 2006). And varimax rotation is used directly to facilitate this process. By assessing the factor matrix, we can find that the variable of SQ2 has a cross-loading in factor 2 and factor 3, and the variable of SQ21 has a cross-loading in factor 1 and factor 5, showed in the 5-factor solution (Appendix 6). The difficulty arises because a variable with several significant loadings (across-loading) must be used in labeling all the factors on which it has a significant loading (Hair et al, 2006). On the other hand, a 4-factor solution (Appendix 7) has a clear separation of the factors, making each variable associating with only one factor. Thus, a 4-factor solution in that matter is superior to a 5-factor solution, as the factors in 5-factor solution may overlap. Therefore, 4 factors are finally retained. Table 5 shows the results of the factor analysis in terms of factor name, the variables loading on each factor. A remarkable point of this result is that the original dimensions of “academic aspects” and “programmes issues” are combined as one new factor “ academic-programme”, the other dimensions of “non-academic aspects”, “comfort” and “access” remain the same. Four factors are:

 Factor 1: non-academic aspects. This factor consists of items that are essential to enable students fulfill their study obligations, and it mainly relates to duties carried out by non-academic staff (this factor is similar to that in Firdaus‟s (2006) HEdPERF scale) .

 Factor 2: academic-programme aspects. The items that describe this factor are mostly the responsibilities of academics, as well as including programme issues (this factor combines two dimensions of “academic aspects” and “programme issues” in HEdPERF scale).

incorporating aspects that could make students feel comfortable and confident of the institutions (this factor is similar to that in Firdaus‟s (2006) HEdPERF scale).  Factor 4: access. This factor consists of items that mainly relate to such issues as

approachability, ease of contact, availability and convenience (this dimension is also similar to that in Firdaus‟s (2006) HEdPERF scale).

All in all, the 4 factors extracted, however, did not conform with the HEdPERF instrument, where five dimensions were identified, namely “non-academic aspects”, “academic aspects”, “reputation”, “access”, “programme issues”. Therefore, hypothesis 1 is rejected.

Table 5. Four factors retained

Factor Variables Non-academic aspects Academic - programme aspects Comfort Access SQ1 (Sincere interest in solving problem) 0.761 SQ2 (Caring and individualized attention) 0.59 SQ3 (Efficient/prompt

dealing with complaints) 0.789

SQ17 (Recreational facilities) 0.629 SQ18 (Library) 0.582 SQ19 (Minimal class sizes) 0.559 SQ20 (Easily employable graduates) 0.64 SQ21 (Equal treatment and respect) 0.638 SQ22 (Fair amount of freedom) 0.531 SQ23 (Confidentiality of information) 0.822 SQ24 (Easily contacted by telephone/Email) 0.578 SQ25 (Student's Union) 0.512 SQ26 (Feedback for improvement) 0.453 SQ27 (Service delivery procedures) 0.751 SQ28 (Variety of programmes/specializatio ns) 0.72 SQ29 (Flexible syllabus and structure) 0.67 4.3.1.3 Reliability analysis

The reliability of the dimensions should be examined after the identification of factors. The rationale of internal consistency is that the individual items should all be measurig the same construct and thus be highly intercorrelated (Hair et al, 2006). Becasuse no single item is a perfect measure of a concept, we must rely on a series of diagnostic measures to assess internal consistency. In this study, two internal consistency estimates of reliability, namely, Cronbach alpha and split-half coefficient were computed for the four service quality constructs. According to Hair et al (2006), an alpha value of 0.60 and above is considered to be the criteria for demonstrating internal consistency. For the split-half coefficient, the split yielding the highest correlation ordinarily gives the most nearly comparable halves (Wagner et al., 1986). The values of both Cronbach alpha and split-half coefficient for all four constructs are shown in Table 6. All the values meet the required criteria, thereby proving that the four dimensions are internally consistent and have satisfactory reliability values. Table 6. Reliability for the four service quality constructs

Dimensions Cronbach alpha split-half coefficient

Academic-programme

aspects 0.896 0.894

Comfort 0.698 0.654

Access 0.782 0.801

4.3.2 The relationship between service quality and satisfaction

4.3.2.1 The associations between four dimensions of service quality and satisfaction

To investigate the relationship between the dimensions of service quality (non-academic aspects, academic-programme aspects, comfort, access) and satisfaction, which is hypothesis 2a, I included four factors as independent variables and satisfaction as dependent variable in the regression. 3

The result of regression analysis, shown in Table 7, reveals that three out of four dimensions of service quality have significant relationship with satisfaction. The R square of this model is 0.494, which indicates that the model explained 49.4% of the original variability. The significant F-ratio ( F = 13.912, P = 0.000) indicates that the results of the regression model could hardly have occurred by chance. Thus, the goodness-of-fit of the model is satisfactory.

According to Table 7, three of four dimensions “non-academic aspects” (Beta = 0.309, p < 0.05), “academic-programme aspects” (Beta = 0.200, p < 0.05), “access” (Beta = 0.202, p < 0.05) are the determinant variables in predicting the overall satisfaction. In contradiction to what was expected, no significant association has been found between the dimension of “comfort” ( p > 0.05) and satisfaction. Therefore, H2a, stating that there are positive associations between all service quality dimensions and satisfaction can be rejected.

In addition, based on Table 7, the variable of “non-academic aspects” was the most important determinant of (students‟) overall satisfaction; it had the highest standardized coefficient value (Beta = 0.309). “Access” (Beta = 0.202), “adademic-programme aspects” ( Beta = 0.200), followed, in descending order of importance.

Table 7. Regression: SQ dimensions on satisfaction

Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta (Constant) 1.618 0.669 2.418 0.017** Nonacademic _{0.382 0.097} 0.309 3.945 0.000** Academic-programme 0.229 0.088 0.200 2.607 0.010** 3

Comfort 0.173 0.110 0.124 -1.575 0.117

Access 0.306 0.119 0.202 2.579 0.011**

Model F=13.912, P=0.000, significant at the 0.01 level R square= 0.494, ** p < 0.05

4.3.2.2 The relative importantce of technical aspects and functional aspects on satisfaction

To test H2b, stating that technical quality is more related with satisfaction than functional dimension,another regression analysis was conducted. The mean scores for technical quality including “non-academic” of 7 items, “academic-programme” of 9 items, and functional quality including “comfort” of 6 items, “access” of 7 items were calculated as predictor variables in the regression.

Table 8 tells about the results of the regression analysis. The R square of this model is 0.439, which is deemed to have sufficiently explanatory power.

According to Table 8, technical quality had a significant influence on satisfaction (Beta = 0.396, p < 0.05). No significant relationship has been found between functional quality (Beta = 0.066, p >0.05) and satisfaction. In other words, technical quality is more associated with satisfaction than functional dimension. Thus, H2b was supported.

Table 8. Regression: the effects of technical quality and functional quality on satisfaction Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta (Constant) 1.912 0.68 2.812 0.005** Technical 0.59 0.126 0.396 4.678 0.000** Functional 0.118 0.151 0.066 0.779 0.437