Mediating Influence of Satisfaction on Reputation, Relationship Benefits and Student Loyalty in Private Higher Education Institutions in Malaysia

Malaysian private higher education institutions are currently operating in a competitive education market. Their profitability and viability are being compromised by uncertain levels of student loyalty. Therefore, it is imperative for managers of private higher education institutions to be aware of the drivers of their students’ loyalty. This study aims to investigate the significant effects of reputation (university management, academic and media reputation) and relationship benefits on student loyalty in private higher education institutions in Malaysia and to determine the mediating effects of satisfaction on the aforementioned relationships. The methodology utilises primary data obtained from questionnaire administered to a sample of 400 students from several private higher education institutions using stratified random probability sampling. Multiple regression technique and mediation analyses were employed to analyse the data via SPSS statistical package. The results revealed that reputation and relationship benefits had significant effects on student loyalty. Empirical evidence concludes that university management and relationship benefits have significant effects on students’ loyalty, whilst academic and media reputation did not. Satisfaction was found to mediate the relationship between the independent variables and student loyalty. The study concludes with the discussion and implications, as well as limitations and suggestions for future research.

The Malaysian private higher education industry is currently operating under a highly competitive environment. Prior to 1996, before the passing of the Private Higher Educational Institutions Act. (Act 555), there were only nine public universities and no private universities, As at November 2015, the latest statistics indicate that there are now 20 public universities, 111 private universities and university colleges, 402 private colleges, 34 polytechnics and 94 community colleges with an estimated student population of 1.2 million (Tapsir, 2016).
The current scenario indicates that the profitability and viability of private higher education institutions (HEIs) in Malaysia are being compromised by the competitive market. Many private HEIs have been forced to shut down, with notable examples being reported in the press such as the closure of Allianz University College of Medical Sciences (Samy et al., 2014) and UNITAR's architecture faculty (Landau, 2017). Thus to survive and prosper, private HEIs have to find some form of competitive advantage to make themselves stand out from their competitors (Shah and Sid, 2013).
Loyalty is an essential component of an organization's success or failure (Aritonang, 2014). For education institutions, student loyalty is important for the many benefits that it brings, such as increased retention, re-patronage and positive word of mouth (Thomas, 2011). Many private HEIs are now realising the importance of student loyalty to their continued survival and are now turning their sights towards maintaining and increasing their students' loyalty levels.
As private HEIs are competing for loyalty using different marketing strategies, this study proposes that loyalty could be influenced by reputation and relationship benefits and mediated by satisfaction.

Student Loyalty
Loyalty is a very important issue for private higher education institutions as their major source of income comes from tuition fees. Having loyal students would provide a steady and solid stream of income for the running and profitability of the higher education institution. Besides this, student loyalty can also contribute towards positive alumni involvement and contribution. After graduation, students who are loyal may carry on supporting his/her education institution through various ways, such as financial support (monetary donations), collaborative partnerships (job placements, internships, giving talks) and positive word of mouth recommendation to potential, present and past students (Hennig-Thurau et al., 2001).
In this study, student loyalty is defined as loyalty of a student during and after his/her time at an education institution. It is usually demonstrated by re-patronage, positive word of mouth and supportive and committed behaviour towards the education institution (Kunanusorn and Puttawong, 2015).
A number of studies have found that reputation has a significant positive effect on student loyalty. In the seminal study conducted by Nguyen and LeBlanc (2001) in Canada, reputation was empirically confirmed as a predictor of student loyalty. According to Fares et al. (2013) student loyalty was positively and significantly affected by reputation of the higher education institution. Similarly, Sung andYang (2009) andWei (2012) concur when they found that the higher education institution's reputation had a significant and positive effect on student loyalty.
Existing research indicate that relationship benefits has a significant positive effect on loyalty. The study conducted by Adidam et al. (2004) on three Midwestern universities in America found that relationship benefits positively affected student loyalty. Wong and Wong (2012) and Holford and White (1997) found similar results by concluding that relationship benefits significantly influenced student loyalty. Gwinner et al. (1998) found relationship benefits had significant effects on customer loyalty in the service industry. Based on these findings, this study constructed the following hypotheses: H1: Reputation has a significant and positive effect on student loyalty. H2: Relationship benefits has a significant and positive effect on student loyalty.

The Mediating Role of Satisfaction
Satisfaction has been shown to be a strong predictor of student loyalty (Ali et al., 2016). In addition to its direct effect on student loyalty, satisfaction has exhibited a positive mediating effect on the relationship between reputation and relationship benefits in the higher education context. Barusman (2014) conducted a study on five private universities in Indonesia and found that reputation has an indirect effect on student loyalty via satisfaction. A similar mediating relationship was confirmed by Thomas (2011) in several universities in India. Satisfaction has also been shown to mediate the relationship between relationship benefits and loyalty. A study in the banking industry by Dimitriadis (2010) found that satisfaction has a mediating effect on relationship benefits and loyalty. Based on these findings, this study constructed the following hypotheses: H3: Satisfaction mediates the relationship between reputation and student loyalty. H4: Satisfaction mediates the relationship between relationship benefits and student loyalty.

Research design
This study utilizes a quantitative cross-sectional research design, which allows the researcher to integrate loyalty literature and the actual surveys as a means by which to collect accurate, less biased and high quality data (Sekaran, 2005).
This study was conducted in the Malaysian education industry and focused on private HEIs. The unit of analysis was individuals and comprised of students enrolled with private HEIs in Malaysia.

Measures
Measurement of the variables in this study was based on 27 items representing reputation, relationship benefits, and satisfaction and student loyalty. All the variables were measured using a 5 point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).
Reputation was measured by nine items which were adopted from Sung and Yang (2009). Six items for relationship benefits were adopted from Wong and Wong (2012). Satisfaction was measured by six items which were adopted from Ali et al. (2016). And lastly, student loyalty was measured by six items which were adapted from Nguyen and LeBlanc (2001).

Sample and Data Collection Procedures
The sampling design for this study was the proportionate stratified random sampling design. The sampling frame was obtained from the Ministry of Higher Education and consisted of 439 private HEIs with a student population of 580,928. These private HEIs were stratified according to their statuses: University, University Colleges, Foreign University Branch campus and College status and placed alphabetically in the respective strata. A private HEI or HEIs were selected within each strata randomly. Approval was obtained from the selected private HEIs to visit their campus to conduct a survey. The respondents for this study were students enrolled at the private HEIs at the time of survey.
The data for this study was collected via a personally administered questionnaire by the researcher. Appointments were made with the respective private HEIs to conduct the survey at their convenience. The researcher personally visited the campuses to conduct the survey. As the required samples size was 400, the similar number of questionnaires were handed out and collected back by the researcher upon respondents' completion. All questionnaires were checked for completeness and usability on the spot by the researcher.

Data Analysis
Various analyses were conducted on the data obtained using Statistical Package for Social Sciences (SPSS). A combination of descriptive and inferential analyses was employed in this study. The data was tested to ensure that it was suitable for subsequent statistical analyses used to test the hypotheses. The goodness of measures analyses was conducted prior to statistical analyses to ensure that they met the requirement in terms of reliability and validity (Hair et al., 2006).

Profiles of the Respondents
The demographic information includes the following characteristic of the respondents: gender, nationality, race, age and years studying at current education institution. The demographic information is presented in Table 1, based on frequency distributions and percentages. From the 400 respondents in this study, 156 (39%) were male and 244 (61%) were female. In terms of nationality, the majority of respondents were Malaysians, 380 (95%). Majority of the respondents were Chinese, 243 (60.8%), followed by Indians, 95 (23.8%), Malays, 41(10.3%) and others, 21 (5.3%). In terms of age group, majority of the respondents, 216 (54%) were aged between 17 -20 years. Lastly, in terms of years of studying at current education institution, slightly more than half of the respondents, 204 (51%) students reported having studied at their current education institution between 1 -2 years.

Factor Analysis
Factor analysis was performed for each category of variables (independent, dependent and mediating), using principal component analysis with varimax rotation to validate the construct validity of the measurements used. The items were selected if their factor loadings were greater than or equal to .50 and cross loadings with less than .35 (Hair et al., 2006).
The factor analysis for reputation produced two factors with eigenvalues greater than 1 contributing to 62.83% to item variance. The factors and their corresponding items selected were then grouped and renamed accordingly. Factor 1 was named university management and Factor 2 was named academic and media reputation. For the rest of the variables, namely relationship benefits, satisfaction and student loyalty, all three clearly loaded on one factor. Relationship benefits ((Eigen value = 2.72; % variance explained = 54.47%), satisfaction (Eigen value = 3.91, % variance explained = 65.20%) and student loyalty (Eigen value = 3.56, % variance explained = 59.35%). The results of the factor analysis are reported in Table 2.

Reliability Analysis
The reliability of the measures was tested after factor analysis was conducted. This was to ensure that the new measures identified for reputation in the previous factor analysis was reliable.
The results in Table 3 shows that all the variables have high reliability as they are all between the range of .78 to .89, with satisfaction at the highest reliability at .891 and relationship benefits at the lowest with .785.

Restatement of Hypotheses
Following the results from the factor analysis, reputation was not a uni-dimensional construct. Subsequently, the following hypotheses were re stated. H1: Reputation has a significant and positive effect on student loyalty. H1a: University management has a significant and positive effect on student loyalty. H1b: Academic and media reputation has a significant and positive effect on student loyalty. H3: Satisfaction mediates the relationship between reputation and student loyalty. H3a: Satisfaction mediates the relationship between university management and student loyalty. H3b: Satisfaction mediates the relationship between academic and media reputation and student loyalty.

Multiple Regression Analysis
To answer hypotheses 1 and 2 of this study, multiple regression analysis was conducted. Table 4 reports the results of the regression analysis and shows the significant effects of reputation (university management, academic and media reputation) and relationship benefits on student loyalty.
The results indicated that the F value of 99.779 is significant at confidence level of 0.05. The R² of .432 indicated that the two independent variables could explain 43.2% of the variance in student loyalty. The small values of the Variance Inflation Factor (VIF) ranging from 1.531 to 1.587 were well below 10.0, indicating no multicollinearity problems. There was also no auto correlation problem in the error terms as the Durbin-Watson value of 2.138 was within the acceptable range of 1.5 to 2.5 (Pallant, 2013).
In terms of effect size, Cohen J. W. (1998) suggested than a R² of .0196 can be considered small, a R² of .13 can be considered medium, and a R² of .25 can be considered large. Thus, based on Cohen J. (1988) conventions, a combined effect of this magnitude by the two independent variables in this study can be considered large.
From Table 4, university management and relationship benefits were found to be significant. Academic and media reputation was not significant. Therefore, the results indicated that Hypothesis 1 was partially supported, as Hypothesis H1a, university management was supported whilst H1b, academic and media reputation was not supported. Hypothesis 2 was supported, indicating that relationship benefits as a strong predictor of student loyalty.

Mediation Analysis
For hypothesis 3 and 4 of this study, PROCESS, a computational tool by Hayes (2012) was used to examine the direct and indirect pathways through which reputation (university management, academic and media reputation) and relationship benefits transmits its effect on student loyalty through an intermediary, satisfaction.
The relationship between university management and student loyalty was mediated by satisfaction. As Figure 1 illustrated, the standardized regression coefficient between university management and satisfaction was statistically significant, b = .711, p < .05, as was the standardized regression coefficient between satisfaction and student loyalty, b = .578, p < .05. The indirect effect was tested using bootstrap approach with 1000 samples and the effect size was .411 with a 95% CI = .3058, .5301). University management was associated with approximately .41 points higher student loyalty score when mediated by satisfaction. Therefore, Hypothesis 3a was supported. Fig-1.Mediating effect of satisfaction on the relationship between university management and student loyalty* p < 0.05 Figure 2 reports the results of the mediation analysis to answer Hypothesis 3b that states satisfaction mediates the effect of academic and media reputation on student loyalty. Results indicated that academic and media reputation was a significant predictor of satisfaction, b = .447, p < .05, and that satisfaction was a significant predicator of student loyalty, b = .709 p < .05. The indirect effect was tested using bootstrap approach with 1000 samples and the effect size was .320 with a 95% CI = .2388, .4095). Academic and media reputation was associated with approximately .32 points higher student loyalty score when mediated by satisfaction. As such, Hypothesis 3b was supported.

Fig-2.
Mediating effect of satisfaction on the relationship between academic and media reputation and student loyalty* p < 0.05 The relationship between relationship benefits and student loyalty was mediated by satisfaction. As Figure 3 illustrated, the standardized regression coefficient between relationship benefits and satisfaction was statistically significant, b = .620, p < .05, as was the standardized regression coefficient between satisfaction and student loyalty, b = .651, p < .05. The results indicated that there was a significant indirect effect of relationship benefits on student loyalty through satisfaction. The indirect effect was tested using bootstrap approach with 1000 samples and the effect size was .400 with a 95% CI = .3181, .5055). Relationship benefits was associated with approximately .40 points higher student loyalty score when mediated by satisfaction. Therefore, Hypothesis 4 was supported. Fig-3. Mediating effect of satisfaction on the relationship between relationship benefits and student loyalty* p < 0.05

Discussion and Implication
This study has identified that university management had significant and positive effect on student loyalty whilst academic and media reputation did not. Overall, the findings in this study is broadly consistent with the results of other studies conducted by Fares et al. (2013); Wei (2012) and Sung and Yang (2009) that determined reputation as a predictor of student loyalty. In the higher education context, this finding is consistent with the fact that higher education is a service and not a product (Walsh et al., 2009). When a customer is dealing with something intangible, the reputation of the service provider is of utmost importance. For students, pursuing an academic qualification is associated with large monetary cost and time spent. Most students spend at least three to four years with an education institution, with some at even longer durations. Parents of students' are willing to pay expensive tuition fees to ensure that their children obtain an academic qualification from a reputable education institution. Therefore, being perceived as being reliable, trust worthy and responsible are important criteria's for students (Awang and Jusoff, 2009).
Relationship benefits was found to have a positive and significant effect on student loyalty. This indicates that students who perceive that they are receiving benefits which are of value to them would be more loyal. This result is consistent with previous studies conducted by Gwinner et al. (1998) and Holford and White (1997) in various different consumer industries. Marketing managers of private HEIs in Malaysia must be aware of what benefits are important to their students and focus their marketing strategies on these benefits. As paying customers, students would expect the best benefits for money spent.
The mediating analyses revealed that satisfaction plays an important role in all the relationships between university management, academic and media reputation and relationship benefits with student loyalty. The management of private HEI's need to monitor the satisfaction levels of their students towards these determinants. If students' are dissatisfied, they may not become loyal to their education institution even if the determinants received is up to their expectations. Therefore, managers of private HEIs must understand their students' needs and react to their concerns in the quest to keep them satisfied.
The results of this study describe for the very first time the direct effect of relationship benefits on student loyalty as well as the mediating effects of satisfaction on the afore mentioned relationship in the higher education context in Malaysia. Other researchers which examined the effect of relationship benefits in the education context did not study the direct effect of relationship benefits and used other mediators such as commitment (Wong and Wong, 2012); (Adidam et al., 2004).

Limitations and Suggestions for Future Research
There are a few limitations in this study. The first limitation is related to the generalizability of the results. The findings from this study cannot be generalized to other industries as it was conducted specifically in the education industry. This study was conducted as a quantitative cross sectional study, thus this could limit the ability to infer causal relationships among the variables of this study.
Based on the limitations of the study, this study suggests a few recommendations for future research. Future research should look at other variables not included in this study such as service quality, image and shared values. As this study was conducted using the quantitative approach, future research could consider combining both the quantitative and qualitative research approach. This would provide more accurate information on the why and how of the complex loyalty formation process. Future research could also be conducted as a longitudinal study so that relationships between student loyalty and their determinants can then be more accurately revealed.

Conclusion
The level of competition amongst private HEIs in Malaysia is expected to increase in the future and it is critical for management of these institutions to determine suitable strategies to overcome this problem. The education institution's success or failure will be dependent on its ability to improve its students' loyalty levels.
This study concluded that university management was the most influential on student loyalty, followed by relationship benefits, whilst academic and media reputation had no effect. Satisfaction was found to mediate the relationships between all the independent variables and the dependent variable.
This study is timely and significant as it provides valuable information on the loyalty formation of students in private higher education institutions. A clearer understanding as to the relationships between reputation, relationship benefits, and satisfaction and student loyalty can help managers of these institutions to formulate accurate marketing strategies to aid them in enhancing their students' loyalty levels, which in turn generates more profits, thus ensuring their viability.