A study on improving employees ’ behaviour towards increasing students ’ loyalty : the mediating role of need understanding , service quality , and intimacy among students in Kenya

Article history: Received: March 23, 2018 Received in revised format: June 1, 2018 Accepted: June 15, 2018 Available online: June 15, 2018 During the past few years, there have been various studies on the relationship between academic behavior, teaching and student retention but not much has been on administration staff and students. This study examined improving employees’ behavioural factors towards increasing students' loyalty. The paper explores the direct mediation and indirect mediation of employee behavioural factors leading to build student loyalty. More specifically the focus is on the constructs of service quality (SQLTY), need understanding (NEEDUND), intimacy (INTIMACY), and student loyalty (STLOY). This study depended on a positivist research paradigm. In this study, over 800 structured questionnaires were administered for research subjects. However, 743 were captured for the final analysis. The analytical tool used in this study is the ADANCO 2.0.1 software and in terms of statistical processing, the PLS-SEM technique was utilized. The two main takeaways from the direct relationships are that INTIMACY is a strong predictor of STLOY, and NEEDUND is also a strong predictor of SQLTY. The indirect mediations of INTIMACY on SQLTY and STLOY, INTIMACY on NEEDUND and STLOY, and SQLTY on NEEDUND and INTIMACY were all not supported. The study, like any academic work, has limitations. Despite these limitations, this study offers theoretical as well as practical values for the research community and administrators of universities and higher education administrators as a whole. For administrators of the universities and the higher education, the study points out the critical needs for administrators to better understand students’ loyalty and behaviours of employees during the service encounter, which could lead to intimacy and eventually student loyalty. The present study is innovated by quantitatively examining how the above-mentioned behavioural factors of employees could lead to INTIMACY and STLOY.


Introduction
The subject of customer loyalty has been well researched in the service industry in general, and it receives huge interest among the academic research community as well as for managers of services (for more detail commentary, see du Toit & Burns, 2016;Ernst & Young [EY], 2017;McMullan & Gilmore, 2008;Werner & Kumar, 2002).There has been a call from researchers like Ruiz et al. (2016) for researchers to introduce fairly newly constructs in testing loyalty since constructs like trust and satisfaction have extensively been researched.Quite recently, researchers like de Waal et al. (2016) heed to this call by studying the behaviours of employees that predict intimacy which, in turns, leads to customer loyalty.They argue that for organizations to achieve their objectives, employees have to understand the importance of the customer to the organization since the organization ceases to exist if there are no customers.They identified behavioural factors such as (service quality (SQLTY), need understanding (NEEDUND), as some of the factors if display by employees towards customers could possibly lead to intimacy (INTIMACY) and subsequently lead to customer loyalty (for more commentary on this, see de Waal et al., 2016).In this study, the customer (herein refer to student) loyalty will be used and for more similar work (see, Helgesen & Nesset, 2007).Also, readers should keep in mind that in this study, student loyalty means the loyalty of the students during their course of study and after graduation from the university.Building on the recommendations of de Waal et al. (2016), this study filled the gap by quantitatively testing these behavioural factors mentioned earlier by examining the behavioural factors of university staff (need understanding (NEEDUND) and service quality (SQLTY) on students' intimacy (INTIMACY) and student loyalty (STLOY).The study further examined direct and indirect mediation between NEEDUND, SQLTY, STINTM, and STLOY.See, Fig. 1 for the various constructs and the relationship tested by the study.
There is an extant researched literature on customer loyalty in the educational sector in general (c.f.Sembiring, 2013;Kheiry et al., 2012;Thomas, 2011;Mohamad & Awang, 2009).What is missing in the customer loyalty research in the educational sector is the kind of behavioural attitude employees of educational institutions, especially universities which the study is focused, should demonstrate towards their students for them to be loyal.There are several studies on customer loyalty (for more details, see Al-Awadi, 2002;Chang et al., 2009;Kuo & Ye, 2009 ;Sembiring, 2013;Kheiry et al., 2012;de Waal et al., 2016;Ruiz, et al., 2016 ), we are yet to have a universal definition on the concept.However, for the purpose of this study, we adopted the definition of Reichheld (2003).Reichheld (2003, p.3) defined loyalty as the "willingness of someone, a customer, an employee, a friend to make an investment or personal sacrifice in [other/order,another] to strengthen a relationship".In line with this definition, the study argues that the behaviour of employees of universities towards students is crucial in building a strong relationship among students.Past researchers have identified intimacy as playing an important role and a dual one between the customer and the organization relationship building process (Treacy & Wiersma,1993).As cited in de Waal et al. (2016), customer intimacy is defined as "making customers feel good whenever they make contact with your company" (Ballou, 2006, p. 5, cited in IBM Global Services, 2006) or "tailoring and shaping products and services to fit the increasingly specific definition of the customer" (Bügel, 2010, p. 65).Empirically, past researchers ( for more details see, Fleming et al., 2005;Lloyd & Luk, 2011;Pitta et al., 2006, Lau, 2000;Xu & Van der Heijden, 2005;Yim et al., 2008) observed that having a good interaction with customers and employees of an organisation impact on customer intimacy.Typically, Yim et al. (2008) study suggest that good customer service from an employee leads customers to have intimacy with the organization which leads to customer loyalty.
In the service literature,, researchers have developed myriad models with scales for testing service quality in general.(cf.Parasuraman et al., 1988;Brogowicz et al., 1990;Cronin & Taylor, 1994;Berkley & Gupta, 1994;Spreng & Mackoy, 1996;Philip & Hazlett, 1997;Sweeney et al., 1997;Oh, 1999;Dabholkar et al., 2000;Frost & Kumar, 2000;Soteriou & Stavrinides, 2000;Broderick & Vachirapornpuk, 2002;Grönroos, 1984Grönroos, , 2000;;Zhu et al., 2002;Santos, 2003;Vijayvargy, 2014).Some examples of these models developed by these researchers are: the information technology (IT) alignment model, the service quality and satisfaction model, PCP (Pivotal, Core, Peripheral) attribute model, the retail service quality and perceived value model, the customer value and customer satisfaction model , the antecedents and mediator model, the internal service quality model, the internal service quality data envelope analysis model, the service quality model for internet banking, the service quality model for IT-related business, and the model of e-service quality.Among these models, the study observed that models such as SERVPERF, SERVQUAL, and SERVPERF have extensively been researched by scholars in the service industry.In the educational research on student loyalty, most of the abovementioned models have been relied on to study students loyalty in universities.However, scholars like Tinto (1975, 1993), and Hennig-Thurau and Klee (1997) developed models such as student drop-out behaviour model and relationship quality-based students loyalty ( RQSL) respectively to test student loyalty..The Tinto's model indicates that two main constructs: students commitment and integration are interconnected and initiate a dynamic process.The model is of the view that students commitment directly impact loyalty.However, the relationship between integration and loyalty is mediated by the student's commitment (for more detail commentary, see Hennig-Thurau et al., 2001).The RQSL model developed by Hennig-Thurau and Klee (1997) broaden the discussion on student loyalty by identifying three main cardinal principles of students perceived quality of teaching, trust in the employees, and students commitment to the university as a direct influence on students loyalty.
Additionally, this study observed that most of the theories in the service quality literature and especially the theories used in the examining student loyalty have not dwell much on the behaviours expected of employees towards students which would create the necessary intimacy which will lead to the loyalty of students towards universities.Based on earlier submission, the study, therefore, have identified this gap that has existed in the student loyalty research by relying on the work of de Waal et al. ( 2016) in examining the behavioural factors of university staff in Kenya which could lead to intimacy and loyalty of students.Currently, with the liberalization of private university education in Kenya and in most countries in Africa, universities are experiencing increasing competition withn the tertiary section of education.Also, coupled with the constant reduction of government funding support especially the public universities has put a huge burden on administrators of these universities.Now, universities in Kenya have to find innovative ways of increasing their students' numbers to enable them to meet their operations cost.This concern is perfectly presented by the Vice-Chancellor Prof Kanyari of one of Kenya's public university in their 2017 graduations.Prof Kanyari indicated that: "This year, JKUAT will be admitting 4,994 government-sponsored students in its September 2017 intake.However, with the ongoing reforms in the education sector, the University anticipates receiving a significantly reduced number of self-sponsored students this year.This will result in an income shortfall of close to Sh1 billion," Prof Kanyari.
That is why it is important that employees who are in daily contact with the students should behave in a manner that will build intimacy and loyalty among students.This is because loyalty, as pointed out in the literature, has various advantages such as customer retention, increased profitability and can be a significantly low costs driver (see Dawes & Swailes, 1999;Duncan & Elliot, 2002;Gupta et al., 2004;Lam & Burton, 2006;Makanyeza & Chikazhe, 2017;Reichheld & Sasser, 1990;Tantakasem & Lee, 2007).This research contributes to academic scholarship on student loyalty in the university sector by testing constructs which are fairly new assumptions in a context which remains underexposed in the student loyalty.Secondly, our study helps inform managers of universities in Kenya and other similar developing countries to spend resources in empowering their employees to behave towards their students in a manner that would lead to intimacy and then loyalty.As hinted, this paper is mainly concern about examining the mediation role of need understanding (NEEDUND), service quality (SQLTY), and student intimacy (INTIMACY) on student loyalty (STLOY).The main objective of the study was divided into the following sub-objective, namely 1) to determined whether NEEDUND will be positively related to INTIMACY, 2) to determine whether NEEDUND will be positively related to SQLTY, 3) to determine whether SQLTY will be positively related to INTIMACY, 4) to determine whether INTIMACY will be positively related to STLOY, 5) to determine whether INTIMACY will positively mediate the relationship between SQLTY and STLOY, 6) INTIMACY will positively mediate the relationship between NEEDUND and STLOY, 7) to determine the indirect effect of NEEDUND on STLOY will be mediated sequentially be SQLTY and INTIMACY.One interesting thing about this study is its departure from past researches on students' loyalty.For the rest of the paper the main issues addressed were a) literature review, b) methodology, c) findings and d) discussion of findings and conclusions

Student Loyalty: An overview
The construct customer loyalty, amongst several definitions in the literature, is stated to be "the strength of the relationship between an individual's relative attitude and repeat patronage" (see Dick & Basu, 1994:99).From a practitioner's point of view, loyalty can be captured broadly as "the willingness of someone-a customer, an employee, a friend-to make an investment or personal sacrifice in other to strengthen a relationship" (see Reichheld, 2003).The literature on students' loyalty has also received fairly some research from the academic community (for example, see Rojas-Méndez et al., 2009;Lee and Anantharaman, 2015).However, the extensiveness of research in the area of student loyalty is not the same as compared with loyalty in general.Currently, we are yet to see a generally acceptable definition from the scholars of student loyalty.As cited in the work of Lee and Anantharaman (2015 p.106) student loyalty is defined as "psychological attachment to their universities founded upon their feelings of identification and affiliation (Verhoef et al., 2002) and it is manifested as behavioural and attitudinal commitment toward an institution" (Tinto, 1993).The issue of student loyalty is a complex one and remains far from completely understood.It is not surprising that previous research on students' loyalty listed factors such as satisfaction, trust, service quality, and commitment as predicting students loyalty (for more details, see Rojas-Méndez et al., 2009;Lee & Anantharaman, 2015).However, in the customer loyalty literature, it has been identified by de Waal et al. (2016) that behavioural factors such service quality, need understanding by employees will lead to intimacy and subsequently leads to customer loyalty.Though the study of de Waal et al. ( 2016) was applied to one media organization in the Netherlands, the trust of this study is to look at how quality services, need understanding behaviours exhibited by university staffs towards students could lead to students intimacy and students loyalty in universities in Kenya.
Researchers like Sirdeshmukh et al. (2002), Yim et al. (2008), Dixon et al. (2010), Bügel (2010), Mechinda and Patterson (2011), Lloyd andLuk (2011), andde Waal et al. (2016) have studied several behavioural factors that employees of services should exhibit towards customers for loyalty to manifest.This is because employees are constantly dealing with customers on a daily basis and any negative attitude they exhibit could have an undesirable consequent on the customer.That is why it is important employees exhibit behaviours that will increase intimacy and loyalty among their customers.For example, in the university setting, where our study is focused, students are constantly dealing with both academic and non-academic staff.So, any adverse behaviour from both category of staff could have a negative effect on the students and the institution as a whole.In contrast, if the employees are able to provide the needed services and understand the needs of the students this could lead to intimacy and loyalty of students.

Proposed Conceptual Framework and Research Hypotheses
The proposed framework for the study is illustrated in Fig. 1 which identifies the various constructs examined on intimacy (INTIMACY) and student loyalty (STLOY).To start with, the study examined the direct mediations between service qualities (SQLTY), need understanding (NEEDUND), intimacy (INTIMACY) and student loyalty (STLOY).Secondly, the study also examined the indirect mediations between SQLTY, NEEDUND, INTIMACY, and STLOY.

Extant Model 1: Direct relations between need understanding (NEEDUND), intimacy (INTI-MACY) and service quality (SQLTY)
Researchers like Yim et al. (2008), Winsted (2000), Mechinda and Patterson (2011), Lloyd and Luk (2011) have conducted studies on behavioural factors and indicated that during the service interaction, employees should bring about a sense of comfort for the customer during the service experience.For example, Lloyd and Luk (2011) empirically studied the behaviour of employees necessary to elicit customer comfort about fashion apparel retailing and casual dining restaurants.The findings show that a favourable interaction behaviour of employees creates a sense of comfort for the customer.According to Lloyd and Luk (2011), a favourable comfort leads to overall quality of service.In a similar study by Mechinda and Patterson (2011) in examining the determinants of customer-oriented behaviour (COB) of service providers within the healthcare shows that service employees should display customer-oriented behaviors towards the customers for customer satisfaction to manifest.Also, de Waal et al. ( 2016) conducted a study in the media sector in the Netherlands and the results show that need understanding of employees and service quality exhibited by employees leads to intimacy between the customer and the service employees which eventually leads to customer loyalty.Since, previous researchers (see, Mechinda & Patterson, 2011;Lloyd & Luk, 2011;de Waal et al., 2016) have shown in different industry contexts that employees' behaviour towards customers' impact on customer loyalty is, therefore, seminal.This study is also carried out in a high contact service setting like the universities since different result could emerge.Based on past researches, this study argues that when employees of universities demonstrate the NEEDUND of students, this could lead to INTIMACY.Also, if employees are able to understand the needs of students and provide the necessary assistance this could also enhance service quality (SQLTY).The study, therefore, put forward the following hypotheses to be tested.

Hla: NEEDUND will be positively related to INTIMACY
H1b: NEEDUND will be positively related to SQLTY

Extant Model 2: Direct relations between service quality (SQLTY), intimacy (INTIMACY) and student loyalty (STLOY)
In the student loyalty literature researchers (see, Annamdevula & Bellamkonda, 2016;Sembiring, 2013;Rojas-Méndez et al., 2009;Mohamad & Awang, 2009) have examined quality service on student loyalty.For example, a study conducted by Annamdevula and Bellamkonda, (2016) which examined 918 students from three oldest universities in the state of Andhra Pradesh in India, shows that SQLTY is a predictor of STLOY.When students are satisfied with a service by the university, this influence the students to be loyal to the university (Annamdevula & Bellamkonda, 2016).Conversely, a study conducted by Rojas-Méndez et al., (2009) of a sample of 752 students (freshman, sophomore, junior and senior) of Chilean public university for two years shows that SQLTY has no significant effect on STLOY.However, de Waal et al. ( 2016) study shows that SQLTY has a positive relationship between INTIMACY and STLOY.They indicated that when employees provide customers with excellent customer service during the service encounter this could lead to intimacy and then eventually customer loyalty.The study is therefore of the view that when employees of universities provide students with excellent services this could lead to INTIMACY, and then leads to STLOY.The following hypotheses are formulated H1c: SQLTY will be positively related to INTIMACY H1d: INTIMACY will be positively related to STLOY

Extant Model 3: INTIMACY will indirectly mediate the relationship between SQLTY and STLOY
The study succinctly discusses the mediating role of INTIMACY on the connection between SQLTY and STLOY.Customer intimacy is defined as "making customers feel good whenever they make contact with your company" (Ballou, 2006, p. 5, cited in IBM Global Services, 2006) or, alternatively, "tailoring and shaping products and services to fit the increasingly specific definition of the customer" (Bügel, 2010, p. 65).Previous empirical work has shown that the interaction between the service employee and the customer is the most important determinant of customer intimacy (Fleming et al., 2005;Lloyd & Luk, 2011;Pitta et al., 2006).In the academic literature on student loyalty, researchers like Annamdevula and Bellamkonda (2016) have studied the mediating role of student satisfaction on service quality and loyalty.However, there is no evidence from academic surveys confirming the mediating role of student intimacy on SQLTY and STLOY.Albeit, de Waal et al. (2016) in the service context, in general, found service quality influencing intimacy and intimacy leading to customer loyalty.The study reasons that for STLOY to occur it must first of all start with SQLTY which will lead to INTIMACY and then to STLOY.So, it is paramount for employees to offer quality services to students to enhance an intimacy between them and then this could lead to student loyalty.This study joins de Waal et al. to show that in the educational sector INTIMACY will indirectly mediate the relationship between SQLTY and STLOY.The study, therefore, proposed the following hypothesis.
H2a: STINTM will positively mediate the relationship between SQLTY and STLOY

Extant Model 4: INTIMACY will indirectly mediate the relationship between NEEDUND and STLOY
Concerning model 4, this study is of the opinion that need understanding on students' loyalty will be sequentially mediated by intimacy.Past researchers (de Waal et al., 2016;Mechinda & Patterson, 2011;Lloyd & Luk, 2011;Yim et al., 2008) suggest that need understanding of customers by employees leads to intimacy and subsequently leads to customer loyalty.The study, therefore, based on these past empirical findings by senior researchers argues that INTIMACY will sequentially mediate the relationship between NEEDUND and STLOY.Hence the following hypothesis is proposed.

Survey and Data
This study used cross-sectional survey research design.The study employed convenient sampling techniques to accrue the required sample size for the study, which is in line with a similar study conducted by (Boateng & Okoe, 2015).Convenient sampling approach facilitates data collection easily and is focused on those who want to part of the process of filling the questionnaire.This study was conducted among students at the University of Nairobi, Kenya.University of Nairobi is the largest public university in Kenya, located at capital city of Nairobi.The study participants were randomly selected from the University of Nairobi, main campus which is in the capital city Nairobi.A team of five research assistants were trained on questionnaire administering and study participant recruitment for the purposes of this survey.
The study distributed a little more than 800 questionnaires which were self-administered with the assistant of five trained undergraduate students of the University of Nairobi.The survey exercise lasted for nearly four months in 2017.At the end of the survey exercise, the study captured 743 responses that were used for the final analysis.Some of the questionnaires were discarded because they were either incomplete or were not filled out properly by respondents and hence could not be used for the analysis.The study maintained high levels of research ethics procedure.Oral informed consent was sought from the participants before completing the survey questionnaire.All participants were free to withdraw at any time of completing the questionnaire.In this study, apart from the demographic characteristics, all the measures used in this study came from previous studies.The study presents the demographic characteristics of respondents who took part in the investigation.The distribution of characteristics of those surveyed (see Table 1) shows most of them were young ages between 18 and 25 representing 53.2% of the sample.Out of the 743 respondents, 57.5 % were male, and 42.5 % were female.The majority were in their third (47.6%) and fourth (22.1%) years at the university.Please refer to Table 1 for more details.

Measures
Except for demographic measures, all the (latent) measures used in this study come from previous studies.Briefly, the measures of Service Quality (SQLTY) had nine measurements items, Need Understanding (NEEDUND) had seven measurement items, Student Loyalty (STLOY) also had seven measurement items and were all taken from these researchers (see, de Waal et.al., 2016;Yim et al., 2008;, Sirdeshmukh et al., 2002;Dixon et al., 2010;Winsted, 2000;Dabholkar et al., 2000;Bügel, 2010), while Student Intimacy ( INTIMACY) had five measurement items taken from these researchers (Yim et al., 2008;, de Waal et al., 2016).The measurement items were all measured on a 5-point Likert Scale.Readers should note that some of the measurement items were slightly modified to suit the objective of this study.In case readers need the exact wordings of the multi-item measures, they may choose to consult the authors privately or refer to the studies of the researchers mentioned.Descriptive statistics of each of the measures are presented in Table 2.

Choice of Analytical Tool
The data was analysed with the help of ADANCO 2.0.1 software (see, Hair et al., 2012Hair et al., , 2017;;Henseler & Djisktra, 2015) and was based on PLS-SEM over covariance-based SEM.ADANCO, like Smart PLS, is a very user-friendly and powerful tool for testing complex relations.Senior researchers like Henseler and Djisktra have advocated for the use of PLS-SEM in testing complex and several relationships in empirical studies like this study.Hence, the choice of using PLS-SEM for this study and moreover, in this formative stage of consumer research (student loyalty) in Kenya context, PLS-SEM is appropriate techniques to use.Furthermore, the study is focused in-sample prediction rather than on theory confirmation.Therefore, the study is convinced just like senior scholars like Hair and colleagues that PLS-SEM will be a more appropriate technique.

Test for Common Method Bias (CMB)
This research recognised the issue of CMB.Any serious concerns about CMB has been adequately addressed in this analysis.Diligent steps were taken during the instrument design and data collection.The respondents were assured of their anonymity and that there were no correct or incorrect answers.Oral informed consent was sought from the participants before completing the survey questionnaire.The fact that all participants were free to withdraw at any time of completing the questionnaire sought to address the concerns of CMB (for more commentary, see MacKenzie & Podsakoff, 2012).Spector (2006) indicated that the notion of CMB has been blown out of proportion in the methodology literature.Similarly, other researchers (Doty & Glick, 1998) suggests that though CMB is a problematic issue in studies that rely purely on self-report data, just like this work, (yet) it does necessarily invalidate empirical results.
Succinctly, the procedure and steps taken in this analysis have partially eliminated any strong concerns about CMB (also see the recommendations by Conway and Lance, 2010).

Measurement Model Assessment
In this study, ADANCO 2.0.1 was used performed construct reliability based on Dijkstra-Henseler's rho, convergent validity, and AVE.Concerning discriminant validity, the study used the more popular Fornell-Lacker criterion (see Fornell & Lacker, 1981), as well as Heterotrait-monotrait (HTMT) ratio of correlations (for more details, see Henseler et al., 2015Henseler et al., , 2016)).Examining the cross-loading shows that cross-loadings also lends numerical support to the discriminant validity of research constructs.The overall results show that the constructs reliability scores differ significantly among one another.The highest HTMT score was 0.80 (i.e., correlation between SQLTY and NEEDUND), which is also below the (suggested) maximum threshold of 0.9.Also, the HTMT inference of the correlation between SQLTY and NEEDUND was also the highest with a value of 0.77 which also happens to be significantly below the cutoff value of 1.The study also performed overall goodness-of-fit (GoF) index using the standardized root mean square residual (SRMR -estimated model) which was found to be 0.04.Readers should bear in mind that there is an ongoing debate about the relevance of reporting fit statistics in PLS-SEM literature.Researchers like (Henseler et al., 2016)

Structural Model Assessment
Though there is a current debate among academic scholars about the use of GoF measures like SRMR statistic in PLS-SEM study, this study presents the SRMR value of saturated (ie, structural) model to be approximately 0.04 which is also close to the SRMR of the estimated model.In PLS-SEM literature, it is observed that SRMR value below 0.08 displays a good fit for the data.From the result of the (direct) path estimates, there is an empirical support for H1d and H1b (for more information on this, see Table 8 for more information).As presented in Table 8, the indirect mediation test offers no support for the mediating effects ranging from H2a, H2b, and H3a.It is important to note that nonparametric bootstrapping technique was employed to test the suggested mediating effects.As earlier on indicated, all the results have been processed using ADANCO with bootstrapped sample values of 999.The study also relied on R-square statistics (cf.Hair et al., 2012) as far as the theoretical model explanatory power was concern.The R-squared value of the main construct, STLOY, was found to be 0.5 suggesting model explains about 50 percent variation in this construct.From the ADANCO software comes another useful statistics like Cohen's f 2 , a measure for effect size, are also presented in the table.
In the academic scholarship, researchers like Henseler et al. (2009) have indicated the following range of effect sizes: 0.02 (small effect), 0.15 (midrange effect), and 0.35 (large effect) by citing the work of senior researcher Cohen (Tables 5 and 6 present more details on this).The study now concentrates on the discussion of the empirical results.

Discussion of Results
The study was based on examining behaviours that university employees should demonstrate towards students which will lead to intimacy and then eventually students' loyalty among students in Kenya.The analysis was based on the interrelations among behaviours toward students.The premise of this study is the analysis of the interrelations among, NEEDUND, SQLTY, INTIMACY and STLOY.The study examined the direct relations, namely H1a, H1b, H1c, and H1d.It further examined the indirect mediations as indicated in H2a, H2b and H2c (for more details refer to Table 8) and then the empirical findings were discussed.
The findings from the direct relationships reveal that INTIMACY is a strong predictor of STLOY and NEEDUND is also a strong predictor of SQLTY.These results support H1b and H1d.In the service industry in general, previous studies conducted by researchers like Yim et al. (2008) andde Waal et al. (2016) indicate that INTIMACY of customers with employees predicts customer loyalty.The study findings of INTIMACY are strong predictor of STLOY and support earlier research work done by researchers (cf.Yim et al., 2008, de Waal et al., 2016).Meanwhile, in the student loyalty literature, constructs such as trust, satisfaction, image, service quality and commitment have been empirically tested on student loyalty by researchers like Annamdevula andBellamkonda (2016), Rojas-Mendez et al. (2009), Helgesen andNesset (2007), Hennig-Thurau et al. (2001).It is important to state that at the moment, empirical studies of behavioural attitude of employees of educational institutions especially universities should demonstrate towards building intimacy and loyalty among students in the context of developing country like Kenya is lacking; although the findings support INTIMACY as a strong predictor of STLOY.This means that when employees of universities in Kenya and other parts of the world demonstrate or exhibit good attitude towards students during their interaction and encounter students will be intimated with them and this will lead to student loyalty.The loyalty literature as observed by researchers (de Waal et al., 2016;Mechinda & Patterson, 2011;Bügel, 2010) have identified that employees demonstrate affectionate ties and good interactions during the service encounter could lead to INTIMACY and subsequently could lead to customer loyalty.In the same vein, if employees of universities demonstrate good behavioural attitude towards students during the service encounter, this could lead to INTIMACY and STLOY.With the current competition among universities for students' enrolments, it is important that administrators pay more attention to equipping staff with excellent customer service to be able to demonstrate behavioural attitude that would lead to intimacy at all time when the service encounters with university employees.With respect to H1d, empirical estimates reveal that NEEDUND significantly correlates with SQLTY.Researchers of customer loyalty in general (see, Bowen & Schneider, 1988) indicated that providing excellent service as well as creating the necessary conducive environment in an effort to provide excellent services is an assurance that customers have received quality services.Frazer Winsted (2000), Yim et al. (2008), Lloyd and Luk (2011), Mechinda andPatterson (2011) andde Waal et al. (2016 ) have observed need understanding (NEEDUND) behaviours like how the employee meet customers' needs and the interests of employee in customers' needs are as some of the behavioural attitudes that demonstrates need understanding.It is not surprising that NEEDUND actually correlates with SQLTY.This is because if employees of universities are able to display these behavioural attitude in terms of understanding the needs of students, and are able to meet these needs then it will be prudent to argue that SQLTY has been achieved.So it is important that staff of universities are able to understand what students need during the service interaction and encounter so that they could display these behaviours in meeting the needs of students.H1a and H1c were not supported by the findings of this research.This study hypothesized that NEEDUND will lead to INTIMACY (H1a).However, our empirical estimates show otherwise.This is not in support of an earlier findings by de Waal et al. (2016).Hence, the findings from this study suggest that employees may display behaviours to students during their service encounter which will lead to NEEDUND but does not necessary lead to INTIMACY.In addition, it could also mean the context and the industry in which the study is carried out will produce different results.With respect to H1c, the study also examined the direct relationship between SQLTY and INTIMACY.The empirical evidence shows that there is no relationship between SQLTY and INTI-MACY when it comes to loyalty among students in Kenya.This is also in sharp contrasts with the findings of de Waal et al. ( 2016) which confirm that when employees exhibit behaviours which display quality service provision during the service encounter could lead to INTIMACY.
Apart from the direct relationships examined, the study also examined the indirect mediations of SQLTY, NEEDUND, INTIMACY, and STLOY.The first mediation examination was to ascertain whether intimacy will mediate the relationship between SQLTY and STLOY (SQLTY → INTIMACY → STLOY).The empirical estimates did not support the claim of intimacy mediating the relationship between service quality and student loyalty.The study also examined the mediation between need understanding, intimacy, and loyalty (NEEDUND → INTIMACY → STLOY).The indirect relationship between need understanding (NEEDUND) through INTIMACY to STLOY was also not confirmed.Finally, on the indirect mediation analysis, the findings revealed that SQLTY will mediate the relationship between NEEDUND and INTIMACY (NEEDUND → SQLTY → INTIMACY).By way of summarizing, the analysis makes two main contributions to the increasing student loyalty (STLOY) research.First, INTIMACY is a strong predictor of STLOY.Second is the fact that need understanding (NEEDUND) is also a strong predictor of student service quality (SQLTY).For the rest of the direct and indirect relationships i.e.H1a, H1c, and H2a-c, because this is an entirely new relationship within the student loyalty scholarship, future research would be required to subject it to empirical testing in different contexts.Conclusively, this study has contributed significantly to the student loyalty and loyalty in general as we extend the frontiers of previous academic thinking by demonstrating some of the behavioural attitudes that employees must demonstrate during the service encounter which could lead to student loyalty and loyalty as a whole.

Managerial implications of the study
For administrators of universities and higher educational institutions i, this research emphasized the critical need for administrators to better understand students loyalty and behaviours of employees.For instance, INTIMACY was a strong predictor of STLOY and provides a valuable guidance to administrators to devote resource and time to train their staff on customer service.This will enable them display behaviours which will lead to INTIMACY and STLOY any time they encounter students in their service delivery during their encounter with students.This study will also help administrators of universities not only in Kenya but elsewhere in the world to develop with the collaboration of all stakeholders a customer service policy to be adhered to by all employees.This will eventually enhance the behaviours of employees toward students during the service delivery.Let also remind ourselves that the complex mechanism involving SQLTY, NEEDUND, INTIMACY, and STLOY involves strategic thinking, patience and investments.This is because the payoff is not immediate; it takes time.Given the increasing competition in the higher education landscape in Kenya and other parts of the world today, this study will enable employees and the university as a whole to be consistent in delivery student loyalty.Finally, the study hopes administrators of universities and higher education, in general, will be able to develop a proper understanding of the role that INTIMACY, as well as strong STLOY, plays when it comes to the implementation of STLOY strategies.

Conclusion, Limitations and Future Research
This study has extended the frontier of knowledge and understanding in the are of employees behaviour towards students loyalty and loyalty as a whole.However, it is important to point out to readers and administrators of universities and higher educational managers to some of the limitation of the present analysis.The first limit is the issue of data collection which though relatively large but cannot be generalized to the entire educational universities in Kenya and elsewhere.Secondly, limiting it to even only students of the University of Nairobi in itself is a limitation.These limitations suggest the need for further study in Kenya and elsewhere in the world.The study is of the view that a mixed approach i.e. qualitative and quantitative approach could be helpful or qualitative approach and could also be worth researching.A further outlet for future study is also to look at the other behavioural factors of employees such as empathy, trust, courtesy, and capability (cf.de Waal et al., 2016), which could be combined with the behavioural factors examined to assess students' loyalty.Finally, this empirical analysis has just shown the starting point of the journey toward behavioural factors of employees in building STLOY in the higher education industry and is far more complex than earlier thought.For administrators of university education and higher education, the study findings are a welcome opportunity for them to reflect upon.As they reflect upon the findings, the study urges them to continue to allocate substantial resources for training and motivating staff to display behaviours towards building INTIMACY and STLOY among students.

Table 2
Descriptive statistics of research measures Source: Empirical data, 2017 Hair et al. (2017)c about reporting GoF index in PLS-SEM study.Meanwhile, researchersHair et al. (2017)indicated that given the newness of this statistics in say PLS-SEM literature, researchers should be cautious about reporting fit statistics and SRMR values.More details of the measurement model are presented in Tables 3, 4a and 4b.

Table 3
Measurement Model

Table 6
Report of R-Squared statistic and Cohen's Effect size Source: Empirical data, 2017

Table 7
Direct effect inference

Table 8
Indirect effects inference