PLS-SEM Based Analysis of Service Quality and Satisfaction in Open Distance Learning in Sri Lanka

Students’ Perceived Service Quality (SPSQ) and Students’ Satisfaction (SSAT) in Open Distance Learning (ODL) are renowned features in Higher Education (HE) field in the present-day. The vast number of research studies have addressed diverse types of variables on different grounds. The ODL students are from various family backgrounds such as cultural, social, technological skill levels, and most importantly defer from basic educational entry qualifications. This must cognize how the HE institutions are able to offer their diversified services together with quality enhanced features to enjoy a competitive edge from the student’s satisfaction point of view. The number of students completing the course, or graduates passing out are diminishing when compared with the increase in number of registered student’s year on year. Most of the universities and HE institutions have faced student persistence and attrition problems and rush to find solutions with the concepts of service quality and satisfaction. The purpose of this study is mainly based on finding out the significant factors affecting the SPSQ and SSAT. This study will address the modified SERVQUAL constructs in relation to the SPSQ and SSAT. The independent variables were Assurance, Empathy, Responsiveness, Reliability, and Website Content and the dependent variables were the Students’ Perceived Service Quality and Satisfaction in ODL in the OUSL. Data was collected by using a self-administered questionnaire from 760 undergraduate students of the Open University of Sri Lanka (OUSL) covering six main regional centers island wide. The OUSL is the unique university operating under the ODL system in Sri Lanka. The analysis was based on descriptive and inferential statistics. The significant relationships are revealed between the SPSQ and Reliability, Responsiveness and Website Content. The other significant relationships are SSAT with Reliability and Website Content and the relationship between SPSQ and SSAT. The R for the SPSQ is 0.244 (24%) and SSAT is 0.549 (55%). The same research could be recommended as a future research with more service quality variables, and in regional centers as a longitudinal data collection method to understand more about service quality and satisfaction which influence student retention and completion.


Introduction
Service Quality and Student Satisfaction in Open Distance Learning (ODL) are becoming more important day by day with the increasing demand of higher education sector.Distance Education (DE) has formed as a supplement method of teaching and learning process over the world.The distance education is defined as a mode of teaching and learning characterized by separation of teacher and learner in time and/or place for most part of the educational transaction, mediated by technology for delivery of learning content with possibility of face-to-face interaction for learner-teacher and learner-learner interaction, provision of two-way didactic communication, and acceptance of industrial process for division of labour, and economies of scale (Open and Distance Learning -are graduating with ODL system with higher education qualifications they will able to use that qualifications to gain higher remunerations (Simpson, E-Learning and the Future of Distance Education, 2013).The theory of student integration by Tinto in 2008 as cited by (Jones D. , 2009 ) " The time of adjustment for a student to their new environment is a time of risk for withdrawal as students may suffer stress and a sense of loss".The problem of student attrition or dropouts has to be solved by the universities or the higher education institutes.The Open University of Sri Lanka also has faced the same situation of high dropouts of registered students (Statistical Handbook 2013, 2014;Strategic Management Plan 2015-2020, 2014).The factors influencing for high dropout rates in OUSL faculty of engineering identified as mismatch with the expectations and entry qualifications, old-employed learners, difficult to update the knowledge with latest technologies, poor language proficiencies, physical distance and jobs and family constrains (Liyanagama, 2014).Consequently in 2009 Jones has uncovered the reasons for the student attrition as selection of the wrong course, lack of timely support/course information, isolation, unbearable workload with assignments and assesment criteria.The 'rivergram' has showed the dropouts of the students of their first course in Open University of United Kingdom (UKOU) and Distance student must be given the support in various time in their learning journey (Simpson, Student Support Service for Success in Open and Distance Learning, Feb 17, 2016).
Whether the students need more support or few , helping a student will be more effective to overcome their problems and reached to thier expected porgress (Simpson, Student Support Service for Success in Open and Distance Learning, Feb 17, 2016)."The quality factors were strongly linked to academic progress which closely related to student retention" (Chakuchichi, 2011, p. 90).The Perceived service quality is defined as" an overall evaluation of the goodness or badness of a product or service" or as an attitude.The consumer satisfaction is "similar to attitude, but it is short-term and results from an evaluation of a specific consumption experience" (Athiyaman, 1997)."Delivering of quality service has become an important goal for most higher education institutions" (Athiyaman, 1997, p. 539).The student retention is defined as "the act of retaining or keeping students on the programme until its natural conclusion or they have achieved their academic goals" (Chakuchichi, 2011, p. 90;Jones, Berry, Gregson, & Smith, 2008)."For the ODL institutions to improve quality of service provision in order to manage student retention.In seeking to improve student retention ODL institutions could work out intervention strategies that impact quality instruction, information computer technology and the economies of scale regarding affordability of education" (Chakuchichi, 2011, p. 97).The high noncompletion rate recorded from the OUSL in very popular programme of Certificate in Preschool Education.The recommended solution was given to minimize this situation was to enhance the quality perspectives and effectiveness of the program (Ariyaratne, Munasinghe, Seneviratne, Rajapaksha, & Dediwala, 2014).

Problem Statement
It has been evident that the completion rate of the registered courses are very low in comparison to the student enrollment rate in the Open University of Sri Lanka (Statistical Handbook -2013-Open University of Sri Lanka (OUSL), July 2014) and consequently the student dropout or attrition rate is extremely high (Strategic Management Plan 2015-2020, 2014).The said low level graduation rate and low continuity rates are serious concerns.Researchers have predicted many reasons and recommendations (Ariyaratne, Munasinghe, Seneviratne, Rajapaksha, & Dediwala, 2014;Ismail, 1997 August;De Zoysa, Munasinghe, Seneviratne, & Mukunthan, 2011;Liyanagama, 2014).The deficiencies in offered service by the institution and perceived by the students is one of the main reason and empirically has proven which will lead to student satisfaction or dissatisfaction (Jurkowitsch, Vignali, & Kaufmann, 2006 ).In marketing perspectives the consumer perceived service quality, satisfaction and behavioral literature has been used to ascertain the customer service quality and satisfaction (Cronin & Taylor, 1992;Parasuraman, Zeithaml, & Berry., 1988;Parasuraman, Zeithaml, & Malhotra, 2005;ZeithamI, Berry, & Parasuraman, 1996).In the higher education enviornment the most famous validated instrument of SERVQUAL (Parasuraman, Zeithaml, & Berry., 1988) has applied these service quality concepts in various higher educational research studies ( (Mbwesa, 2014;Ahmed & Masud, 2014;Al-Alak & Alnaser, 2012).In 2011 Udo, Bagchi, & Kirs has used modified SERVQUAL instrument by replacing the dimension of Tangibility by Website Content (Udo, Bagchi, & Kirs, 2008) to reflect the e-learning and online settings in ODL.To find the significant service quality dimensions which would affect Students' Perceived Service Quality and Satisfaction empirically which could be enhanced the ODL environment in OUSL for their students' retention and completion.

Research Objectives
1. To determine the significant factors contributing towards Students' Perceived Service Quality and Satisfaction in ODL in OUSL.
2. To ascertain the relationship between Students' Perceived Service Quality and Satisfaction in ODL relating to OUSL. 3. To model the significant factors in the final model.

Research Questions
Based on the service quality literature the dimensions of modified SERVQUAL are Assurance, Empathy, Reliability, Responsiveness, and Website Content which could be considered to develop research questions to test the association with student satisfaction in ODL (Mantovani, 2012;Udo, Bagchi, & Kirs, 2008;Udo, Bagchi, & Kirs, 2011;Stodnick & Rogers, January 2008).Research questions of the study are as follows:

Significance of the Study
The results of this study enlighten the path for the academia in ODL in the OUSL.The higher management and policy makers could understand the ranking order of the affective significant factors.They can do the necessary changes conferring to those results in the teaching and learning process to take the strategic decisions to develop the ODL system.
The policy makers can change the policies to enhance the teaching-learning process while implementing the quality assurance process to every program.It is very important in assessing whether every change must link-up with support for the learners from their initial stage to fulfil their requirements and successfully complete their learning journey.

Literature Review
The service quality concept originated in a marketing environment, but it has gradually moved in to the service industry as well as in the educational environment.Since education is a long-term service it is very important to study the facts connected with service quality and student satisfaction.

Service Quality
In the Service Quality paradigm, it can be find the different types of SQ items and instruments based on the nature of the service and service organisations.The perceived service quality model was constructed under the three dimensions of Functional quality, Technical quality, and Reputational quality of the organisation (Grönroos, 1984).In 1985 Parasuraman, Zeithaml, and Berry (Parasuraman, Zeithaml, & Berry, 1985) conceptualised a gap model for Service Quality with ten dimensions and subsequently in 1988 they reduced it to a 5 item scale (SERVQUAL) (Parasuraman, 1985) as a scale for measuring consumer perceptions of perceived service quality.It consisted of the dimensions of Assurance, Responsiveness, Empathy, Reliability, and Tangibility.In 1990 Carman has followed generic qualities of the SERVQUAL instrument in three service settings a tyre retailer, a business school and a dental school.With the factor analysis he has identified the underlying dimensions between five to seven.Cronin and Taylor (1992) contradict the framework of SERVQUAL and proposed a 'performance only' new serice quality measurement scale was called SERVPERF.In 1993 Harvey & Green, 1993 conceptulised the nature of the concept of quality in relation to higher education " Quality can be viewed as exceptional, as perfection, as fitness for purpose, as value for money and as transformative" and it depend on the stakeholder preferences.In 1997 Peterson, Kovel-Jarboe, & Schwartz have presented a conceptual frame work with key elements which could be practised in higher education and hypothesised between the the quality improvements and the student retention.Managing service quality in higher education (Hill, 1995).In 1997 Athiyaman has presented eight services/service attributes as the determinants of service quality in higher education and conceptualised the basis of consumer satisfaction and perceived quality in the higher education sector and their importance.Consequently in the same year by LeBlanc & Nguyens examined the concept of service quality in business education and identified seven significanr factors which influence student evaluations of service quality.The "descending order of importance of the factors are: reputation, administrative personnel, faculty, curriculum, responsiveness, physical evidence and access to facilities" (LeBlanc & Nguyen, 1997, p. 72).In 1997 Joseph & Joseph have examined the determinants of business students' perceptions of service quality in education in New Zealand and revealed the career opportunities, programme issues, cost/time, physical aspects, location, and other were most significant factors relating to service quality.In 2000, Oldfield & Baron has itentified three dimensions of requisite elements, acceptable elements and functional elements or which are practical in nature.In 2004 Rashid & Harun has found the the dimensions of serice quality as gender, ethnicity, type of academic programs, and location of learning centers in an ODL institution in Malaysia.The new measurement scale was developed for higher education enviornment called HEdPERF (Higher Education PERFormanceonly) (Abdullah, 2006).The six items in the scale were Non-academic aspects, Academic aspects, Reputation, Access, Program issues, and Understanding.With the development of the information technology a new distinctive process of people-technology interactions has added to the research studies and developed multiple-itemscale (E-S-QUAL) for measuring the service quality delivered by Websites e-service quality (Parasuraman, Zeithaml, & Malhotra, 2005).In 2006, Prybutok, & Huang presented e-service quality model which was consisted of individual differences, e-service convenience, Website service quality, risk, e-satisfaction, and intention.They have empirically proved all the constuts were significant except "individual differences" in this measurement scale.The SERVQUAL intrument was applied to the traditional classroom enviornment in a innovative nature by Stodnick & Rogers, in 2008 for the first time to evaluate students' perceived service quality associated with students' satisfaction with the course, instructor, learning and expected grade.In 2008 Udo, Bagchi, & Kirs have operationalised web service quality constructs as perceived risk, web content and service convenience.Moreover they have further analysed the relationship with customer satisfaction and behavioral intentions in an e-business environment.In 2011, Udo, Bagchi, and Kirs has empirically presented the modified SERVQUAL constructs by replacing the " Tangibility " dimension with " Website Content " and found significant with four variables except the dimension of Reliability in e-service quality ODL enviornment.Consequently, in 2012, Mantovani has used the same instrument and found the significant dimensions as Assurance, Reliability, and Website Content.In 2012, Çerri has applied the SERVQUAL in Albanian public universities to investigate the service quality consequences and concluded that all five dimensions are significant and the antecedents of service quality.

Student Satisfaction
The definition of Student Satisfaction was specified by Udo, Bagchi,and Kirs,(2011) as "overall assessment of the student learning experience".In 2006 Arambewela & Hall (Arambewela & Hall, 2006) has examine the relationship between the SERVQUAL constructs and satisfaction among four cohorts of Asian international postgraduate students studying in Australian universities.In 2008, Hasan, Ilias, Rahman, & Razak have examined the relationship between SERVQUAL and students satisfaction of bachelors degree students studying at a private HEIs in Malaysia and found that all 5 dimensions were positively significant with student satisfaction.In 2011, Wei and Ramalu have examined the relationship between service quality and student satisfaction in a university of Malaysia by using SERVQUAL instrument and found the significant factors as, Responsiveness, Assurance and Empathy.In 2014 Mansori, Vaz and Ismail used the same indtrument for the private universities and colleges in Malaysia and revealed that Tangibility, Empathy, and Responsiveness were significant with student satisfaction.In 2015, Sembiring has used the SERVQUAL instrument which directly influence the students' satisfaction in ODL perspectives and found the significant relationship with all variables and the students' satisfaction.Assessment of factors which affect Student Satisfaction upon Service Delivery in the newly established universities in Sri Lanka (Wijesiri, 2016).In Zambia, in higher education, service quality performance dimensions (tangibility, reliability, responsiveness, empathy and assurance) are positively significant with overall customer satisfaction (Mwiya, et al., 2017).

Service Quality and Students' Satisfaction in ODL
In 2011 ,Udo, Bagchi and Kirs have proposed a modified SERVQUAL model to evaluate a relationship between e-learning quality in distance education and students' satisfaction and found significant.Followed by the same diemensions, in 2012, Mantovani has empirically proved significant relationship between perceived service quality in distance education and students' satisfaction.Consequently, the studies in higher education institutions disclosed that service quality has a positive significant effect on student's satisfaction (Jiewantoa, Laurens, & Nellohc, 2012 ;Meštrović, 2017).

Methodology
Based on the literature the modified SERVQUAL five (5) dimensions of Assurance, Empathy, Responsiveness, Reliability and Website Content were considered as the independent variables (IV).The dependent variables (DV) for the study were Students' Perceived Service Quality and Satisfaction in ODL in the OUSL.The Instrument was mainly based on literature (Stodnick & Rogers,2008 ;Udo, Bagchi, & Kirs ,2008, 2011) and minor modifications were made.The scale of the research instrument except the demographic questions rated with the 5-point Likert scale; 1-Strongly Disagree to 5-Strongly Agree.The Primary data collection method would be a survey based on a self-administered questionnaire consisting of 46 questions which have been validated and tested for reliability in the pilot survey.The six main regional centers in the OUSL were selected and the respondents were 760 undergraduate students who have more than one-year university service experience.The valid responses were only 744 which covered 98% of the original sample.The Data analysis consists of descriptive and inferential statistics.The Structural Equation Modeling (SEM) was employed by using the 'Partial least square' (PLS) to check the significance of each of the dimensions in the model proposed for evaluating the Students' Perceived Service Quality and Satisfaction.The significant constructs will be used only for model fitting.There are two main statistics model styles known as 'parametric' and 'non-parametric' which implies that a distribution is assumed for the population.An assumption is made when performing a hypothesis test that the data is a sample from a certain distribution; commonly, normal distribution.Non-parametric implies that there is no assumption of the specific distribution of the population which means the non-parametric test results are more robust against the violation of the assumptions (Field, 2011, pp. 264-315;Saunders, Lewis, & Thornhill, 2011, pp. 449-453).
The following eleven hypotheses were tested.

H 1A :
There is an association between Assurance and Students' Perceived Service Quality in ODL in OUSL 2. H 2A : There is an association between Assurance and Students' Satisfaction in ODL in OUSL.

H 1B :
There is an association between Empathy and Students' Perceived Service Quality in ODL in OUSL.

H 2B :
There is an association between Empathy and Students' Satisfaction in ODL in OUSL.

H 1C :
There is an association between Responsiveness and Students' Perceived Service Quality in ODL in OUSL.

H 2C :
There is an association between Responsiveness and Students' Satisfaction in ODL in OUSL.

H 1D :
There is an association between Reliability and Students' Perceived Service Quality in ODL in OUSL.

H 2D :
There is an association between Reliability and Students' Satisfaction in ODL in OUSL.

H 1E :
There is an association between Website Content and Students' Perceived Service Quality in ODL in OUSL.

H 2E :
There is an association between Website Content and Students' Satisfaction in ODL in OUSL.

H 3 :
There is an association between Students' Perceived Service Quality and Students' Satisfaction in ODL in OUSL. ijbm.ccsen

Data
The resea Rogers, J research i with 1-St self-admi from the undergrad only 744 The test of construct validity was conducted using PLS -it has reduced restrictions on measurement scales, sample size, and residual distributions, compared to the covariance-based SEM.Smart PLS was also used to create the component-based regression/path analysis model.PLS has the advantage of being able to model multiple dependent and independent variables while handling multi-collinearity among the independent variables.It is also robust in handling missing data and the basis of cross-products involving the response variable(s), hence resulting in stronger predictions (Chin, 1998;Hair, Hult, Ringle, & Sarstedt, 2017).

Survey Material
The survey items were mainly based on literature which were validated from the related studies.Minor modifications were made to reflect the ODL environment.The research material consisted of two (2) sections.
Section-1 was the demographic data section and the other one consisted of 46 questions which were rated using the 5-point Likert scale, with 1-Strongly Disagree to 5-Strongly Agree (Appendix O consists of Section-2).The survey instrument was mainly developed to collect data and to test the hypotheses based on the proposed conceptual frame work.

Assurance, Empathy, Responsiveness, and Reliability
These four constructs were mainly based on the original SERVQUAL (Parasuraman, Zeithaml, & Berry., 1988) 2012) used same variables for the distance education e-learning enviornment in Brazil.There were 26 questions in section-2 based on these four(4) constructs.The Assurance in ODL in the OUSL mainly focused on knowledge and courtesy of the staff and has seven(7) items.The Empathy in ODL in the OUSL assessed the individualised caring and attention and consisted with 6 (six) items.The responsiveness in ODL in the OUSL evaluated as the willingnes to help the the students and make available a prompt service and measured from six (6) items.The construct of Reliability in ODL in the OUSL guaged through the reliable promised service consistently and accurately from seven(7) items (Udo, Bagchi, & Kirs, 2011).

Website Content
The eight (8) items for this construct were taken from (Udo, Bagchi, & Kirs, 2011) and validated from the pilot survey.The questions are intended to measure the usefulness, quality, relevancy of the intended information and how multimedia features integrated to the website; the efficient way of applying the audio, video graphics and animation for on-line lessons.

Perceived Service Quality (PSQ)
Perceived service quality is described in terms of the "customer's assessment of the overall excellence or superiority of the service" (Mantovani, 2012;Zeithaml, 1988).Measures of Perceived Service Quality have been developed in previous studies (Mantovani, 2012;Udo, Bagchi, & Kirs, 2011).There were three items which reflect overall perceptions of quality, course content and superiority of the service when compared with other institutions.

Data Analysis and Results
PLS-SEM results are evaluated using a systematic process.The main purpose of this method is to maximise the explained variance (R 2 ) of the endogenous latent variables in the PLS path model.As a result, the evaluation quality of the PLS-SEM measurement and structural models focuses on the predictive capabilities of the models.
The systematic evaluation of these criteria follows a separate assessment of a two-step process of the measurement model and structural model.The structural model will be evaluated after establishing the reliability and validity of the constructs.There are two types of measurement model specifications: Reflective and Formative measurement models (Hair, Hult, Ringle, & Sarstedt, 2017).
In the reflective model, the relationship (causality) goes from the construct to its measures.Reflective indicators could be viewed as a representative sample of all the possible items available within the conceptual domain of the construct.All indicator items are caused by the same construct and should be highly correlated with each other.The construct must have sufficient reliability, even the single item exchange or drop (Hair, Hult, Ringle, & Sarstedt, 2017).In formative measurement model, the construct forms the causal indicators.The most important feature of this model is the formative indicators cannot be interchangeable.The omitting of an indicator potentially alters the nature of the construct (Hair, Hult, Ringle, & Sarstedt, 2017).For the current study, the conceptual model consists of the reflective measurement model which has the relationships from construct to the indicators.
The results of the outer loadings of the reflective constructs (Appendix A) shows some items are having less than 0.7.As a rule of thumb, the outer loadings should be at least 0.7 or more.The Indicator Reliability which is the square value of Outer Loadings of the items, and which is giving the minimum value as 0.50 that will cover the 50% of the value as the variance extracted from the item (Hair, Hult, Ringle, & Sarstedt, 2017).The indicators which indicated the poor Indicator Reliability (<0.50) or Outer loadings (<0.7) were dropped.The dropped indicators were (ASS7P=0.655,EMP5P= 0.660, EMP6P= 0.668, REL5P=0.684,REL6P =0.620, REL7P=0.688,RES6P= 0.625, SAF8P= 0.669, SAF9P=0.666).
As a rule of thumb, the Cronbach's Alpha and Composite Reliability values must be greater than 0.7 and the AVE must be higher than 0.50.The results of Cronbach's Alpha (CR), Composite Reliability and Average Variance Extracted (AVE) have reached up to the required standard (Appendix B) ( (Hair, Hult, Ringle, & Sarstedt, 2017).
The assessment of the discriminant validity the Fonell-Larcker criterion, the square root of the AVE of each construct should be greater than the constructs' highest correlation with any other construct in the model.The results of the Fonell-Larcker criterion (Appendix D) presented the square root of the reflective constructs' AVE on the diagonal values are greater than any value of the correlation between the constructs in the off diagonal raw and column position.Another discriminant method is Cross Loadings of the indicators loadings.All loadings on its assigned construct must be higher than all other cross loadings of the other constructs.The results showed (Appendix C) all loadings with the assigned constructs have reached the highest correlation with their own indicators.The third method of the discriminant validity is the Heterotrait-Monotrait (HTMT) values.The bootstrap (5000 samples) confidence intervals values should be significantly different from one (1).The columns labelled 2.5% and 97.5% show the lower and upper boundary of the 95% bias corrected confidence interval.As the results showed (Appendix E) neither of the lower or upper boundary confidence intervals are included in the value of 1 (Hair, Hult, Ringle, & Sarstedt, 2017;Henseler, Ringle, & Sarstedt, 2015).

Structural Model Fit Analysis
After analysing the measurement model, the structural model will be assessed.As a first step collinearity must be accessed.The critical levels of the Tolerance values are less than 2 and the Variance Inflation Factor (VIF) are less than 5 to avoid the multicollinearity problems.The values of the VIF values in the (Appendix F) and all values are less than or equal 2 and can be confirmed there is no issues of multicollinearity problems (Hair, Hult, Ringle, & Sarstedt, 2017).The key criteria for assessing the structural model in PLS-SEM are the significance of the path coefficients.The 1 st step of the structural model path coefficients were checked against the rule of thumb from 5000 samples of bootstrap t values and the p values.The critical t values for two tailed tests are 1.65 (significance level 10%), 1.96 (significance level 5%) and 2.57 (significance level 1%).When p values are considered, if the significance level is 5%, the p values must be smaller than 0.05 to consider the relationship as a significant at 5% level.The significant relationships are revealed between SPSQ and Reliability, SPSQ and Responsiveness, SPSQ and Website Content.And consequently, the Students SAT and Reliability and Students SAT and Website Content.And the Students' Perceived Service Quality and Students' Satisfaction (Appendix I).The non-significant relationships were deleted from the model and final model was generated again with the 5000 samples of bootstrap method (Hair, Hult, Ringle, & Sarstedt, 2017).The Fig. 2 shows final model of the study with only significant relationships and significant outer loadings of each construct (Hair, Hult, Ringle, & Sarstedt, 2017).
The most common method of evaluating the structural model is the coefficient of determination (R 2 value) which is used to measure the predictive power of the model.This value represents the exogenous latent variable's combined effects on the endogenous latent variable.That is the amount of variance in the endogenous constructs explained by all of the exogenous constructs linked to it (Hair, Hult, Ringle, & Sarstedt, 2017).The R 2 for the Students' Perceived Service Quality is 0.244(24%) and can be considered as rather weak and Students Satisfaction is 0.549 (55%) can be considered as moderate (Appendix J).
The SRMR is the square root of the sum of the squared differences between the model-implied and the empirical correlation matrix (Henseler, Hubona, & Ray, 2016).A value of 0 for SRMR would indicate a perfect fit, but models can yield SRMR values of 0.064 and higher (Henseler, et al., 2014).Therefore, a cut-off value of less than 0.08 as proposed by (Hu & Bentler, 1999).For this model the SRMR is 0.064 and considered as a good fit.
The RMS theta builds on the outer model residuals, which are the differences between predicted indicator values and the observed indicator values.RMS theta values below 0.12 indicate a well-fitting model, whereas higher values indicate a lack of fit (Henseler, et al., 2014).For this model RMS theta is 0.117 and considered as good fit (Hair, Hult, Ringle, & Sarstedt, 2017).
The summary of the finalised measurement model is depicted in Table 1.All the significant constructs included in the final model are tabulated with their Outer Loadings, Indicator Reliability, AVE (Appendix K), Composite Reliability (Appendix L), Cronbach's Alpha (Appendix M) and HTMT (Appendix N) values.All the constructs have indicated strong reliability and validity values.(Mbwesa, 2014, p. 86).The delivery of a quality service continually is very difficult but it can be risen the organisational revenues (Zeithaml, 1988).As a future research the relation between student retention and attrition based on the revenu of the institution can be studied.

Copyrights
Copyright for this article is retained by the author(s), with first publication rights granted to the journal.
This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).WSC6. WSC7.
The website uses audio elements properly.
The website uses video elements properly.
Service Quality; Reflect perceptions of quality, clarity of instruction, currency of information, and functionality of the features on the website.OLQ1(Udo et al.,2011)Over all service quality is excellent Likert scale (1..5) OLQ2 (Researcher) Overall quality of the course content is satisfactory OLQ3 (Researcher) Excellent service quality offered compare to other institutions 7.Satisfaction: The overall assessment of the student's distance learning experience, including overall pleasure and satisfaction with the to say that ''I am satisfied with my decision to enroll with this distance programme''Would you agree to say that ''My choice to enroll this programme was a wise one?''Would you agree to say that ''I think I did the right thing when I paid for this learning service?''Would you agree to say that ''I feel that my experience with distance learning has been enjoyable?''Thisdistance learning course meets my expectations.My Overall experience is better than I originally anticipated.I am Overall satisfied with the programmes and services offered by the Institute.I am satisfied with the distance course since it will give me a better chance in my career development.I am delighted with the distance course and its contents.Likert scale(1..5) 1. How does Assurance impacts Students' Perceived Service Quality and Students' Satisfaction in ODL in the OUSL? 2. How does Empathy impacts Students' Perceived Service Quality and Students' Satisfaction in ODL in the OUSL? 3. How does Responsiveness impacts Students' Perceived Service Quality and Students' Satisfaction in ODL in the OUSL? 4. How does Reliability impacts Students' Perceived Service Quality and Students' Satisfaction in ODL in the OUSL? 5. How does Website Content impacts Students' Perceived Service Quality and Students' Satisfaction in ODL in the OUSL? 6.How Students' Perceived Service Quality does impacts Students' Satisfaction in ODL in the OUSL?

Table 1 .
The results summary of the final measurement model VIF) of the constructs in the final model are distinctly below the threshold value of 5. Therefore, no collinearity issues of the predictor variables in the structural model (Appendix H).The R 2 value for the endogenous variables of Students' Perceived Service Quality is 0.244 and Students' Satisfaction is 0.549 and can be considered as weak and moderate respectively as compared with the threshold values.All the constructs in the finalised structural model are significant and the path coefficients and significance levels are shown in Fig.2.Values are given in (Appendix I).The order of the relative importance of the significant drivers of the Students' Perceived Service Quality is; Website Content (0.323), Reliability (0.214) and Responsiveness (0.079).The ranking order for the significant exogenous variables of Students' Satisfaction are Website Content (0.415) and Reliability (0.323).There is a significant relationship between the Students' Perceived Service Quality and the Students' Satisfaction (0.224).The summary of the hypotheses are tabulated in Table2.

Table 2 .
The summary of the hypotheses (Appendix G and Appendix I) A