Higher Education Quality and Student Satisfaction: Meta-Analysis,Subgroup Analysis and Meta-Regression

Abstract

This article reports the findings of a systematic literature review and meta-analysis carried out to identify various quality dimensions in higher education that contributes to student satisfaction. The results of heterogeneity generated from the meta-analysis were further explored via subgroup analysis and meta-regression. To fulfil these objectives, we summarized 122 qualitative studies and 40 quantitative studies that were relevant and important. Most of the included studies for meta-analysis (N = 40) assessed the relationship between different quality dimensions in higher education institutions and student satisfaction. All six dimensions were found to be positively associated with the dependent variable (student satisfaction). Administrative service quality was found to have the highest correlation with student satisfaction, followed by student support quality, lecturer quality, teaching quality, curriculum quality, and physical evidence quality. The three categorical variables (student, university, and country) were included to test for moderation individually and collectively—these categorical variables, when taken together, moderates all the relationships except lecturer quality–student satisfaction. To our understanding, the current research will be the first one to examine the relationship between six higher education quality dimensions and student satisfaction based on a meta-analysis and meta-regression. The current analysis offers more novel details on the detection of significant moderating variables (type of student, type of university, and country of the study). These consolidated findings help us in identifying the existing state of literature, that is, what we consider and what we do not.

Keywords

Quality higher education student satisfaction meta-analysis meta-regression subgroup analysis moderators

Introduction

Quality is a wicked concept¹; it was also termed as “contradictory” and “chameleon-like” by Vidovich.² Quality is now an important goal for all organizations globally.³ It leads to excellence, and “excellence is the progression to achieve the greatest heights”.⁴ It is complex to define quality singly and comprehensively.⁵ Several scholars tried to define quality in different ways; a few definitions are as follows: “a high degree of goodness”⁶; “doing the right things right”⁶; and “the totality of features and characteristics of a product or service that bear on its ability to satisfy stated or implied needs”.⁷

Cheng and Tam⁸ concluded that the quality of education is quite an ambiguous and divisive term. A vast number of people with divergent agendas are interested in education; they are all likely to have their conceptualizations of the term quality. The controversy on the status of students as customers is still contentious.^{9, 10} It is a “stakeholder-relative” term, that is, perceptions about quality differ from person to person.^{8, 11, 12} Therefore, such a wide spectrum of stakeholders increases the wickedness of this issue. Freeman defined stakeholders as “any group or individual who is affected by or can affect the achievement of an organisation’s objectives”.¹³ All the stakeholders are concerned about the education quality provided by the educational institutions with which they are associated. Whether they are teachers, students, principals, top management officials, or government (policymakers), etc., they all are concerned about the teaching quality, course and curriculum quality, and institutional quality.^{3, 14, 15} The satisfaction of all stakeholders is preeminent for higher education institutions (HEIs).¹⁶ Institutions should strive to cater to the needs and demands of all the interest-holders as their satisfaction with services can only make services of the institutions credible.¹⁰ “To improve the quality of higher education, we need to extend beyond the traditional paradigm of higher education quality assurance and move toward quality culture.”¹⁷ It is imperative to carry out the holistic approach in building a quality culture for HEIs by taking into account the needs, trust, and social contract between all the stakeholders.¹⁷ For institutions to sustain and survive, stakeholders’ satisfaction is the key.¹⁰ Their views and recommendations should be heard and considered while taking important decisions for educational institutions.¹⁰ This study tries to answer three search questions:

Research Question 1: What are the important quality dimensions in the higher education sector?

Research Question 2: What is the relationship between these important quality dimensions and student satisfaction?

Research Question 3: Are there any significant moderating variables, moderating these relationships?

Research Background

Education plays a powerful role in the evolution, advancement of any country and also in improving the standard of living of its people.¹⁸ The economic success of any country is dependent upon its quality of education.¹⁹ The higher education sector is growing tremendously worldwide with increasing academic institutions and student’s enrolment, thereby increasing competition between HEIs.²⁰ Students are considered to be an important part of HEIs and are primary consumers of HEIs.^21–26 Student satisfaction is a topic of deeper interest to researchers and policymakers in the area of Higher Education.²⁷ Studies in this category focused on defining and assessing the determinants of educational experience that influence the students’ satisfaction.^{24, 28}

Most of the scholars have paid attention to students, as they are considered to be the primary customer and stakeholders of HEIs.^{12, 29, 26, 30, 31} These arguments were also supported by Crawford²² and Gallowlay.²³ Most of the past studies describe quality for HEIs from the point of students.^32–35 A few studies take academicians’ points of view into account.^36–44 But the same set of quality factors cannot describe quality from the viewpoint of other stakeholders of higher education systems such a administrators, employers, parents, and researchers.³⁷ All the stakeholders, including current students, have a broader view of the quality of education at HEIs. In another study by Hill,⁴⁵ it was observed that students relate quality in HEIs with a supportive and responsive learning environment. Srikanthan and Dalrymple⁴⁶ postulated that to ensure quality, managers and teachers should provide a supportive and interactive learning environment for students. In another study by Al-Amri et al.,³⁶ it was observed that students focused more on teaching-learning for rating HEIs in terms of quality, employers were found to be concerned with graduates’ research acumen, industry skills and social skills, and teaching faculty were found to be sensitive about accreditation processes and quality improvement.

Student Satisfaction

Tinto found that if students are satisfied, their retention rate is more and dropout’s rates are lower.⁴⁷ Mahmud et al.⁴⁸ postulated that colleges/universities should focus on their vision and mission statements, rules, regulations, and procedures and should try to align these with student satisfaction to become more student-centric or student-oriented.

Satisfaction levels of customers with regards to quality dimensions of HEIs have gained noticeable attention.⁴⁹ There is an increasing need to understand the factors that affect student satisfaction.⁵⁰ As consumerism is on the rise in HEIs, focusing on student satisfaction has become a crucial issue.⁵¹ If the quality of important dimensions of HEIs is taken care of, it can have a great impact on the overall satisfaction of students.⁵² Scholars defined satisfaction in several ways:

Satisfaction is something an individual desires as an outcome of a task or job he performs that fulfils his esteem needs.⁵³

A cognitive or affective reaction that emerges in response to a single or prolonged set of service encounters.⁵⁴

Student satisfaction is widely recognised as a short-term behaviour arising from an assessment of the student’s learning experience.²⁷

Student satisfaction is a simple but dynamic phenomenon. Measuring student satisfaction can be difficult due to the complexities of higher education and the uncertainty revolving around “students as customers”.⁵⁵ It is challenging to satisfy students, but student satisfaction can give a competitive edge to institutions. Campuses that routinely assess and act on measurements of student satisfaction tends to enjoy the highest rate of institutional and academic performance.

Scales Used in the Extant Literature

Several scales have been developed by scholars to measure customer satisfaction regarding higher education quality or service quality of higher education. SERVQUAL (SERvice QUALity), SERVPERF (SERvice PERFormance), HEdPERF (Higher Education PERFormance), and HiEdQUAL (Higher Education QUALity) are some of the important scales developed by past researchers to measure customer satisfaction with higher education quality. The SERVQUAL scale was developed by Parasuraman et al.^{56, 57} After a few years, Cronin and Taylor⁵⁸ developed the SERVPERF scale for measuring customers’ perceptions towards quality. Clemes et al.⁵⁹ observed that SERVPERF gave better results than SERVQUAL. Despite several critiques, the SERVQUAL model is considered to be having more applicability, flexibility, and practicability when compared to other models/scales. All the other scales (SERVPERF, HEdPERF, HiEdQUAL) are based on SERVQUAL only. SERVQUAL can be used in different areas for measuring service quality; therefore, Ho and Wearn⁶⁰ developed HETQMEX, that is, “Higher Education Total Quality Management Excellence Model”. Abdullah⁶¹ developed a specific scale known as the HEdPERF scale as a measurement instrument for mapping out different dimensions of higher education quality. The HiEdQUAL scale was developed by Annamdevula and Bellamkonda⁶² for evaluating education quality for HEIs. The dimensions used in these scales were reviewed critically, and an attempt was made to gather quantitative studies establishing a relationship between different quality dimensions and student satisfaction.

Quality Dimensions

In order to answer the first research question, this study searched for various quality dimensions that affect or influence student satisfaction from the existing literature. Six dimensions were identified from the extant literature, and they were lecturer’s quality (LQ), teaching quality (TQ), curriculum quality (CQ), physical evidence quality (PEQ), student support quality (SSQ), and administrative services quality (ASQ). These dimensions were observed as important for student satisfaction (see Table 1).

Table 1.

Previous Studies Reported a Relationship Between Various Quality Dimensions and Student Satisfaction^79–104

S.No	Quality Dimensions	Alternate Descriptors	Supporting Literature
1	Lecturer’s Quality (LQ)	Lecturer Quality, Faculty Credentials, Teaching Faculty Qualifications, Teaching Faculty expertise, Teachers’ Characteristics, Assurance	Augustine et al. (2016), Butt & Rehman (2010), DeShields et al. (2005), Gruber et al. (2010), Hanaysha et al (2011), ), Kajenthiran & Karunanithy (2015), Lim Kok Seng & Ling (2013), Mai (2005), Martirosyan (2015), Muhammed et al (2010), Mwiya et al. (2017), Parahoo & Tamim (2012), Sharabati et al. (2019), Trang et al. (2018), Weerasinghe & Fernando (2018), Wei & Ramalu (2011), Wilkins & Balakrishnan (2011), Zopiatis et al. (2014)
2	Teaching Quality	Quality of instruction, pedagogy, teaching methods, reliability	Augustine et al. (2016), Butt & Rehman (2010), DeShields et al. (2005), Elliott (2002), Farahmandian et al. (2013), Gruber et al. (2010), Gul et al. (2019), Hanaysha et al. (2011), Ijaz et al. (2011), Kajenthiran & Karunanithy (2015), Karma & Julin (2015), Ko & Chung (2014), Mahmud et al. (2018), Mai (2005), Martirosyan (2015), Muhammed et al. (2010), Mwiya et al. (2017), Ngamkamollert & Ruangkanjanases (2015), O’Driscoll (2012), Singh & Singla (2018), Tessema et al. (2012), Wei & Ramalu (2011), Wong et al. (2016)
3	Curriculum Quality	Overall design and delivery, usefulness, scheduling, content, availability, study program, course content, course structure	Allam (2018), Butt & Rehman (2010), DeShields et al. (2005), Farahmandian et al. (2013), Gruber et al. (2010), Lim Kok Seng & Ling (2013), Mahmud et al. (2018), Mai (2005), Martirosyan (2015), Muhammed et al. (2010), Sakthivel et al. (2005), Tessema et al. (2012), Trang et al. (2018), Osman & Saputra (2019), Weerasinghe & Fernando (2018), Wilkins & Balakrishnan (2011), Wong et al. (2016)
4	Physical-Evidence Quality	Availability, access, physical aspects, usefulness, IT support, Infrastructure, tangibles	Abu Hasan et al. (2009), Allam (2018), Al-Sheeb et al. (2018), Butt & Rehman (2010), Darawong & Sandmaung (2019), Elliott (2002), Farahmandian et al. (2013), Gamage et al. (2008), Gruber et al. (2010), Gul et al. (2019), Hanaysha et al. (2011), Hanssen, & Solvoll (2015), Ijaz et al., (2011), Kajenthiran & Karunanithy (2015), Lim Kok Seng & Ling (2013), Mahmud et al. (2018), Mai (2005), Martirosyan (2015), Mwiya et al. (2017), Ngamkamollert & Ruangkanjanases (2015), O’Driscoll (2012), Parahoo & Tamim (2012), Sakthivel et al. (2005), Sheikh Ali & Mohamed (2014), Siming et al. (2015), Singh & Singla (2018), Songsathaphorn et al. (2014), Weerasinghe & Fernando (2018), Wei & Ramalu (2011), Wilkins & Balakrishnan (2011), Zopiatis et al. (2014)
5	Student Support Quality	Accessibility, professionalism, helpfulness, responsiveness, understanding, empathy, advising, support services	Augustine et al. (2016), Darawong & Sandmaung (2019), DeShields et al. (2005), Elliott (2002), Farahmandian et al. (2013), Gruber et al. (2010), Hanaysha et al. (2011), Ijaz et al. (2011), Kajenthiran & Karunanithy (2015), Law & Cheung (2013), Muhammed et al. (2010), Mwiya et al. (2017), Ngamkamollert & Ruangkanjanases (2015), Roberts (2010), Sakthivel et al. (2005), Sheikh Ali & Mohamed (2014), Songsathaphorn et al. (2014), Tessema et al. (2012), Wei & Ramalu (2011), Zopiatis et al. (2014)
6	Administrative Services Quality	Responsiveness, management quality, institutions’ responsiveness, Institutional factors	Allam (2018), Augustine et al. (2016) Butt & Rehman (2010), Darawong & Sandmaung (2019), DeShields et al. (2005), Gul et al. (2019), Hanaysha et al. (2011), Ijaz et al. (2012), Kajenthiran & Karunanithy (2015), Mai (2005), O’Driscoll (2012), Parahoo & Tamim (2012), Sakthivel et al. (2005), Sharabati et al. (2019), Ushantha & Kumara (2016), Weerasinghe & Fernando (2018), Wei & Ramalu (2011)

Source: The author.

Lecturer’s Quality: Lecturers’ Quality and their credentials are essential for assuring quality in higher learning institutions.⁶³ In a study by Butt and Rehman,²⁰ it was found that student satisfaction was positively correlated with the teacher’s expertise. In general, the main function of the teaching faculty is to equip students with the pragmatic expertise that is most needed and tailored to the current era of technological advancement.⁶³

Teaching Quality: Students considered teaching and learning to be very critical to their satisfaction.^{20, 64} Teaching quality has a positive and strong relationship with student satisfaction.⁶⁵ Teaching styles and knowledge of the teaching faculty have a strong influence on student satisfaction.⁶⁶ Quality of instruction by teaching faculty is considered to be the essential dimension for the satisfaction of students with the overall quality of education.⁶⁷

Curriculum Quality: Courses offered by colleges/universities play an important role in enhancing student satisfaction.²⁰ The reasonability and clarity of the courses enhance student satisfaction.⁶⁶ The curriculum should be such that it meets the demand of the industry over time.⁶⁸ Curriculum planning is an important dimension for higher education quality as it influences learner behaviour and satisfaction.⁶⁹ Student satisfaction is an interplay of many factors, and one important factor is curriculum; therefore, universities should foster curriculum development.⁷⁰

Physical Evidence Quality: The presence of classroom facilities affects student’s satisfaction positively.²⁰ In a research study carried out by Gul et al.,⁷¹ tangibles, that is, resources and facilities, were found to be one of the major factors affecting student satisfaction. This was also observed in studies by Abu Hasan et al.,⁷² Ijaz et al.,⁷³ and Kajenthiran and Karunanithy.⁷⁴ If institutions focus on their resources and facilities, this helps in teachers’ instruction, thereby increasing student satisfaction.⁷⁵ HEIs should not restrict themselves to the improvement of only academic aspects and course content; if they want to increase overall student satisfaction, then focus should also be on non-academic aspects and facilities.^{76, 77}

Student Support Quality: An interaction between students and teaching staff is essential as it will make students feel that they belong to an institution, thereby enhancing their satisfaction level with the university and helping them achieve academic success. An institution’s responsiveness to students has a great impact on student satisfaction.⁷¹ In another study by Chuah and Ramalu,⁷⁵ it was found that student support, that is, care of students at the institutional level, is positively related to student satisfaction with the institution as a whole. This dimension came out to be one of the most important and strongest dimensions in the study. Student support services are essential services, and they enhance student college experiences, thereby enhancing their overall satisfaction.⁷⁷

Administrative Services Quality: Universities have learned that to maintain sustainability in the long term, they need to provide the best services, and the quality of these services will give universities a competitive edge over other universities.⁷⁸ Students view administrative services as authoritative and expect clear directions, advice, and friendliness.⁶³ Institutions should provide the needed administrative services, thereby maintaining good governance and ensuring the quality of an institution.¹⁰

Few scholars measure the quality of education and check their satisfaction from the students’ point of view. We have a smaller number of studies, both theoretical and empirical, regarding student satisfaction with higher education quality. And also, studies are divergent in terms of quality dimensions and their scope. Student satisfaction is taken as the major variable determining the quality of higher education in the present research. Therefore, this study attempts to do a meta-analysis in order to answer the second research question and see the consistency of results in the past studies relating to student satisfaction and various quality dimensions. The meta-analysis will help us in getting the combined effect size for different quality dimensions and student satisfaction.

Materials and Methods

This study employs the methodology as discussed by Tranfield et al.,¹⁰⁵ as it allows transparency and repetition and helps in avoiding research bias. There are three stages to perform the systematic literature review: (a) plan the review; (b) conduct the review; and (c) reporting and disseminating.^{105, 106} The present research included only papers that met the inclusion criteria of the research protocol. Articles and papers were explored on Scopus, Emerald, and Google Scholar databases and those published in the English language in peer-reviewed journals were included. Search terms included were “quality higher education” and “stakeholders’ satisfaction” or “student satisfaction”. Qualitative synthesis was done via SLR and Quantitative synthesis of the relevant studies was done via Meta-Analysis. CMA version 3 for analysis¹⁰⁷ and Meta-Essentials for generating forest plots were used.¹⁰⁸

PRISMA tool (see Figure 1) was used to conduct this systematic literature review and meta-analysis. PRISMA stands for “Preferred Reporting Items for Systematic reviews and Meta-Analyses”. In addition to the research protocol, Figure 1, based on PRISMA methodology, indicates a schematic process that represents the technique adopted, showing the number of records in each step. The study protocol restricts the study conducted between 1990 and 2020. This timeline was deemed acceptable due to the fragmented evidence of the publications before 1990. Subsequently, a search using the “targeted keywords” returned 845 papers and articles (332 in Google Scholar, 305 in Emerald, and 208 in Scopus). Google Scholar, Emerald, and Scopus databases were selected as their coverage, and systematic methodology culminated in a curated selection of records.

Figure 1.

Source: The author.

Summary of Data Included

Information on the variables was derived from each study and was entered into the meta-analysis. Most of these included studies were performed in South Asian Countries (India, Pakistan, Bangladesh, and Sri Lanka), South-East Asian countries (Thailand and Malaysia), West Asian countries (UAE and Dubai), East Asian countries (China and Taiwan), North American countries (USA), African countries (Zambia and Kenya), and European countries (Ireland and Germany). The sample size ranged from 91 to 5,223. Papers were written between 2002 and 2019. Most of the included studies (N = 40) assessed the relationship between different quality dimensions in HEIs and student satisfaction. The correlational data was extracted from different studies to conduct a meta-analysis. We searched for t-statistic and chi-square values from studies which were not having correlation values. The t-statistic and chi-square values were converted into correlational values by using statistical formulas as prescribed by Hunter and Schmidt.¹⁰⁹

Results

Meta-Analysis

Meta-Analysis is a systematic procedure for synthesizing quantitative results of different empirical studies with the aim of exploring the dispersion of effect sizes. Glass¹¹⁰ brought the technique of Meta-Analysis into the social sciences. This approach was accepted by medical sciences, and there is plenty of medicinal literature based on a meta-analysis. Currently, the use of meta in social sciences is in its nascent stage, but it is quickly taking over the current researches.¹¹¹ Meta-Analysis is aimed at finding an effect size for the studies. There should be at least two studies for conducting Meta-Analysis and at least 10 studies for conducting meta-regression.¹¹² These conditions are met in this present study. The statistical power may be lacking in a single study due to the small size of the sample, but when previous studies are combined, it increases the statistical power and also combining all the studies can enhance the precision of an estimate.¹¹³ An effect size is a quantifiable measure of the strength of the experimental effect and is the “unit of currency” in the meta-analyses.¹⁰⁷ It is the strength of covariation between two variables belonging to a homogeneous population.¹¹⁴ The higher the effect size, the stronger is the relationship between the two variables under consideration. “How strong is the association between a and b?” So it is a number that estimates the degree to which two variables are related.¹¹⁴ Effect sizes can be a correlation coefficient between two variables, regression coefficient, the mean difference or the risk of a particular event.¹¹⁵ The Pearson correlation coefficient (r) that ranges between –1 to +1 is widely used as an effect size when paired quantitative data are available.¹¹⁶ As per Cohen,¹¹⁷ the effect size of 0.20 is considered to be small, 0.40 as a medium, and 0.60 as large. There are two models for conducting a meta-analysis, and they are the fixed-effect model and random effect model. The fixed-effect size model assumes that all studies on which meta-analysis is performed share a “common or true” effect size. It means that the factors that are capable of influencing an effect size are the same across all the studies; hence, there is true effect size. For this reason, this model is known as the fixed-effect model, and the fixed effect size is denoted by theta (θ).^{107, 118} The age, education, health, and many other such variables can be different across studies; therefore, it is improbable to assume true effect size for all the studies. In this situation, it is better to perform a meta-analysis based on the random-effects model, which assumes that true effect sizes are normally distributed.¹¹⁸ This meta-analytic study synthesizes the previous studies based on the random-effects computational model as studies are similar, but we cannot say that they are identical. Studies have a different group of respondents belonging to different age groups, different countries, and different HEIs, and also studies are conducted in different time frames, therefore, it is better to perform meta-analysis based on a random-effects model.

Overall Association of Higher Education Quality Dimensions with Student Satisfaction

Table 2 illustrates the average “random-effect” correlations between different quality dimensions and student satisfaction. Quality dimensions were shown to have medium to broad correlations with student satisfaction. In Table 2, “SS” represents student satisfaction, “LQ” is lecturer’s quality, “TQ” is teaching quality, “CQ” is curriculum quality, “PEQ” is physical evidence quality, “SSQ” is student support quality, and “ASQ” represents administrative services quality. These quality dimensions were identified during the literature search and were included in meta-analyses for conducting quantitative synthesis and also to reach an overall conclusion of their relationship with student satisfaction.

Table 2.

Student Satisfaction with Observed Quality Dimensions

Relations	n	k	r+	LL	UL	Sig₁	Q	I²	Sig₂	Tau	Tau²
LQ-SS	6205	18	0.474	0.364	0.570	0.000	463.575	96.333	0.000	0.281	0.079
TQ-SS	15269	22	0.456	0.367	0.537	0.000	839.274	97.498	0.000	0.252	0.064
CQ-SS	9210	15	0.437	0.339	0.525	0.000	303.971	95.394	0.000	0.219	0.048
PEQ-SS	10320	28	0.398	0.305	0.482	0.000	761.286	96.453	0.000	0.277	0.077
SSQ-SS	12512	19	0.477	0.374	0.568	0.000	734.563	97.550	0.000	0.271	0.073
ASQ-SS	6166	17	0.493	0.374	0.596	0.000	528.486	96.972	0.000	0.303	0.092
Where, n = Total number of participants, k = Total number of effect sizes included in meta-analysis of each relation, r+ = average correlations as per random effects model, LL = Lower Limit (95%), UL = Upper Limit (95%), Sig₁ = Degree of significance for the effect size, Q-I²-Tau-Tau² = Tests of heterogeneity, Sig₂= Degree of significance for heterogeneity.

Source: The author.

ASQ was found to have the highest correlation (r+ = 0.493) with student satisfaction, followed by SSQ (r+ = 0.477), LQ (r+ = 0.474), TQ (r+ = 0.456), CQ (r+ = 0.437), and PEQ (r+ = 0.398). Physical evidence quality (PEQ) was found to have the weakest relationship with student satisfaction as per the results of meta-analyses. The results are found to be significant (p-value/sig₁ <0.05) and there is no zero between upper and lower limits. Therefore, all the six hypotheses that is, H₁, H₂, H₃, H₄, H₅, and H₆ are accepted as meta-analysis results that reveal that all the six dimensions positively influence student satisfaction.

Since a certain amount of difference in the measured effect size is predicted, we performed tests to find if the real effect size differs across studies. Therefore, we conducted different tests of heterogeneity, that is, Q, I², Tau, and Tau². “The Q-statistic (also referred to as “Cochrane’s Q”) is the weighted sum of squared differences between the observed effects and the weighted average effect”.¹⁰⁷ “I² is a measure for the proportion of observed variance that reflects the real differences in effect size”.¹⁰⁷ It is a relative measure and expresses results in percentage. Borenstein⁸¹ suggested that I² should be used as a measure of heterogeneity, especially when using subgroup analysis and meta-regression. “Both Tau and Tau² are a measure of the dispersion of true effect sizes between studies in terms of the scale of the effect size”.¹¹⁹ When heterogeneity is very high, the results of the meta-analysis can be explored by conducting subgroup analysis and meta-regression.¹²⁰ All the tests of heterogeneity (see Table 2) have shown a substantial amount of variability for each “effect size” thereby, justifying the use of subgroup analysis and meta-regression in search of moderators.

Subgroup Analysis

Subgroup Analysis is conducted to divide the studies into different subgroups based on certain characteristics that bring about differences in the effect sizes between subgroups.¹¹⁹ In this case, we have divided our studies based on the “type of student,” that is, undergraduate (UG) and postgraduate (PG); type of university that is, public university (PU) or private university (PR); and country of the study, that is, South East Asian countries (SEAC), South Asian Countries (SAC), West Asian countries (WAC), East Asian countries (EAC), European countries (EC), North American countries (NAC) and African countries (AFC). Subgroup analysis was conducted using comprehensive meta-analysis (CMA) Version 3.0.¹²¹ Table 3 summarizes the effect sizes and heterogeneity percentages for different subgroups. UG students were found with a high correlation for associations of lecturer’s quality, teaching quality, physical evidence quality, and administrative services quality with student satisfaction. In comparison, PG students observed a higher correlation between curriculum quality and support services quality with student satisfaction. Studies that had private universities reported a higher correlation of quality dimensions with student satisfaction when compared to public universities. Several countries were observed in different studies, and it was found that studies that were conducted in Southeast Asian countries reported a higher correlation between all the quality dimensions and student satisfaction.

Table 3 also shows heterogeneity statistics for all the subgroups, and no group with lower heterogeneity was observed except PG students showing only 12.63% of heterogeneity in the association of curriculum quality and student satisfaction. Such high heterogeneity statistics indicate that there are still a few latent characteristics that cause heterogeneity in these studies apart from the student, university, and country.

Table 3.

Sub-Group Analysis

Association	Groups →	Type of student		Type of University		Country of Study
Association	Effect size & heterogeneity statistics ¯	PG	UG	PU	PR	SEAC	SAC	WAC	EAC	EC	NAC
LQ-SS (H1)	r+	0.441	0.466	0.369	0.542	0.580	0.443	0.575	–	0.460	0.300
	I² (%)	55.88	96.96	97.28	93.87	79.59	89.82	96.77	–	97.44	80.07
TQ-SS (H2)	r+	0.485	0.497	0.423	0.534	0.670	0.404	0.532	0.596	0.495	0.293
	I² (%)	67.57	97.13	97.96	98.08	83.53	96.60	97.09	94.57	85.40	98.71
CQ-SS (H3)	r+	0.480	0.453	0.407	0.528	0.504	0.431	0.374	–	–	0.398
	I² (%)	12.63	96.42	91.42	94.02	71.10	98.13	93.98	–	–	91.06
PEQ-SS (H4)	r+	0.439	0.454	0.344	0.446	0.567	0.400	0.341	–	0.261	0.119
	I² (%)	93.98	91.78	95.18	97.31	69.87	95.38	94.21	–	92.36	96.83
SSQ-SS (H5)	r+	0.590	0.506	0.455	0.460	0.618	0.529	0.372	0.754	0.390	0.520
	I² (%)	78.73	94.80	96.88	97.79	78.17	78.45	85.48	93.33	94.39	99.53
ASQ-SS (H6)	r+	0.345	0.491	0.343	0.629	0.703	0.414	0.673	–	0.500	0.213
	I² (%)	94.28	96.98	95.06	94.29	97.42	95.41	82.67	–	86.62	89.35
Where, r+ = average correlations as per random effects model & I² = Test of heterogeneity

Source: The author.

Meta-Regression

In the primary research, we use multiple regression, which is a statistical method to determine the relationship between “covariates and outcome variables”. The subject is the unit of analysis for such studies, and moderators and outcome variables are determined for each subject. In the same way, this approach can be applied to meta-analysis with few changes. Studies will be the unit of analysis where covariates and outcomes are calculated for each such analysis. This technique, when applied to meta-analysis, is termed meta-regression. In the present study, the third research question is answered via conducting meta-regression. Regression is based on calculations between the effect size and covariates which can be categorical or continuous.¹¹⁹ In our study, there is a high level of heterogeneity; therefore, we are required to use a random-effects model for meta-regression also. The fixed model only takes into account the within-study error, that is, error by chance/sampling error. But the random-effects model is used when heterogeneity is quite high, and we measure the between study error, that is, the error which is not by chance or by sample size; it is there due to some other reasons. In the Tables 4–9, we will test three categorical variables for moderation.

Meta-Regression: Lecturers Quality–Student Satisfaction (LQ–SS)

Moderators used here are as follows: (a) type of student; (b) type of university; and (c) country.

Table 4 summarizes the effect of three categorical variables, that is, the type of student, type of university, and country of the study, on the lecturer’s quality and student satisfaction relationship. It also shows the coefficient, standard error (SE), upper and lower limits, z-values, and p-values for both the intercept and the covariates. The total variance in the null model, that is, lecturer’s quality and student satisfaction relationship without moderators, is 0.0789. When the “type of student” (Model 1) is introduced as a moderator in this relationship, we can see that no variance is explained by the moderator as R² = 0.00. Most importantly, the p-value of the model is 0.9195, which is insignificant as p > .05; therefore, there is no evidence that the type of student moderates this relationship. The type of university (Model 2) as a moderator explains 22% of the variance (R² = 0.22), but the p-value (p = .2202) is insignificant; therefore, this categorical variable does not moderate this relationship. The country of the study (Model 3) was also tested for moderation. Though it explained 29% of the variance (R² = 0.29), the p-value (p = .1085) was found to be insignificant. Therefore, there is no evidence that the country of the study moderates this relationship. In Model 4, all three categorical variables were taken together to test whether multiple covariates moderate this relationship or not. Though all three variables together explained 34% of the total variance (R² = 0.34) but p-value (p = .1653) is found to be insignificant. Therefore, three variables together do not moderate the relationship between the lecturer’s quality and student satisfaction. If we compare the null model with Models 1–4, we can see that there is no such significant change in the values of heterogeneity expressed by I² in each case. Therefore, none of the covariates moderates the relationship between the lecturer’s quality and student satisfaction.

Table 4.

Meta-Regression (LQ–SS)

Moderator Analysis (Random Effects Model)
Set	Covariate	Coefficient		S.E.		LL (95%)	UL (95%)	Z-value	2-sided p-value
Moderator/Covariate: Type of Student
	Intercept	0.5795		0.1765		0.2335	0.9255	3.28	0.0010
Type of Student	PG	–0.0961		0.2774		–0.6398	0.4476	–0.35	0.7289
Type of Student	UG	–0.0738		0.1953		–0.4565	0.3089	–0.38	0.7054
Model 1: p = 0.9195 Tau² = 0.0869, Tau = 0.2948, I² = 96.75 %, Q = 462.09 Total variance in true effects = 0.0789, R² = 0.00
Moderator/Covariate: Type of University
	Intercept	0.4177	0.1470		0.1296		0.7058	2.84	0.0045
Type of University	PR	0.1898	0.1680		–0.1394		0.5190	1.13	0.2584
Type of University	PU	–0.0329	0.1863		–0.3982		0.3323	–0.18	0.8598
Model 2: p = 0.2202 Tau² = 0.0616, Tau = 0.2481, I² = 95.13%, Q = 307.79 Total variance in true effects = 0.0789, R² = 0.22, Variance explained = 0.0174
Moderator/Covariate: Country
	Intercept	0.1619	0.1693		–0.1700		0.4937	0.96	0.3391
Country	EC	0.3355	0.2945		–0.2418		0.9127	1.14	0.2547
	NAC	0.1436	0.2431		–0.3327		0.6200	0.59	0.5546
	SAC	0.3163	0.2087		–0.0928		0.7254	1.52	0.1297
	SEAC	0.5101	0.2101		0.0984		0.9218	2.43	0.0152
	WAC	0.4898	0.2022		0.0936		0.8860	2.42	0.0154
Model 3: p = 0.1085 Tau² = 0.0561, Tau = 0.2369, I² = 94.29%, Q = 210.17 Total variance in true effects = 0.0789, R² = 0.29, Variance explained = 0.0228
Model 4: All Moderators (Type of student, university, and country) taken together gives p-value = 0.1653, R² = 0.34, Tau² = 0.0525, Tau = 0.2290, I² = 94.03%, Q = 133.91
Null Model (with no moderators/covariates): Tau² = 0.0789, Tau = 0.2810, I² = 96.33%, Q = 463.57, p=0.000, k (total number of studies) = 18

Source: The author.

Meta-Regression: Teaching Quality–Student Satisfaction (TQ-SS)

Moderators used here are as follows: (a) type of student; (b) type of university; and (c) country.

Table 5 summarizes the effect of three categorical variables, that is, the type of student, type of university, and country of the study, on the teaching quality and student satisfaction relationship. When the type of student (Model 1) is introduced as a moderator in this relationship, it can be seen that it explains only 5% of the variance (R² = 0.05), and the p-value (p = .1824) is insignificant. Therefore, there is no evidence that the type of student moderates this relationship. There is a similar case with the type of university (Model 2) with R² = 0.00 and p-value = .4305. The country of the study (Model 3) has a significant p-value (p = .0380) but explains only 5% of the variance (R² = 0.05). Therefore, we can determine that the country moderates this relationship but not fully, and other hidden variables moderate this relationship. The null model shows a total variance of 0.0635, and Model 4 shows that when all three categorical variables are introduced together as moderators, they explain 41% of the total variance (R² = 0.41) with a significant p-value of 0.0018. Table 5 also shows that there is not much change in heterogeneity levels (I²) in all four models when compared to the null model. Therefore, we can conclude that country of the study individually, and all three variables together moderate the relationship between teaching quality and student satisfaction.

Table 5.

Meta-Regression (TQ–SS)

Moderator Analysis (Random Effects Model)
Set	Covariate	Coefficient		S.E.		LL (95%)	UL (95%)	Z-value	2-sided p-value
Moderator/Covariate: Type of Student
	Intercept	0.2850		0.1250		0.0401	0.5299	2.28	0.0226
Type of Student	PG	0.2397		0.1767		–0.1066	0.5860	1.36	0.1748
Type of Student	UG	0.2589		0.1421		–0.0195	0.5374	1.82	0.0684
Model 1: p = 0.1824 Tau² = 0.0603, Tau = 0.2455, I² = 96.71 %, Q = 577.20 Total variance in true effects = 0.0635, R² = 0.05, Variance explained = 0.0032
Moderator/Covariate: Type of University
	Intercept	0.4171	0.1068		0.2078		0.6264	3.91	0.0001
Type of University	PR	0.1776	0.1461		–0.1087		0.4639	1.22	0.2241
Type of University	PU	0.0360	0.1509		–0.2597		0.3317	0.24	0.8113
Model 2: p = 0.4305 Tau² = 0.0765, Tau = 0.2766, I² = 97.67%, Q = 814.59 Total variance in true effects = 0.0635, R² = 0.00, Variance explained = 0.0000
Moderator/Covariate: Country
	Intercept	0.1936	0.1693		–0.1502		0.5374	1.10	0.2697
Country	EAC	0.4947	0.2485		0.0076		0.9817	1.99	0.0465
	EC	0.3525	0.2497		–0.1369		0.8420	1.41	0.1581
	NAC	0.1084	0.2155		–0.3141		0.5309	0.50	0.6150
	SAC	0.2341	0.1997		–0.1572		0.6254	1.17	0.2410
	SEAC	0.6167	0.2171		0.1911		1.0423	2.84	0.0045
	WAC	0.3993	0.3062		–0.2009		0.9995	1.30	0.1922
Model 3: p = 0.0380 Tau² = 0.0603, Tau = 0.2456, I² = 96.92%, Q = 486.82 Total variance in true effects = 0.0635, R² = 0.05, Variance explained = 0.0032
Model 4: All Moderators (Type of student, university, and country) taken together gives p-value = 0.0018, R² = 0.41, Tau² = 0.0376, Tau = 0.1938, I2 = 93.70%, Q = 174.73
Null Model (with no moderators/covariates): Tau2 = 0.0635, Tau = 0.2520, I2 = 97.50%, Q = 839.27, p=0.000, k (total number of studies) = 22

Source: The author.

Meta-Regression: Curriculum Quality–Student Satisfaction (CQ-SS)

Moderators used here are as follows: (a) type of student; (b) type of university; (c) country

Table 6 summarizes the effect of three categorical variables, that is, the type of student, type of university, and country of the study, on curriculum quality and student satisfaction relationship. When the type of student (Model 1) is introduced as a moderator in this relationship, it can be seen that it explains no variance, and the p-value is also insignificant (p = .2673). Therefore, there is no evidence of the type of student moderating this relationship. The type of university (Model 2) also shows very little variance; therefore, R² is set to be zero. Though the p-value (p = .0230) is significant, it does not moderate much of the relationship and explains very little variance; therefore, other moderators need to be searched. The country of study (Model 3) also does not explain much of the variance, and the p-value is also insignificant. When all three categorical variables are taken together, they explain only 8% (R² = 0.08) of the total variance (0.0479) at a significant p-value of .0191. These multi-covariates somewhat moderate this relationship but not fully, and that is why there is a need to look out for other significant moderators. One more thing can be observed from Table 6, and it is that heterogeneity levels remain approximately the same for the null model and the other four models with moderators.

Table 6

Meta-Regression (CQ–SS)

Moderator Analysis (Random Effects Model)
Set	Covariate	Coefficient		S.E.		LL (95%)	UL (95%)	Z-value	2-sided p-value
Moderator/Covariate: Type of Student
	Intercept	0.0671		0.2567		–0.4361	0.5703	0.26	0.7938
Type of Student	PG	0.4478		0.2867		–0.1142	1.0098	1.56	0.1183
Type of Student	UG	0.4216		0.2693		–0.1062	0.9494	1.57	0.1174
Model 1: p = 0.2673 Tau² = 0.0618, Tau = 0.2486, I² = 95.29 %, Q = 254.85 Total variance in true effects = 0.0479, R² = 0.00
Moderator/Covariate: Type of University
	Intercept	0.1665	0.1339		–0.0959		0.4288	1.24	0.2137
Type of University	PR	0.4208	0.1550		0.1171		0.7245	2.72	0.0066
Type of University	PU	0.2530	0.1886		–0.1166		0.6226	1.34	0.1797
Model 2: p = 0.0230 Tau² = 0.0502, Tau = 0.2241, I² = 94.18%, Q = 206.36 Total variance in true effects = 0.0479, R² = 0.00, Variance explained = 0.00
Moderator/Covariate: Country
	Intercept	0.6328	0.3021		0.0407		1.2250	2.09	0.0362
Country	NAC	–0.2223	0.3494		–0.9071		0.4626	–0.64	0.5247
	SAC	–0.1705	0.3314		–0.8200		0.4790	–0.51	0.6070
	SEAC	–0.0831	0.3503		0.7696		0.6034	–0.24	0.8125
	WAC	–0.2393	0.3507		–0.9267		0.4481	–0.68	0.4950
Model 3: p = 0.9393 Tau² = 0.0893, Tau = 0.2988, I² = 96.38%, Q = 276.26 Total variance in true effects = 0.0479, R² = 0.00, Variance explained = 0.00
Model 4: All Moderators (Type of student, university, and country) taken together gives p-value = 0.0191, R² = 0.08, Tau² = 0.0443, Tau = 0.2105, I² = 92.42%, Q = 79.21
Null Model (with no moderators/covariates): Tau² = 0.0479, Tau = 0.2190, I² = 95.39%, Q = 303.97, p = 0.000, k (total number of studies) = 15

Source: The author.

Meta-Regression: Physical Evidence Quality–Student Satisfaction (PEQ–SS)

Moderator analysis for three moderators are as follows: (a) type of student; (b) type of university; and (c) country.

Table 7 summarizes the effect of three categorical variables, that is, the type of student, type of university, and country of the study, on the physical evidence quality and student satisfaction relationship. When the type of student (Model 1) is introduced as a moderator in this relationship, it can be seen that it explains 39% of the variance (R² = 0.39) with a significant p-value (p = .0364). Therefore, it can be reported that the type of student moderates this relationship. In Models 2 and 3, the type of university and the country of the study, respectively, do not explain much of the variance, and p-values are insignificant. Accordingly, it can be said that both the categorical variables, that is, the type of the university and the country of the study, do not moderate this relationship. Model 4 in Table 7 shows that when all the three categorical variables are introduced together as moderators in the relationship, they explain 55% (R² = 0.55) of the total variance expressed by the null model (0.0767) with a significant p-value (p = .0022). There are no significant changes in the heterogeneity levels in Models 1–4 when compared to the null model. To conclude, the type of student individually and all the categorical variables together moderates the relationship between physical evidence quality and student satisfaction.

Table 7.

Meta-Regression (PEQ–SS)

Moderator Analysis (Random Effects Model)
Set	Covariate	Coefficient		S.E.		LL (95%)	UL (95%)	Z-value	2-sided p-value
Moderator/Covariate: Type of Student
	Intercept	0.2465		0.0793		0.0911	0.4019	3.11	0.0019
Type of Student	PG	0.2234		0.1775		–0.1244	0.5712	1.26	0.2080
Type of Student	UG	0.2438		0.0955		0.0566	0.0364	2.55	0.0107
Model 1: p = 0.0364 Tau² = 0.0466, Tau = 0.2159, I² = 94.08 %, Q = 422.09 Total variance in true effects = 0.0767, R² = 0.39, Variance explained = 0.0301
Moderator/Covariate: Type of University
	Intercept	0.3543	0.1133		–0.1322		0.5764	3.13	0.0018
Type of University	PR	0.1242	0.1377		–0.1456		0.3940	0.90	0.3669
Type of University	PU	0.0048	0.1668		–0.3221		0.3317	0.03	0.9770
Model 2: p = 0.5644 Tau² = 0.0865, Tau = 0.2941, I² = 96.69%, Q = 756.13 Total variance in true effects = 0.0767, R² = 0.00, Variance explained = 0.00
Moderator/Covariate: Country
	Intercept	0.4655	0.1651		0.1419		0.7891	2.82	0.0048
	EC	–0.1960	0.2103		–0.6081		0.2161	–0.93	0.3513
Country	NAC	–0.3463	0.2290		–0.7951		0.1025	–1.51	0.1305
	SAC	–0.0426	0.1801		–0.3956		0.3105	–0.24	0.8132
	SEAC	0.1724	0.1901		–0.2002		0.5449	0.91	0.3645
	WAC	–0.1188	0.1958		–0.5025		0.2649	–0.61	0.5440
Model 3: p = 0.0547 Tau² = 0.0485, Tau = 0.2203, I² = 93.53%, Q = 340.15 Total variance in true effects = 0.0767, R² = 0.37, Variance explained = 0.0281
Model 4: All Moderators (Type of student, university, and country) taken together gives p-value = 0.0022, R² = 0.55, Tau² = 0.345, Tau = 0.1858, I² = 90.50%, Q = 189.57
Null Model (with no moderators/covariates): Tau² = 0.0767, Tau = 0.2769, I² = 96.45%, Q = 761.29, p = 0.000, k (total number of studies) = 28

Source: The author.

Meta-Regression: Student Support Quality–Student Satisfaction (SSQ-SS)

Moderators used here are as follows: (a) type of student; (b) type of university; and (c) country.

Table 8 summarizes the effect of three categorical variables, that is, the type of student, type of university, and country of the study, on the physical evidence quality and student satisfaction relationship. When the type of student (Model 1) is introduced as a moderator in this relationship, it can be seen that it explains very little variance (R² = 0.15) and the p-value is also insignificant (p = .0501). Therefore, it does not moderate this relationship. The type of university (Model 2) also explains very little variance with an insignificant p-value (p = .7383). Hence, there is no evidence that the type of university plays the role of a moderator. Though the p-value of the country of the study (Model 3) is significant (p = .0240), it does not explain much of the total variance. But, when we compare the null model with Model 4 (multiple covariates), it is interesting to observe that all moderators together explain 67% (R² = 0.67) of the total variance (0.0732) with a significant p-value (p = .0000). Accordingly, it can be concluded that all the three categorical variables together moderate the relationship between student support quality and student satisfaction and explain 67% of the total variance. There is not much change in the heterogeneity levels of the four models with moderators when compared to the null model.

Table 8.

Meta-Regression (SSQ–SS)

Moderator Analysis (Random Effects Model)
Set	Covariate	Coefficient		S.E.		LL (95%)	UL (95%)	Z-value	2-sided p-value
Moderator/Covariate: Type of Student
	Intercept	0.2602		0.1283		0.0088	0.5116	2.03	0.0425
Type of Student	PG	0.4537		0.1989		0.0640	0.8435	2.28	0.0225
Type of Student	UG	0.2976		0.1483		0.0069	0.5883	2.01	0.0448
Model 1: p = 0.0501 Tau² = 0.0623, Tau = 0.2495, I² = 95.57%, Q = 361.31 Total variance in true effects = 0.0732, R² = 0.15, Variance explained = 0.0109
Moderator/Covariate: Type of University
	Intercept	0.6144	0.1382		0.3436		0.8852	4.45	0.0000
Type of University	PR	–0.1222	0.1667		–0.4489		0.2046	–0.73	0.4636
Type of University	PU	–0.1200	0.1782		–0.4693		0.2293	–0.67	0.5007
Model 2: p = 0.7383 Tau² = 0.0726, Tau = 0.2694, I² = 97.08%, Q = 547.98 Total variance in true effects = 0.0732, R² = 0.01, Variance explained = 0.0006
Moderator/Covariate: Country
	Intercept	0.4238	0.1656		0.0993		0.7483	2.56	0.0105
	EAC	0.5584	0.3525		–0.1324		1.2492	1.58	0.1131
	EC	–0.0120	0.3276		–0.6541		0.6301	–0.04	0.9708
Country	NAC	–0.1779	0.2327		–0.6340		0.2782	–0.76	0.4445
	SAC	0.1612	0.2188		–0.2675		0.5900	0.74	0.4611
	SEAC	0.3060	0.2029		–0.0917		0.7038	1.51	0.1316
	WAC	–0.5903	0.3329		–1.2427		0.0621	–1.77	0.0761
Model 3: p = 0.0240 Tau² = 0.0779, Tau = 0.2792, I² = 97.48%, Q = 476.46 Total variance in true effects = 0.0732, R² = 0.00, Variance explained = 0.00
Model 4: All Moderators (Type of student, university, and country) taken together gives p-value = 0.0000, R² = 0.67, Tau² = 0.0240, Tau = 0.1549, I² = 88.44%, Q = 69.21
Null Model (with no moderators/covariates): Tau² = 0.0732, Tau = 0.2706, I² = 97.55%, Q = 734.56, p = 0.000, k (total number of studies) = 19

Source: The author.

Meta-Regression: Administrative Support Quality–Student Satisfaction (ASQ-SS)

Moderators used here are as follows: (a) type of student; (b) type of university and (c) country.

Table 9 summarizes the effect of three categorical variables, that is, the type of student, type of university, and country of the study, on the physical evidence quality and student satisfaction relationship. When the type of student (Model 1) is introduced as a moderator in this relationship, it can be seen that it explains no variance (R² = 0.00) with an insignificant p-value (p = .8202). The type of university (Model 2) when introduced as a moderator it explains 39% (R² = 0.39) of the total variance with a significant p-value (p = .0046). The country of the study (Model 3) also explains 37% (R² = 0.37) of the variance with a significant p-value (p = .0051). It is interesting to observe that when all the three categorical variables are introduced as moderators in the relationship, they explain 80% (R² = 0.80) of the total variance (0.0889) with a significant p-value (p = .0000). In Model 4, heterogeneity has reduced from I² = 96.89% (null model) to I² = 85.58%. Hence, it can be concluded that all categorical together moderates the relationship between administrative service quality and student satisfaction and explain 80% of the variance.

Table 9.

Meta-Regression (ASQ–SS)

Moderator Analysis (Random Effects Model)
Set	Covariate	Coefficient		S.E.		LL (95%)	UL (95%)	Z-value	2-sided p-value
Moderator/Covariate: Type of Student
	Intercept	0.6083		0.2312		0.1551	1.0615	2.63	0.0085
Type of Student	PG	–0.2485		0.3969		–1.0265	0.5294	–0.63	0.5312
Type of Student	UG	–0.0707		0.2469		–0.5547	0.4133	–0.29	0.7747
Model 1: p = 0.8202 Tau² = 0.1009, Tau = 0.3177, I² = 97.20%, Q = 499.23 Total variance in true effects = 0.0889, R² = 0.00, Variance explained = 0.00
Moderator/Covariate: Type of University
	Intercept	0.3324	0.1683		0.0025		0.6624	1.97	0.0483
Type of University	PR	0.4069	0.1892		0.0360		0.7777	2.15	0.0315
Type of University	PU	0.0270	0.1914		–0.3482		0.4022	0.14	0.8879
Model 2: p = 0.0046 Tau² = 0.0544, Tau = 0.2333, I² = 94.80%, Q = 269.33 Total variance in true effects = 0.0889, R² = 0.39, Variance explained = 0.0345
Moderator/Covariate: Country
	Intercept	0.2312	0.1697		–0.1014		0.5638	1.36	0.1730
	EC	0.3181	0.2983		–0.2666		0.9028	1.07	0.2862
Country	NAC	–0.0168	0.2436		–0.4942		0.4605	–0.07	0.9449
	SAC	0.2086	0.1967		–0.1770		0.5942	1.06	0.2889
	SEAC	0.6386	0.2219		0.2036		1.0735	2.88	0.0040
	WAC	–0.5873	0.2240		0.1482		1.0264	2.62	0.0088
Model 3: p = 0.0051 Tau² = 0.0563, Tau = 0.2374, I² = 94.87%, Q = 214.60 Total variance in true effects = 0.0889, R² = 0.37, Variance explained = 0.0326
Model 4: All Moderators (Type of student, university, and country) taken together gives p-value = 0.0000, R² = 0.80, Tau² = 0.0174, Tau = 0.1320, I² = 85.58%, Q = 48.56
Null Model (with no moderators/covariates): Tau² = 0.0889, Tau = 0.2982, I² = 96.89%, Q = 513.89, p = 0.000, k (total number of studies) = 17

Source: The author.

Discussion

The first research question was regarding the important quality dimensions in the higher education sector. After the qualitative synthesis of the extant literature, six important dimensions were identified namely, lecturer’s quality (LQ), teaching quality (TQ), curriculum quality (CQ), physical evidence quality (PEQ), student support quality (SSQ), and administrative services quality (ASQ). To answer the second research question regarding the relationship between student satisfaction and different quality dimensions, authors carried out meta-analyses to find the combined effect size based on previous important studies. To our understanding, the current research will be the first one to examine the relationship between six higher education quality dimensions and student satisfaction based on a meta-analysis. All six dimensions were found to be positively associated with the major variable (student satisfaction). ASQ was found to have the highest correlation with student satisfaction, followed by SSQ, LQ, TQ, CQ, and PEQ. The PEQ was found to have the weakest relationship with student satisfaction as per the results of meta-analyses. These results are somewhat similar to the studies by Butt and Rehman,²⁰ Gruber et al.,¹²² Parahoo and Tamim,¹²³ Sakthivel et al.,¹²⁴ thereby suggesting that there might be similar relationships of quality dimensions with student satisfaction. The current analysis offers more novel details on the detection of significant moderating variables to answer the third research question. The type of students, that is, undergraduate (UG)and postgraduate (PG) students do not play the role of the moderator in 5 of the total six dimensions. It only moderates the relationship between physical evidence quality and student satisfaction. Undergraduate students show a higher correlation with PEQ–SS as compared to PG students. The type of university as a moderator, that is, public universities (PU) and private universities (PR), moderates two relationships. But it is not playing a significant role as a moderator in the curriculum quality and student satisfaction as it slightly moderates this relationship, and therefore, we need to search for more valid moderators. This variable also moderates the relationship between administrative service quality and student satisfaction and explains only 39% of the variance. The country of study as a moderator very slightly moderates the relationship between teaching quality and student satisfaction and student support quality and student satisfaction. It explains very little and insignificant variance in both the relationships. This categorical variable was also found to moderate the relationship between administrative service quality and student satisfaction and explains 37% of the variance. Still, there is a need to search for more valid and significant moderators.

Multiple Variables (Student, University, and Country) as Moderators

These categorical variables, when taken together, moderates all the relationships except LQ–SS. In the case of teaching quality and student satisfaction, all variables together play the role of moderator and explain 41% of the total variance. It also moderates the relationship between curriculum quality and student satisfaction (CQ–SS) but explains very little variance. Apart from this, it was interesting to observe that these categorical variables together moderate the relationship between physical evidence quality and student satisfaction (PEQ–SS); Student support quality and student satisfaction (SSQ–SS); and administrative service quality and student satisfaction (ASQ–SS). They explain 55% of the variance in PEQ–SS, 67% of the variance in SSQ–SS, and 80% of the variance in ASQ–SS. The current review requires further examination to determine and examine more moderators.

Strengths and Limitations

Strengths of the present meta-analytical review include an extensive search approach directed at both reported and unpublished research work, a robust review and selection procedures, and usage of defined standards¹²⁵ for designing, conducting, and reporting the results of the meta-analysis. Despite the inclusion of homogeneous studies, results still reported high levels of heterogeneity for all the associations. This suggests the existence of other moderators not included here. In comparison, our study used a relatively small number of studies for meta-analysis, which may hinder our ability to find major variations in moderator analyses. Even with this small sample of studies, however, we found proof of moderation for some of the very important correlations between different quality dimensions of higher education and student satisfaction. This study relied on a small sample of studies and was able to generate only three moderators via meta-regression that can moderate different relationships. Future researchers can take into account a greater number of studies and can determine more moderators. Besides, increasing the number of studies would make it possible to conduct meta-analysis again on those constructs which are still not confirmed. Finally, we realized that the study of student satisfaction in the context of higher education has a long way to go. Instead of opinionating the findings, this research tries to provoke thoughts. We also urge scholars to participate in novel studies in various fields and to consider other variables that have not been explored in the current study. This meta-analytic study may also be used as a point of reference with new findings observed by the researchers.

Theoretical and Practical Implications

As far as theoretical implications are concerned, we believe that the uniqueness of these results will add to the existing literature in the field of higher education. This meta-analytic study presents a comprehensive qualitative and quantitative synthesis of the available studies on the antecedents of student satisfaction. As opposed to the conventional studies, the meta-analytical study here adopts more clear descriptions of topics examined in studies analysing various higher learning institutions around the world. Our study overcomes the potential assumptions and shortcomings associated with current published work (scale of the study, type of study, statistical reliability, and soundness), thereby producing reliable estimations of the combined effect size of every association. Consolidated and robust findings allow us to identify the state of existing literature, that is, what we consider and do not consider concerning the measurement of student satisfaction with higher education quality. Also, this study has tried to address questions that have become prevalent among HEIs’ management due to the spread of new universities-for instance: What are the important factors that affect student satisfaction? and how powerful are these antecedents that cause student satisfaction? This can provide deeper insights into the study of student satisfaction in higher learning institutions.

The quality of education in Indian institutions of higher learning is a major concern. Better performance levels in Indian higher education are required due to the dynamic nature and demand for higher quality education, as well as a rise in competitive pressure. This can be achieved by a thorough understanding of students’ needs and expectations, and also the emphasis they place on factors discovered by the study such as teaching, teachers, administrative services, curriculum, student support services, and infrastructure. Therefore, concerning practical implications, the findings showed that university officials should make better resource allocation decisions to deliver better educational quality (teachers, teaching, and curriculum), including student support services, administrative services and infrastructure facilities. Also, institutions should place a bigger priority on improving their administrative services as it was found to be highly correlated with student satisfaction. Satisfaction with administrative departments is comprised of contentment with primary administrative services, ancillary administrative services, service providers’ attitudes, and general administration services offered by the institution. Implementing the above-mentioned practical implications might boost student satisfaction in academic institutions which would benefit not only students but also the institutions. The higher levels of student satisfaction will correspond to academic success which further has a bearing on the performance of students who are graduating and those who will be deployed in the workforce and society. For institutions, past research states that satisfied students recommend their institutions to prospective students. Also, student satisfaction leads to better student retention in higher education. Therefore, higher education administrators better allocate resources and devise methods to improve the quality of HEIs and enhance student satisfaction.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Forest Plots

References

Krause

. Addressing the wicked problem of quality in higher education: Theoretical approaches and implications. High Educ Res Develop. 2012 Jun 1; 31(3):285–297. DOI: 10.1080/07294360.2011.634381.

Vidovich

. That chameleon “quality”: The multiple and contradictory discourses of “quality” policy in Australian higher education. Discourse Stud Cultur Polit Educ. 2001 Aug 1; 22(2):249–261. DOI: 10.1080/01596300120072400.

Mishra

. Quality assurance in higher education: An introduction. Bangalore: National Assessment and Accreditation Council; 2007.

Qamar

. Status of quality in higher education varying perceptions. In: Higher education India: Issues Related to Expansion, Inclusiveness Quality and Finance. New Delhi: University Grants Commission; 2008. pp. 199–207.

Reeves

, Bednar

. Quality as symphony. Cornell Hotel Rest ADM Q. 1995 Jun; 36(3):72–79.

Procter

. Longman dictionary of contemporary English. London: Longman; 1978.

BSI-BS4778-2. Quality vocabulary—Part 2: Quality concepts and related definitions. London: Engineering360; 1991a.

Cheong

Cheng Y

, Ming

Tam W

. Multi–models of quality in education. Qual Assur Educ. 1997; 5(1):22–31. DOI: 10.1108/09684889710156558.

Sax

. Students as “customers.” On the Horizon. 2004; 12(4):158–160.

10.

Manyaga

. Standards to assure quality in tertiary education: The case of Tanzania. Qual Assur Educ. 2008 Apr 25; 16(2):164–180. DOI: 10.1108/09684880810868448.

11.

Filippakou

. The idea of quality in higher education: A conceptual approach. Discourse Stud Cultur Polit Educ. 2011; 32(1):15–28. DOI: 10.1080/01596306.2011.537068.

12.

Harvey

, Green

. Defining quality. Assess Eval High Educ. 1993; 18(1):9–34. DOI: 10.1080/0260293930180102.

13.

Freeman

. Strategic management: A stakeholder approach. Cambridge: Cambridge University Press; 1984.

14.

Nedwek

, Neal

. Performance indicators and rational management tools: A compatative assessment of projects in North America and Europe. Res High Educ. 1994 Jan 1; 35(1):75–103. DOI: 10.1007/BF02496663.

15.

Quinn

, Lemay

, Larsen

, Johnson

. Service quality in higher education. Total Qual Manag. 2009 Feb 1; 20(2):139–152. DOI: 10.1080/14783360802622805.

16.

Mahapatra

, Khan

. A framework for analysing quality in education settings. Eur J Eng Educ. 2007 May 1; 32(2):205–217. DOI: 10.1080/03043790601118606.

17.

Yingqiang

, Yongjian

. Quality assurance in higher education: Reflection, criticism, and change. Chin Educ Soc. 2016 Mar 3; 49(1–2):7–19. DOI: 10.1080/10611932.2016.1192382.

18.

Manzoor

. Measuring student satisfaction in public and private universities in Pakistan. Global J Manag Bus Res. 2013 Apr 9; 13(3):5–16.

19.

Allam

. Students’ perception of quality in higher education: An empirical investigation. Manag Sci Lett. 2018; 8:437–444. DOI: 10.5267/j.msl.2018.4.002.

20.

Butt

, Rehman

. A study examining the students satisfaction in higher education. Procedia Soc Behav Sci. 2010; 2(2):5446–5450. DOI: 10.1016/j.sbspro.2010.03.888.

21.

Hill

. Managing service quality in higher education: The role of the student as primary consumer. Qual Assur Educ. 1995 Sep 1; 3(3):10–21. DOI: 10.1108/09684889510093497.

22.

Crawford

. Total quality management. London: Committee of Vice-Chancellors & Principals of the Universities of the United Kingdom; 1991.

23.

Galloway

. Quality perceptions of internal and external customers: A case study in educational administration. TQM Mag. 1998; 10(1):20–26. DOI: 10.1108/09544789810197774.

24.

Voss

, Gruber

, Szmigin

. Service Quality in higher education: The role of student expectations. J Bus Res. 2007 Sep 1; 60(9): 949–959. DOI: 10.1016/j.jbusres.2007.01.020.

25.

Ahmad

. Evaluating student satisfaction of quality at international branch campuses. Assess Eval High Educ. 2015; 40(4):488–507. DOI: 10.1080/02602938.2014.925082.

26.

Guilbault

. Students as customers in higher education: Reframing the debate. J Mark High Educ. 2016 Jul 2; 26(2):132–142. DOI: 10.1080/08841241.2016.1245234.

27.

Athiyaman

. Linking student satisfaction and service quality perceptions: The case of university education. Eur J Mark. 1997; 31(7):528–540. DOI: 10.1108/03090569710176655.

28.

Wilkins

, Stephens

Balakrishnan MS

. Assessing student satisfaction in transnational higher education. Int J Educ Manag. 2013 Feb 15; 27(2):143–156. DOI: 10.1108/09513541311297568.

29.

Alves

. The measurement of perceived value in higher education: A unidimensional approach. Serv Ind J. 2011; 31(12):1943–60. DOI: 10.1080/02642069.2011.550042.

30.

Harvey

, Knight

. Transforming higher education. Buckingham: Society for Research into Higher Education and Open University Press; 1996.

31.

Rizkallah

, Seitz

. Understanding student motivation: A key to retention in higher education. Sci Ann Econ Bus. 2017 Mar 1; 64(1):45–57. DOI: 10.1515/saeb-2017-0004.

32.

Gift

, Bell–Hutchinson

. Quality assurance and the imperatives for improved student experiences in higher education: The case of the University of the West Indies. Qual High Educ. 2007 Jul 1; 13(2):145–157. DOI: 10.1080/13538320701629178.

33.

Nabaho

, Aguti

, Oonyu

. Unravelling quality in higher education: What say the students? Afr Educ Rev. 2019 Sep 3; 16(5):102–119. DOI: 10.1080/18146627.2016.1224600.

34.

Teeroovengadum

, Nunkoo

, Gronroos

, Kamalanabhan

, Seebaluck

. Higher education service quality, student satisfaction and loyalty: Validating the HESQUAL scale and testing an improved structural model. Qual Assur Educ. 2019 Oct 14; 27(4):427–445. DOI: 10.1108/QAE-01-2019-0003.

35.

Tiwari

, Kumar

. Promoting quality at institutions of higher education in India: Role of faculty and students. Discussant. 2018 Jan; 6(1):7–18.

36.

Al-Amri

, Mathew

, Zubairi

, Jani

. Optimal standards to measure the quality of higher education institutions in Oman: Stakeholders’ perception. SAGE Open. 2020; 10(3):1–11. DOI: 10.1177/2158244020947440.

37.

Asif

, Awan

, Khan

, Ahmad

. A model for total quality management in higher education. Qual Quant. 2013; 47(4):1883–1904. DOI: 10.1007/s11135-011-9632-9.

38.

Beerkens

, Udam

. Stakeholders in higher education quality assurance: Richness in diversity? High Educ Pol. 2017; 30:1–19.

39.

Cao

, Foster

, Yaoyuneyong

, Krey

. Hedonic and utilitarian value: The role of shared responsibility in higher education services. J Mark High Educ. 2019; 29(1):134–152. DOI: 10.1080/08841241.2019.1605439.

40.

Cavallone

, Ciasullo

, Manna

, Palumbo

. A tale of two stakeholders: Achieving excellence by merging quality expectations in Higher Education institutions. Stud High Educ. 2020 Mar; 13:1–5.

41.

Dicker

, Garcia

, Kelly

, Mulrooney

. What does “quality” in higher education mean? Perceptions of staff, students and employers. Stud High Educ. 2019 Aug 3; 44(8):1425–1441. DOI: 10.1080/03075079.2018.1445987.

42.

Iacovidou

, Gibbs

, Zopiatis

. An exploratory use of the stakeholder approach to defining and measuring quality: The case of a Cypriot higher education institution. Qual High Educ. 2009 Jul 1; 15(2):147–165. DOI: 10.1080/13538320902995774.

43.

Kettunen

, Kantola

. Strategic planning and quality assurance in the Bologna Process. Perspectives. 2007 Jul 1; 11(3):67–73. DOI: 10.1080/13603100701428205.

44.

Stupnisky

, Hall

, Pekrun

. Faculty enjoyment, anxiety, and boredom for teaching and research: Instrument development and testing predictors of success. Stud High Educ. 2019 Oct 3; 44(10):1712–1722. DOI: 10.1080/03075079.2019.1665308.

45.

Hill

, Lomas

, MacGregor

. Students’ perceptions of quality in higher education. Qual Assur Educ. 2003 Mar 1; 11(1):15–20. DOI: 10.1108/09684880310462047.

46.

Srikanthan

, Dalrymple

. A conceptual overview of a holistic model for quality in higher education. Int J Educ Manag. 2007 Apr 3; 21(3):173–193. DOI: 10.1108/09513540710738647.

47.

Tinto

. Leaving college: Rethinking the causes and cures of student attrition. Chicago: University of Chicago Press; 1987 Mar 24.

48.

Mahmud

, Khan

, Lima

. Measuring students’ satisfaction in the Universities of Bangladesh: An empirical analysis. Int J Mul Res Dev. 2018; 5(6):30–38.

49.

Ibáñez

, Hartmann

, Calvo

. Antecedents of customer loyalty in residential energy markets: Service quality, satisfaction, trust and switching costs. Serv Ind J. 2006 Sep 1; 26(6):633–50. DOI: 10.1080/02642060600850717.

50.

Niamatullah

, Xu

JGD

, Shafi

. Factors leading to students’ satisfaction in the higher learning institutions. J Educ Pract. 2015; 6(31):114–118.

51.

Chung Sea Law DC. Quality assurance in post–secondary education: The student experience. Qual Assur Educ. 2010 Sep 28; 18(4):250–70. DOI: 10.1108/09684881011079125.

52.

Muhammed

Malik E

, Danish

, Usman

. The impact of service quality on students’ satisfaction in higher education institutes of Punjab. J Manag Res. 2010 Jul 1; 2(2):1–11.

53.

Kotler

, Clarke

. Marketing for health care organizations. Hoboken, NJ: Prentice Hall; 1986.

54.

McDougall

GHG

, Levesque

. Customer satisfaction with services: Putting perceived value into the equation. J Serv Mark. 2000 Sep 1; 14(5):392–410. DOI: 10.1108/08876040010340937.

55.

Elliott

. Key determinants of student satisfaction. J Coll Stud Retention: Res Theory Prac. 2002; 4(3):271–279. DOI: 10.2190/B2V7-R91M-6WXR-KCCR.

56.

Parasuraman

, Berry

, Zeithaml

. Refinement and reassessment of the SERVQUAL scale. J Retail. 2002; 67(4):114.

57.

Parasuraman

, Zeithaml

, Berry

. A conceptual model of service quality and its implications for future research. J Mark. 1985 Sep; 49(4):41–50. DOI: 10.1177/002224298504900403.

58.

Cronin

, Taylor

. A reexamination and extension. J Mark. 1992; 56(3):55–68.

59.

Clemes

, Gan

CEC

, Kao

. University student satisfaction: An empirical analysis. J Mark High Educ. 2008; 17(2):292–325. DOI: 10.1080/08841240801912831.

60.

, Wearn

. A higher education TQM excellence model: HETQMEX. Qual Assur Educ. 1996 Jun 1; 4(2):35–42. DOI: 10.1108/09684889610116049.

61.

Abdullah

. HEdPERF versus SERVPERF: The quest for ideal measuring instrument of service quality in higher education sector. Qual Assur Educ. 2005; 13(4):305–328. DOI: 10.1108/09684880510626584.

62.

Annamdevula

, Bellamkonda

, Development of HiEdQUAL for Measuring Service. Quality in Indian higher education sector. Int J Innov Manag Tech. 2012 Aug 1; 3(4):412.

63.

Ashraf

, Osman

AZR

, Ratan

SRA

. Determinants of quality education in private universities from student perspectives: A case study in Bangladesh. Qual Assur Educ. 2016; 24(1):123–138. DOI: 10.1108/QAE-09-2013-0040.

64.

Kärnä

, Julin

. A framework for measuring student and staff satisfaction with university campus facilities. Qual Assur Educ. 2015 Feb 2; 23(1):47–66. DOI: 10.1108/QAE-10-2013-0041.

65.

Wong

, Tong

, Wong

. The relationship between institution branding, teaching quality and student satisfaction in higher education in Hong Kong. J Mark HR. 2017; 4(1):169–188.

66.

Martirosyan

. An examination of factors contributing to student satisfaction in Armenian higher education. Int J Educ Manag. 2015 Mar 9; 29(2):177–191. DOI: 10.1108/IJEM-09-2013-0143.

67.

Mai

. A comparative study between UK and US: The student satisfaction in higher education and its influential factors. J Mark Manag. 2005 Aug 1; 21(7–8):859–878. DOI: 10.1362/026725705774538471.

68.

Gibbs

. Dimensions of quality. York: Higher Education Academy; 2010.

69.

Azma

. Qualitative Indicators for the evaluation of universities performance. Procedia Soc Behav Sci. 2010; 2(2):5408–5411. DOI: 10.1016/j.sbspro.2010.03.882.

70.

Prakash

. Quality in higher education institutions: Insights from the literature. TQM J. 2018 Oct 8; 30(6):732–748. DOI: 10.1108/TQM-04-2017-0043.

71.

Gul

, Jan

, Shah

. The impact of service quality on students satisfaction in higher education institutes of Khyber Pakhtunkhwa. Rev Econ Dev Stud. 2019; 5(1):217–224. DOI: 10.26710/reads.v5i1.589.

72.

Abu

Hasan HF

, Ilias

, Rahman

, Razak

. Service Quality and student satisfaction: A case study at private higher education institutions. Int Bus Res. 2009; 1(3):163–175.

73.

Ijaz

, Irfan

, Shahbaz

, Awan

, Sabir

. An empirical model of student satisfaction: Case of Pakistani public sector business schools. J Qual Tech Manag; 2011; 7(2): 91–114.

74.

Kajenthiran

, Karunanithy

. Service quality and student satisfaction: A case study of private external higher education institutions in Jaffna, Sri Lanka. J Bus. 2015; 1(2):46–64.

75.

Wei

Chuah C

, Ramalu

Sri S

. Students satisfaction towards the university: Does service quality matters? Int J Educ. 2011; 3(2):1–5.

76.

Gamage

, Suwanabroma

, Ueyama

, Hada

, Sekikawa

. The impact of quality assurance measures on student services at the Japanese and Thai private universities. Qual Assur Educ. 2008 Apr 25; 16(2):181–198. DOI: 10.1108/09684880810868457.

77.

DeShields

, Kara

, Kaynak

. Determinants of business student satisfaction and retention in higher education: Applying Herzberg’s two–factor theory. Int J Educ Manag. 2005 Mar 1;19(2):128–139. DOI: 10.1108/09513540510582426.

78.

Aly

, Akpovi

. Total quality management in California public higher education. Qual Assur Educ. 2001; 9(3):127–131. DOI: 10.1108/09684880110399077.

79.

Al-Sheeb

, Hamouda

AMS

, Abdella

. Investigating determinants of student satisfaction in the first year of college in a public university in the state of Qatar. Educ Res Int. 2018; 1–14. DOI: 10.1155/2018/7194106.

80.

Asif

, Searcy

. A composite index for measuring performance in higher education institutions. Int J Qual Rel Manag. 2014; 31(9):983–1001. DOI: 10.1108/IJQRM-02-2013-0023.

81.

Augustine

, Tanui

, Kalai

. Educational service quality and students’ satisfaction in public universities in Kenya. Int J Educ Soc Sci. 2016; 3(10):37–48.

82.

Darawong

, Sandmaung

. Service Quality enhancing student satisfaction in international programs of higher education institutions: A local student perspective. J Mark High Educ. 2019; 29(2):268–283. DOI: 10.1080/08841241.2019.1647483.

83.

Farahmandian

. Perceived service quality and student satisfaction in higher education. IOSR JBM. 2013; 12(4):65–74. DOI: 10.9790/487X-1246574.

84.

Hameed

, Amjad

. Students’ satisfaction in higher learning institutions: A case study of COMSATS Abbottabad, Pakistan. Iran J Manag Stud. 2011 Mar 1; 4(1):63–77.

85.

Hanaysha

, Abdullah

, Warokka

. Service quality and students’ satisfaction at higher learning institutions: The competing dimensions of Malaysian universities’ competitiveness. JSAR. 2011: 1–10. DOI: 10.5171/2011.855931.

86.

Hanssen

, Solvoll

. The importance of university facilities for student satisfaction at a Norwegian University. Facilities. 2015 Oct 5; 33(13/14):744–759. DOI: 10.1108/F-11-2014-0081.

87.

, Chung

. Teaching quality, learning satisfaction, and academic performance among hospitality students in Taiwan. World J Educ. 2014; 4(5):11–20. DOI: 10.5430/wje.v4n5p11.

88.

Law

, Cheung

. Measuring student satisfaction in distance education in Hong Kong. Bus Inf. 2013; 7(9):443–470.

89.

Lim

Kok Seng E

, Pei

Ling T

. A statistical analysis of education service quality dimensions on business school students’ satisfaction. Int Educ Stud. 2013; 6(8):136–146. DOI: 10.5539/ies.v6n8p136.

90.

Mukhopadhyay

. Quality in teacher education: Various parameters and effective quality management. IOSR J Human Social Science. 2014; 19(2):66–71. DOI: 10.9790/0837-19216671.

91.

Mwiya

, Bwalya

, Siachinji

, Sikombe

, Chanda

, Chawala

. Higher education quality and student satisfaction nexus: Evidence from Zambia. Creat Educ. 2017 Jun 8; 8(7):1044–1068. DOI: 10.4236/ce.2017.87076.

92.

Ngamkamollert

, Ruangkanjanases

. Factors influencing foreign students’ satisfaction toward international program in Thai universities. Int J Inf Educ Tech. 2015 Mar 1; 5(3):170–178. DOI: 10.7763/IJIET.2015.V5.497.

93.

O’Driscoll

. What matters most: An exploratory multivariate study of satisfaction among first year hotel/hospitality management students. Qual Assur Educ. 2012 Jul 6; 20(3):237–258. DOI: 10.1108/09684881211240303.

94.

Osman

, Saputra

. A pragmatic model of student satisfaction: A viewpoint of private higher education. Qual Assur Educ. 2019 Apr 1; 27(2):142–165. DOI: 10.1108/QAE-05-2017-0019.

95.

Styron

Jr . Student satisfaction and persistence: factors vital to student retention. Res High Educ J. 2010 Mar 1; 6:1.

96.

Sharabati

, Alhileh

, Abusaimeh

. Effect of service quality on graduates’ satisfaction. Qual Assur Educ. 2019 Jul 1; 27(3):320–337. DOI: 10.1108/QAE-04-2018-0035.

97.

Sheikh

Ali AY

, Mohamed

. Service quality provided by higher education institutions in Somalia and its impact on student satisfaction. Eur J Bus Manag. 2014; 6(11):143–148.

98.

Singh

, Singla

. Stakeholders satisfaction regarding service quality in higher management education. Int J Manag Stud. 2018; 5(3[6]):2231–2356. DOI: 10.18843/ijms/v5i3(6)/01.

99.

Songsathaphorn

, Chen

, Ruangkanjanases

. A study of factors influencing Chinese students’ satisfaction toward Thai universities. J Econ Bus Manag. 2014; 2(2):105–111. DOI: 10.7763/JOEBM.2014.V2.107.

100.

Tessema

, Ready

, Yu

. Factors affecting college students’ satisfaction with major curriculum: Evidence from nine years of data. Int J Hum Soc Sci. 2012 Jan; 2(2):34–44.

101.

Trang

, Kha

, Duyen

, Linh

. Research on factors affecting the postgraduate students’ satisfaction in the quality of training services in accounting at the training institutions in Hanoi. Am J Educ Res. 2018 May 4; 6(5):512–518.

102.

Ushantha

CRA

, Kumara

SPA

. A quest for service quality in higher education: Empirical evidence from Sri Lanka. Serv Mark Q. 2016 Apr 2; 37(2):98–108.

103.

Weerasinghe

IMS

, Fernando

RLS

. Critical factors affecting students’ satisfaction with higher education in Sri Lanka. Qual Assur Educ. 2018 Feb 5; 26(1):115–130. DOI: 10.1108/QAE-04-2017-0014.

104.

Zopiatis

, Theodosiou

, Constanti

. Quality and satisfaction with culinary education: Evidence from Cyprus. J Hosp Tour Educ. 2014 Apr 3; 26(2):87–98. DOI: 10.1080/10963758.2014.900383.

105.

Tranfield

, Denyer

, Smart

. Towards a methodology for developing evidence–informed management knowledge by means of systematic review. Br J Manag. 2003 Sep; 14(3):207–22. DOI: 10.1111/1467-8551.00375.

106.

Tarí

. Research into quality management and social responsibility. J Bus Ethics. 2011 Sep; 102(4):623–638. DOI: 10.1007/s10551-011-0833-x.

107.

Borenstein

(Ed.). Introduction to Meta-Analysis. Hoboken: John Wiley & Sons; 2009.

108.

Suurmond

, van Rhee

, Hak

. Introduction, comparison, and validation of Meta–Essentials: A free and simple tool for meta–analysis. Res Synth Methods. 2017 Dec; 8(4):537–553. DOI: 10.1002/jrsm.1260, PMID 28801932.

109.

Hunter

, Schmidt

. Methods of meta-analysis: Correcting error and bias in research findings. SAGE Publications; 2004 Apr 19.

110.

Glass

. Primary, secondary, and meta-analysis of research. Educ Res. What is quality in higher education? 1976 Nov; 5(10):3–8. DOI: 10.3102/0013189X005010003.

111.

Davis

, Mengersen

, Bennett

, Mazerolle

. Viewing systematic reviews and meta-analysis in social research through different lenses. Berlin: SpringerPlus. 2014 Dec; 3(1):511. DOI: 10.1186/2193-1801-3-511, PMID 25279303.

112.

Ryan

. Meta-analysis: Cochrane consumers and communication review group; 2016. http://cccrg.cochrane.org

113.

Chong-ho

. Meta-analysis and effect size; 2010. http://www.creative-wisdom.com/teaching/WBI/es.shtml

114.

Schäfer

, Schwarz

. The meaningfulness of effect sizes in psychological research: Differences between sub-disciplines and the impact of potential biases. Fron Psyc Front Psychol. 2019 Apr 11; 10:813. DOI: 10.3389/fpsyg.2019.00813, PMID 31031679.

115.

Cohen

. The statistical power of abnormal-social psychological research: A review. J Abnorm Soc Psychol. 1962; 65:145–153. DOI: 10.1037/h0045186, PMID 13880271.

116.

Cohen

. Set correlation and contingency tables. Appl Psychol Measure. 1988; 12(4):425–434. DOI: 10.1177/014662168801200410.

117.

Cohen

. Statistical power analysis. Curr Dir Psychol Sci. 1992; 1(3):98–101. DOI: 10.1111/1467-8721.ep10768783.

118.

Borenstein

, Hedges

, Higgins

JPT

, Rothstein

. A basic introduction to fixed-effect and random-effects models for meta-analysis. Res Synth Methods. 2010; 1(2):97–111. DOI: 10.1002/jrsm.12, PMID 26061376.

119.

Fletcher

. What is heterogeneity, and is it important? BMJ. 2007; 334(7584):94–96. DOI: 10.1136/bmj.39057.406644.68, PMID 17218716.

120.

Hak

, van Rhee

, Suurmond

. How to interpret results of meta-analysis. 2016 Mar. DOI: 10.2139/ssrn.3241367.

121.

Borenstein

, Hedges

, Higgins

, Rothstein

Comprehensive meta-analysis Version 3. Englewood, NJ: Biostat; 2014.

122.

Gruber

, Fuß

, Voss

, Gläser–Zikuda

. Examining student satisfaction with higher education services: Using a new measurement tool. Int J Pub Sec Manag. 2010 Mar 2; 23(2):105–123. DOI: 10.1108/09513551011022474.

123.

Parahoo

, Tamim

. Determinants of student satisfaction in higher education: An empirical study in Dubai. Int J Serv Econ Manag. 2012 Jan 1; 4(4):282–297. DOI: 10.1504/IJSEM.2012.050949.

124.

Sakthivel

, Rajendran

, Raju

. TQM implementation and students’ satisfaction of academic performance. TQM Mag. 2005 Dec 1; 17(6):573–589. DOI: 10.1108/09544780510627660.

125.

Moher

, Shamseer

, Clarke

, Ghersi

, Liberati

, Petticrew

, Shekelle

, Stewart

, PRISMA-P Group. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev. 2015 Dec; 4(1):1. DOI: 10.1186/2046-4053-4-1, PMID 25554246.