Developing a Comprehensive Higher Education Service Quality Model in Ethiopian Context

The existing literature lacks a comprehensive service quality model to measure higher education service quality, particularly from the Ethiopian higher education context. The main purpose of this study was to address this literature gap. To serve the study purpose, the existing generic service quality models mainly designed for market and higher education sectors were thoroughly reviewed. Through an extensive literature review, a comprehensive higher education service quality model was developed. The content validity and reliability of the instrument were empirically tested in three public universities representing 30% of universities located in the Southern part of Ethiopia. Five experienced experts who were selected from three different sample universities judged the content validation of the instrument. The experts’ judgment in each item was quantitatively calculated using the Content Validity Index. The pilot study was conducted on 40 regular undergraduate students who were purposively selected from the different University departments to further analyze the reliability of the instrument while the internal consistency of the instrument was checked at Cronbach’s alpha coefficient of 0.5 using Statistical Package for Social Sciences v.20. The findings of the study revealed that the newly developed model was valid and internally consistent. Since the present model is comprehensive and empirically tested in the Ethiopian higher education context, the model better measures the Ethiopian higher education service quality. However, further studies should be conducted using Exploratory Factor Analysis (EFA) to strengthen the present findings.


Introduction
The Higher Education Institutions (HEIs) nowadays have been influenced by marketization.Different push factors affect HEIs to apply marketing practices in their institutions such as increased competition in higher education, decreased government funding, and the increased cost of education (Guilbault, 2016).For instance, the present government of Ethiopia is also designing different self-administration strategies and policies for public universities.In fact, in the 2023 academic year, Addis Ababa University (one of the oldest and largest universities in Ethiopia) has been officially differentiated as a self-administered university.The remaining first-generation public universities will be self-administered universities in the coming years.The self-administering system of universities is expected to increase competition among HEIs because universities that provide better services are expected to attract many students as compared with low service providers.In educational settings, there are various definitions for the term 'educational service quality' (Schneider & White, 2004); the most recent refers to the difference between students' expectations and their actual perceptions of service delivery (Silva et al., 2017).In today's dynamic higher education sector, educational service quality is considered an important element to ensuring the quality of higher education.According to Onditi and Wechuli (2017), education service quality is a key determinant to measure the quality of higher education performance while Malik et al. (2010) pointed out that it is fundamental and an important parameter of educational excellence in general and higher education in particular.
While the study of Raju and Bhaskar (2017) concluded that educational service quality is essential for a high standard of education and a good image of HEIs, Al-Dulaimi (2016) assert the need for universities to achieve objectives and promote the effectiveness of the education system.Hence, this study also contends that measuring higher education service quality is one of the key elements to ensure the quality in the institutions.The main purpose of this study was to develop a comprehensive Higher Education Service Quality Model [HESQM] to measure Ethiopian higher education service quality.

Background and Research Context
The HEIs across the world have been experiencing a quality assurance process (Hasbullah & Yosuff, 2017) that puts service quality as one of the most important elements in ensuring the quality of higher education (Đonlagić & Fazlić, 2015).In Ethiopia, the last fifteen years saw rapid expansion of HEIs in different parts of the country (Kedir, 2009).The number of public universities increased from 2 in 1991 (Tesfaye, 2011) to 36 in 2015(MoE, 2016).At present, there are 49 public universities and 128 accredited private HEIs in the country (Ministry of Education [MoE], 2018).The most recent data released by Ethiopian MoE (2018) show that the total number of undergraduate students who are attending their education in different Ethiopian public universities reached more than 825, 003 (520,177 males and 304, 826 females) in 2018.
The radical expansion of higher education in Ethiopia had mainly affected the quality of higher education in general and the provision of necessary educational services in particular.
In recognizing these challenges, the government introduced various quality improvement initiatives and reform programs in higher education (Ayalew et al., 2009).Though different quality improvement initiatives and reform programs have been implemented to improve the quality of Ethiopian higher education, the recent empirical studies found that these initiatives and reform programs have brought a low impact on Ethiopian higher education quality (Alemayehu & Solomon, 2017;MoE, 2018).Therefore, designing a comprehensive higher education service quality model from the Ethiopian higher education context is imperative to (1) measure the quality of higher education, (2) take appropriate measures to improve the quality of Ethiopian higher education and (3) identify areas that need further attention.

Existing Service Quality Models
The higher education service quality measures are still underdeveloped because their measurements are almost adopted from models designed for business sectors (Marimuthu & Ismail, 2012).In the literature, many prior studies have employed generic service quality measures, especially Service Quality (SERVQUAL) and Service Performance (SERPERF) models to examine higher education quality (Tuan, 2017;Wei & Ramalu, 2011;Manea & Iatagan, 2015;Mwiya et al., 2017).Among the generic service quality models, SERVQUAL is the most popular service quality measure.The modified SERVQUAL model, which describes the difference between customer expectations about the service provided and their perception after taking the service, consists of five dimensions; tangibles, reliability, responsiveness, assurance and empathy and includes a total of 22 items (Parasuraman et al., 1985) (see appendix table A1).
Although SERVQUAL is well-known model in business and higher education sectors, the SERVQUAL model is not free from scholarly critics.According to Cronin and Taylor (1992), customer's expectations before experiencing the service is difficult to conceptualize.They further claim that there is little evidence, either theoretical or empirical, to support the notion of the "expectations minus performance" gap as a basis for measuring service quality.This situation led Cronin and Taylor (1992) to look for another alternative model.Cronin and Taylor (1992) modified the SERVQUAL model and proposed another model called Service Performance [SEVPERF] that measures performance only.The five SERPERF model dimensions are worded the same as SERVQUAL but do not repeat the set of statements as expectation items.Because of the complex nature of higher education service quality, many higher educations service quality researchers (e.g., Abdullah, 2006a;Annamdevula & Bellamkonda, 2012) argue against employing generic service quality models in the higher education sector.
Research on service quality in higher education settings has generally revolved around major two issues: (1) measurement method and (2) dimensions or facets of higher education service quality (Yildiz, 2012;Kontic, 2014).The nature of higher education services is composed of multi-dimensional constructs or variables.As a result, different scholars propose different models to measure higher education service quality (Cerri, 2012).For example, Abdullah (2006a) argues that the use of existing marketing sector service quality models may not be applicable in all service sectors, particularly higher education and proposes a new measure of higher education service quality called Higher Education Performance [HEdPERF].As an adopted model from Cronin and Taylor (1992) performance-only or Service Performance [SEVPERF] approach, the HEdPERF model has six dimensions, namely, non-academic aspects, academic aspects, reputation, access, program issues and understanding.In his later work, Abdullah (2006b) modified the existing HEdPERF instrument by comparing three service quality measurements, such as HEdPERF, SERVPERF and the moderating scale of HEdPERF-SERVPERF within a higher education setting.In the modified HEdPERF instrument, Abdullah (2006b) indicated five distinct factors, namely, non-academic aspects, academic aspects, reputation, access and programme issues by excluding understanding from the former HEdPERF instrument.The modified HEdPERF consists of 41 items; 13 items were taken from SERVPERF, and the remaining 28 items were developed via literature review (see appendix table A2).Researchers like Annamdevula and Bellamkonda (2012) developed another new higher education service quality measure called the Higher Education Quality [HiEdQUAL] model.Annamdevula and Bellamkonda (2012) also contend that since higher education services are a complex combination of various factors, using the existing generic service quality measures such as SERVQUAL, SERPERF and Evaluated Performance (EP) is not easy to apply in higher education sectors.Therefore, they introduced Higher Education Quality [HiEdQUAL] model which consists of 27 items, with five dimensions, viz., teaching and course content, administrative services, academic facilities, campus infrastructure and support services (see appendix table A3).
Other higher education service quality researchers like Teeroovengadum et al. (2016) argue that although the existing higher education service quality models are empirically tested in higher education, none of them integrated the notion of transformative quality in the development of service quality models.Teeroovengadum, et al. (2016) introduced the new higher education service quality measure called Higher Education Quality [HESQUAL], which consists of five primary dimensions and nine-sub dimensions such as administrative quality (attitude and behavior, administrative processes), support facilities quality, core educational quality (curriculum, attitude and behavior, competence and pedagogy), transformative quality and physical environment quality (support infrastructure, learning setting and general infrastructure) and included a total of 48 items (see appendix table A4 & A5).

Limitations of the existing service quality models and rationale
The literature review shows that both generic service quality models designed for business sectors as well as service quality models principally developed for the higher education sector have attempted to measure higher education service quality.However, this paper argues that using a single model either a generic service quality model or a service quality model mainly designed to measure higher education, is not comprehensive and responsive enough to measure Ethiopian higher education service quality context.Each generic as well as higher education service quality model consists of too short dimensions and items or constructs that do not include all basic services provided by Ethiopian universities.It is a fact that, Ethiopian higher education does not work in isolation; however, in some existing higher education service quality models; many services that have been provided in Ethiopian higher education are missing.For example, Ethiopian higher education provides free dormitory and cafeteria services for undergraduate regular students and in some universities including postgraduate students.Free dormitory and cafeteria services are not common services in most countries.In addition, the literature lacks study that developed a comprehensive model from the Ethiopian higher education context.Table 1 and   Table 2 shortly summarise the strengths and weaknesses of the existing generic service quality models and models designed for higher education service quality.

Conceptual framework
The existing literature lacks a comprehensive service quality model that measures higher education service quality, particularly from the Ethiopian higher education context.
In this study, a combination of the existing generic service quality models mainly designed for business sectors, such as SERVQUAL, SERVPERF and models particularly designed for higher education, namely, HEdPERF, HiEdQUAL, HESQ were thoroughly reviewed and adapted to develop the comprehensive model.In the newly developed HESQM, seven major dimensions and nine sub-dimensions were identified.The major and sub-dimensions include academic service quality (academic staff's attitude and behavior, academic staff's competence, academic facilities and resources, academic program issues, quality of instructional practices, quality of library services), administrative services (administrative staffs' attitude and behavior, quality of administrative processes and procedures, administrative staffs' competence), quality of general infrastructure, quality of support services and facilities, quality of students' welfare services, university access to students and university reputation (see appendix tables A6-A10).

Method
This study intensively reviewed research documents that are relevant to t h e research objectives through different web searches.The main web searches employed in the review mainly include the University online library and Google Scholar.Through the University online library, researchers accessed pertinent research materials from well-known databases such as science direct Elsevier, EBSCO, Project Muse, Springer Link, SAGE Journals and the Francis and Taylor Group.In the review, except for very few publications, the most recent studies (published between 2005 and 2022) were thoroughly reviewed totaling 90 peer-reviewed past studies focused on higher education service quality.Through an extensive literature review, a comprehensive Higher Education Service Quality Model [HESQM] model was developed intended to measure Ethiopian higher education service quality.
The content validity of the questionnaire was conducted to check whether the designed items measure or represent the intended study objective or content area and it is mainly carried out by experts (Creswell, 2012).Furthermore, the reliability of the questionnaire was also checked to ensure the internal consistency of the instruments.Figure 1 summarises the development of a new higher education service quality model.3 summarises the profile of content validity evaluators.

Qualitative results of instruments content validation
The oral discussion with the instruments' content validity evaluators generated comments to further improve the quality of the tool.The main comments include length of the directions, language clarity, repetition of items, items that are not reflecting main and subtitles, proposing new items to be added, consistency and order of items, unnecessary use of conjunctions and length of the directions.After considering the comments, the researchers thoroughly re-read all items and identified some potential errors.The consistency and order of items, language clarity problems, and unnecessary use of conjunctions in items and length of the directions were corrected.Three new items that pertain to students with disability services were also constructed and added under the general campus infrastructure dimension.

Quantitative analysis of instruments content validity
On top of the inputs received via oral discussion with content validity evaluators, the content validity of the questionnaire was also statistically checked.The five content validity evaluators selected from three sample universities rated the relevance of each educational service quality item.The content validity was calculated using the Content Validity Index [CVI].The acceptance rate of content validity rate varies according to the number of content validity evaluators involved.For example, if two experts are involved, the acceptable value of CVI is 0.80, whereas total number of nine experts, the acceptable value is 0.78.Table 4 summarises scholars' recommendations about to the acceptance range of the content validity index of the items.

Calculating content validity index
The content validity index can be measured in two ways: (1) Item Level Content Validity Index [I-CVI] and (2) Scale-Level Content Validity Index [S-CVI].The I-CVI can be mathematically calculated as the number of experts' agreements in each item divided by the total number of expert raters.For example, the first item, 'My Lecturers/Professors have a positive attitude towards students', all five content validity assessors ranked as 4=Highly Relevant [HR].Thus, the I-CVI of this item was calculated as 5÷5=1.As presented in table 5, the result of CVI was evaluated according to Denise et al. (2007).If three to five experts are involved in the content validation of the questionnaire, the acceptable values of CVI should be 1.On the other hand, S-CVI is calculated by adding each acceptable I-CVI result or total relevant divided by the total number of items.In this study, items yielded acceptable (1) I-CVI were included to calculate the S-CVI and items which scored 1 I-CVI were retained and items which scored 0.4 I-CVI were deleted from the questionnaire.

Quantitative results of instruments content validation
A total of 136 items intended to measure education service quality were administered to content validity evaluators.As shown in table 5, I-CVI was computed for each item with 126 items yielding I-CVI (1) acceptable value 1 according to Denise et al.
(2007) rule of thumb while 11 items that yielded a low I-CVI (0.4) were deleted from the questionnaire.After making intensive revisions on each item, a total of 126 items were retained.In addition, based on the written comments forwarded by content validity evaluators, the other three new items that reflect students with disability services were added under the quality of general infrastructure subsection.Therefore, a total of 129 items intended to measure Ethiopian Higher Education Service Quality were ready for pilot testing.Table 6 summarises the item's revision based on I-CVI result.

Pilot study
Upon the completion of the instruments' content validation, the reliability of the questionnaire was checked.A pilot study was conducted at Hawassa University.There is no consistent number of participants for the pilot study; different scholars suggest a different number of participants for the pilot study.For example, Gay and Mills (2012) and Johnson and Christensen (2014) suggest a minimum of 5 to 10 groups of people for piloting the instrument while Saunders (2009) argues that the number of people with whom to pilot depends on the research question(s), the size of the research project, the time and money resources available, and how questionnaire was initially designed.Considering these arguments, the pilot study was conducted to forty (40) regular undergraduate students at Hawassa University; 25 males and 15 females.The participants for the pilot study were purposefully selected from different departments.The researchers with the research assistants properly disseminated the questionnaire to the volunteer participants with a brief orientation about the purpose of the study and the process of filling out the questionnaire.Since the questionnaire sections were somewhat lengthy, one-week period was given to the participants to return the questionnaire.
All the participants filled out and returned the questionnaire.
The questionnaire in English with a brief introduction and instructions has two parts.
The first part contains respondents' demographic information and general directions on how to fill out the questionnaire.The second part includes items about the measurement of educational service quality.

Reliability Analysis
The reliability of each main dimension and sub-dimension were checked.The reliability result was assessed according to George and Mallery's (2003) rule of digits: > 0.90 = Excellent; 0.80 -0.89 = Good; 0.70 -0.79 = Acceptable; 0.60 -0.69 = Questionable; 0.50 -0.59 = Poor; < 0.50 = Unacceptable.As summarized in table 7, the reliability results indicate that the majority of the facets of education service quality yielded an acceptable alpha value, except for thirteen items that explain the attributes of educational service quality.After making all these corrections, a total of one hundred sixteen (116) items were retained.Table 6 summarises the reliability result of the questionnaire.

Summary of Major Findings and Discussion
In the previous studies, many international and local researchers have used a single service quality model to measure higher education service quality.Among the local studies, Solomon (2012) employed the SERVQUAL Model to examine stakeholders' perception of Ethiopian higher education service quality while Gelilawit (2019) employed SERPERF to examine the effect of service quality on students' satisfaction.Furthermore, Semira (2019) and Girum (2017) used HEdPERF to investigate the effect of service quality on students' satisfaction.Since higher education service quality is composed of multidimensional constructs, using a single model cannot fully measure the complex nature of Ethiopian higher education service quality.Consequently, in the present study, a comprehensive Higher Education Service Quality Model [HESQM] was developed to measure Ethiopian higher education service quality.Through an extensive literature review, seven main and nine sub-dimensions of higher education service quality and included a total of 136 items were identified.
The newly developed HESQM was empirically tested in Ethiopian higher education through the purposefully selected experts from the fields of educational planning and management, curriculum and instruction, and psychology, have evaluated the content validity of the instrument.Furthermore, the reliability of the instrument was also analyzed to further check the internal consistency of the instrument.In any research, ensuring the validity and reliability of the instrument is an important procedure as a valid instrument measures what the questionnaire is supposed to measure and reliable instruments are internally consistent (Ghazali, 2016).Based on the test results of content validity and reliability, constructs that do not meet the intended objectives of the study were removed from the questionnaire.
After making all necessary corrections on the questionnaire, a comprehensive HESQM consisting a total of 116 items was retained.Since using a single model is unable to fully explain the Ethiopian higher education service quality, the newly developed comprehensive HESQM model is better at explaining educational service quality from the Ethiopian higher education context.

Conclusion and Recommendations
The results of the study confirmed that the newly developed comprehensive HESQM instrument was valid and internally consistent.The findings of the present study further assured that HESQM is a comprehensive model to measure Ethiopian higher education service quality.Because the present model is comprehensive and empirically tested in the Ethiopian higher education context, the model better measures the Ethiopian higher education service quality.Thus, future researchers may use the newly developed compressive HESQM to conduct research that is associated with Ethiopian higher education service quality and quality assurance and service quality.The Ethiopian higher education quality assurance agencies administrators, educational policy and decision makers and experts can incorporate additional service quality dimensions that have not been touched in their higher education quality assurance yardsticks.Furthermore, each Ethiopian HIE has 'Institutional Quality Enhancement Directorate Director'.This office follows up on the educational quality of their respective institutions.Hence, the present study provides significant information to institutional quality assurance experts who evaluate the institutional quality of education in providing a comprehensive model to evaluate the educational quality of the institution.

Practical implications
The present HESQM is a comprehensive model to measure Ethiopian higher education service quality.Since using a single model is unable to fully explain the Ethiopian higher education service quality, the new comprehensive model is better at explaining Ethiopian higher education service quality.

Social implications
The existing literature lacks a comprehensive service quality model to measure higher education service quality, particularly from the Ethiopian higher education context, the place of current research.Measuring higher educational service quality is one of the most important elements to ensure the quality of higher education.The findings of the present research provide valuable insights to HEI management bodies, higher quality assurance agencies and Ministry of Education to learn a comprehensive model that better measures higher education service quality.Thus, the present may help the practitioners who measure higher education service quality by providing a comprehensive model.

Implications for Future Research
In this study, quantitative research approach was employed.Further comparable studies should conduct a mixed research approach to better determine the present findings.
In addition, future studies should be conducted using Exploratory Factor Analysis [EFA] to strengthen the present findings though the new HESQM model's content validity and reliability have been empirically tested in Ethiopian higher education.In general, the higher education service quality is a combination of various factors or variables.Therefore, this study also encourages future researchers to still re-look at other known models such as Total Quality Management [TQM], Service Driven Market Orientation [SERVMO], and

Figure 1
Figure 1 Quality Management of Excellence [HETQM], to further incorporate missing constructs in the present Higher Education Service Quality Model [HESQM].

Table 1
Strengths and weaknesses of existing generic service quality models Source: Authors own work; where; SERVQUAL=Service Quality; SERPERF=Service Performance

Table 3
Content validity evaluators' profile

Table 4
Summary of Rule of Thumb of Content Validity Index[CVI]

Table 5
Content validity index of education service quality questionnaire

Table 6 Summary of items revision Items I-CVI Result Potential Errors Identified Based on I-CVI Result Actions Taken
Source: Authors own work; where; Q=Question; I-CVI=Item Level Content Validity Index ISSN 2719-0633 (Print) 2719-0641 (Online) | 93

Table 7
Reliability results of education service quality questionnaire (N=40)

Table A2
Summary of modified higher education performance modelAuthors own work; Note: Total Dimensions of HEdPERF: 5; No. of constructs: 41