Julie D. Gregorio
University Research Center
New Era University
How to Cite:
Gregorio, J. D. (2024). Service Quality in Higher Education : Graduating Students’ Ratings Across Different Colleges. NEU Knowledge Journal: A Compilation of Researches of New Era University Faculty, Staff, Students, and Administrators, 1(1), 1-18.
ABSTRACT
This study examined graduating students’ perceptions of university service quality across ten colleges at a private university. A total of 235 respondents completed a 10-item University Experience Survey (UES) using a 7-point Likert scale. Because normality was not met, the One-Sample Wilcoxon Signed Rank Test was used, which rejected the null hypothesis, Z = 13.22, p < .001, suggesting that the observed median AVG was significantly higher than the hypothesized median of 4.00. Differences across the ten colleges were tested using the Kruskal-Wallis 1-way ANOVA, which indicated significant variation in satisfaction levels, H (9) = 79.87, p < .001. Post-hoc pairwise comparisons revealed that several colleges had significantly different satisfaction scores. The College of Business Administration differed significantly from several colleges in students’ ratings of university services. Specifically, COC, COA, CEA, and CICS had significantly higher ratings than CBA, whereas CAS had significantly lower ratings than CBA. The findings indicate that service quality is experienced unevenly across academic units and emphasize the need for targeted, college-sensitive improvement strategies. More broadly, the study demonstrates the value of non-parametric diagnostic analysis in higher education service evaluation when survey data are non-normal.
Keywords: University service quality, student satisfaction, Kruskal-Wallis test, non-parametric methods, college comparison, educational research
INTRODUCTION
Service quality in higher education has increasingly become a focal point for institutions seeking to enhance their competitive advantage, with student satisfaction playing a critical role in determining the overall educational experience (Alves & Raposo, 2007; Parasuraman, Zeithaml, & Berry, 1985). Student satisfaction directly influences retention, graduation rates, and institutional reputation (Tsinidou, Gerogiannis, & Fitsilis, 2010). The diversity of student populations, especially across different academic affiliations, can lead to varying service quality expectations and satisfaction levels, requiring targeted improvements in service delivery.
The present study aims to examine the perceptions of graduating students from ten colleges within a private university regarding service quality. Specifically, it seeks to answer three research questions: (1) How satisfied are students with overall university services? (2) Do satisfaction levels vary significantly across different colleges?
Related Literature
Service quality in higher education is multi-faceted, encompassing both academic and non-academic services. Parasuraman, Zeithaml, and Berry’s (1985) SERVQUAL model has been widely utilized in assessing service quality, highlighting the dimensions of tangibility, reliability, responsiveness, assurance, and empathy. These dimensions have been adapted to the context of higher education, where students expect high standards in academic delivery, administrative support, and campus facilities (Alves & Raposo, 2007).
Previous research has demonstrated the importance of understanding variations in student satisfaction across different academic departments or colleges. For instance, Douglas, McClelland, and Davies (2008) found that students’ perceptions of quality varied significantly depending on their field of study, with business students typically reporting higher satisfaction than those in arts and humanities. Such findings suggest that institutions must tailor service improvements to the unique expectations of each academic group (Tsinidou et al., 2010).
The methodological rigor in assessing service quality perceptions is crucial, particularly in addressing challenges such as non-normal data distributions. Non-parametric methods, such as the Mann-Whitney U test and Kruskal-Wallis H test, offer robust alternatives when assumptions of normality and homogeneity of variance are violated (Field, 2018). These methods are especially useful in educational research, where student satisfaction data often exhibit skewed distributions (Bohrnstedt & Knoke, 1988).
Research Questions
1. RQ1: Are the median ratings of university experience aspects significantly different from a neutral value of 4 (Neither Poor nor Good)?
Null Hypothesis (H01): The median rating of each aspect is equal to 4.
Alternative Hypothesis (Ha1): The median rating of each aspect is not equal to 4.
2. RQ2: Are there significant differences in student satisfaction levels across the 10 colleges at NEU?
Null Hypothesis (H₀2): There are no significant differences in student satisfaction levels across the 10 colleges at NEU.
Alternative Hypothesis (Hₐ2): There are significant differences in student satisfaction levels across the 10 colleges at NEU (Ηₐ2: At least one pair of median is different).
METHODOLOGY
Research Design
This study employed a cross-sectional quantitative research design to assess graduating students’ satisfaction with university services across ten different colleges within a private university. A survey methodology was utilized, wherein a standardized questionnaire was administered to the respondents. The study’s quantitative approach enabled the systematic examination of differences in satisfaction levels based on college affiliation.
Participants
The participants in this study were 235 graduating students from ten colleges at the university. These colleges included:
1. College of Informatics and Computing Science (CICS)
2. College of Accountancy (COA)
3. College of Arts & Sciences (CAS)
4. College of Business Administration (CBA)
5. College of Communication (COC)
6. College of Criminology (CCR)
7. College of Education (CED)
8. College of Engineering and Architecture (CEA)
9. College of Nursing (CON)
10. School of Graduate Studies (SGS)
The sample size of 259 was initially computed using G*Power 3.1.9.7 with an a priori F test ANOVA: Fixed effects, omnibus, one-way. The parameters for the power analysis were set with an effect size of 0.25, alpha of 0.05, and power of 0.80 for 10 groups (Cohen, 1988). Although the target was 259 respondents, 235 valid responses were collected, yielding a return rate of 90.73%.
Sampling Technique
The study used proportionate stratified random sampling techniques to ensure the sample was representative of the population of graduating students. The sample was stratified based on college affiliation, ensuring that each college was represented in the sample. Within each stratum, systematic random sampling was employed, with research assistants instructed to select respondents at an interval of four seats among the graduates.
To ensure the accurate and unbiased collection of data, 18 student-volunteers were trained as research assistants. These student-volunteers were provided with detailed instructions on how to administer the survey. Specifically, they were instructed to distribute the questionnaires to graduating students only after the students had returned to their assigned seats following their appearance on stage. This timing was chosen to minimize distractions and ensure that the respondents could focus fully on the questionnaire.
The surveyors were explicitly instructed to remind respondents not to confer or communicate with others while completing the questionnaire. This was done to maintain the integrity of the responses and ensure that each participant’s feedback was independent and unbiased.
Instrument
The primary data collection tool was the University Experience Survey (UES), a 10-item questionnaire designed to measure student satisfaction across various university services. The items in the questionnaire were:
1. Quality of Education (EDUCATION)
2. Policies and Rules on Discipline (DISCIPLINE)
3. Enrollment Procedure (PROCEDURE)
4. Canteen Services (CANTEEN)
5. Cashier Services (CASHIER)
6. Clinic Services (CLINIC)
7. Guidance and Counseling Services (GUIDANCE)
8. Library Services (LIBRARY)
9. Registrar Services (REGISTRAR)
10. Sports and Recreation Services (SPORTS)
Respondents rated these items using a 7-point Likert scale, where 7 = Excellent, 6 = Very Good, 5 = Good, 4 = Neither Poor nor Good, 3 = Poor, 2 = Very Poor, and 1 = Extremely Poor. This scale was chosen to provide a nuanced understanding of students’ perceptions across a range of services and experiences within the university
Validity and Reliability
Before the administration of the UES, the content validity of the questionnaire was evaluated by three experts in test construction, who reviewed the items and provided feedback. Their comments and suggestions were incorporated into the final version of the questionnaire to ensure that it effectively captured the constructs of interest.
For reliability, internal consistency was estimated using Cronbach’s alpha. The questionnaire was tested with 50 respondents, resulting in a Cronbach’s alpha of 0.90, which exceeds the minimum acceptable value of 0.70 (Bohrnstedt & Knoke, 1988). This high level of reliability indicates that the instrument consistently measured student satisfaction across different university services.
Data Analysis
The collected data were analyzed using non-parametric statistical techniques in IBM SPSS such as the One-Sample Wilcoxon Signed Rank Test and Kruskal-Wallis H test. To identify specific differences between college affiliations, pairwise comparisons were conducted using the Mann-Whitney U test with Bonferroni corrections to account for multiple comparisons.
Ethical Considerations
In conducting this study, ethical standards were rigorously upheld. Participation was voluntary, and respondents were assured of the confidentiality and anonymity of their responses. The research also adhered to the principles of academic integrity and transparency throughout the data collection and analysis process.
RESULTS AND DISCUSSION
Descriptives
The graduating student respondents reported generally high satisfaction with university services. The average overall rating was M = 5.93, SD = 0.82, with a median of Mdn = 6.00, IQR= 1.30, and a 95% confidence interval of 5.82 to 6.03. The distribution was slightly negatively skewed, indicating that responses were concentrated toward the higher end of the scale.
Tests of normality below (Table 2) showed that the overall ratings were not normally distributed. The Kolmogorov–Smirnov test was significant, D (235) = 0.127, p < .001, and the Shapiro–Wilk test was also significant, W (235) = 0.943, p < .001. Because the assumption of normality was violated, a one-sample Wilcoxon signed-rank test was used instead of a one sample t test.
One-Sample Wilcoxon Signed Rank Test
To determine whether the median of the average overall rating (AVG) differed significantly from the neutral midpoint value of 4; a One-Sample Wilcoxon Signed Rank Test was conducted. The test was chosen due to the non-normal distribution of the AVG scores, as indicated by the results of the Kolmogorov-Smirnov and Shapiro-Wilk tests, based on an asymptotic significance (2-sided test). Given that the p-value is less than the alpha level of .05, the null hypothesis was rejected.
The test statistic of 27,241.50 and a standard error of 1,029.49 further support this finding. The positive Z-value indicates that the observed median AVG was significantly higher than the hypothesized median of 4.00. This result suggests that the median overall satisfaction rating from the respondents is significantly above neutral, implying a generally favorable evaluation of their university experience.
Research Question 2
Assumptions Check
The assumption of normality was tested using the Kolmogorov-Smirnov and Shapiro-Wilk tests for each college affiliation group. The results indicated that the AVG ratings were not normally distributed in several groups, including CBA,CAS, and CCR (p < .05). Because the distribution of the AVG ratings did not meet the assumptions for parametric analysis, differences in ratings across college affiliations were examined using the Kruskal–Wallis test.
Due to the violations of normality and homogeneity of variance, the nonparametric equivalent of a one-way ANOVA, the Kruskal-Wallis 1-Way ANOVA, was conducted.
Kruskal-Wallis 1-Way ANOVA
Kruskal-Wallis 1-Way ANOVA test was employed to evaluate differences in average ratings among the various college affiliations. The Kruskal-Wallis test yielded a significant result, H(9) = 79.872, p <.001 (Table 4). This result indicates that there are significant differences in the ratings of the university services among the colleges. Given that the p-value is less than the alpha level of .05, the null hypothesis was rejected. The pairwise comparisons provide further insights into which specific colleges show significant differences.
Post-Hoc Pairwise Comparisons
To determine which specific colleges differed in their average ratings, post-hoc pairwise comparisons were conducted using the Mann-Whitney U test with a Bonferroni correction for multiple comparisons. The adjusted significance level for these comparisons was set at α = 0.05/45 = 0.0011 (Table 5). Several significant differences emerged:
- CAS vs. CBA: A significant difference was observed, Z = -5.730, p < 0.001, indicating that CAS students rated the services lower than CBA students.
- CAS vs. SGS: Another significant difference was found, Z = -4.526, p < 0.001, suggesting that CAS students had lower ratings compared to SGS students.
- COC vs. CBA: There were significant differences, Z = 5.549, p < 0.001, with COC students rating the services higher than CBA students.
- COC vs. SGS: There were significant differences, Z = -4.294, p < 0.001, with COC students rating the services lower than SGS students.
- COA vs. CBA: Significant differences were observed, Z = 4.73, p < 0.001, with COA students rating the services higher than CBA students.
- CEA vs. CBA: There were significant differences, Z = 5.842, p < 0.001, with CEA students rating the services higher than CBA students.
- CICS vs. CBA: Significant differences were observed, Z = 4.484, p < 0.001, suggesting that CICS students rated the services higher than CBA students.
Other pairwise comparisons did not yield significant results after Bonferroni correction.
Overall, the findings suggest that graduating students were satisfied with university services, but their evaluations differed across college affiliation. The strong overall rating is consistent with the idea that students value both academic and non-academic services when evaluating their university experience. In higher education, satisfaction is shaped not only by instruction but also by administrative efficiency, accessibility of support services, and the physical learning environment. The relatively high median score in this study suggests that most students felt positively about these areas, although the spread across colleges indicates that some groups experienced the services more favorably than others.
The significant Kruskal–Wallis result confirms that college affiliation is associated with different satisfaction levels. This may reflect differences in how services are delivered, managed, or accessed across colleges. For example, some colleges may have more responsive administrative systems, stronger student support structures, or better alignment between student needs and available services. These differences matter because they can influence not only satisfaction but also retention, engagement, and institutional trust.
The post hoc comparisons suggest that the College of Business Administration stood out in several pairwise comparisons, often giving lower ratings than other colleges. This may point to areas where service delivery should be reviewed more carefully. By contrast, higher ratings in some other colleges, such as the College of Engineering and Architecture, may indicate better service coordination or a more favorable student experience in that college. These findings do not necessarily imply poor performance overall, but they do suggest uneven student experiences that merit institutional attention.
Limitations and Future Research
The non-normal distribution of the data and the violation of homogeneity of variance justified the use of non-parametric analyses. Although these methods are appropriate and robust for the present data, they provide less information about effect structure than parametric alternatives. Future studies should consider larger samples and, where assumptions are met or appropriately modeled, analytical approaches that permit more detailed examination of group differences. In addition, qualitative inquiry may help explain the contextual and institutional factors underlying variation in ratings across colleges.
Future research using the University Experience Survey (UES) may also benefit from longitudinal or repeated-measures designs to examine students’ ratings of multiple service domains across time or across related service areas. Such approaches would allow a more nuanced assessment of which university services are perceived more favorably or less favorably, and would provide stronger evidence for targeted service improvement.
REFERENCES
Alves, H., & Raposo, M. (2007). Conceptual model of student satisfaction in higher education. Total Quality Management & Business Excellence, 18(5), 571-588.
Bohrnstedt, G. W., & Knoke, D. (1988). Statistics for social data analysis (2nd ed.). Peacock Publishers.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nded.). Lawrence Erlbaum Associates.
Douglas, J., McClelland, R., & Davies, J. (2008). The development of a conceptual model of student satisfaction with their experience in higher education. Quality Assurance in Education, 16(1), 19-35.
Field, A. (2018). Discovering statistics using IBM SPSS statistics (5th ed.). SAGE Publications.
Parasuraman, A., Zeithaml, V. A., & Berry, L. L. (1985). A conceptual model of service quality and its implications for future research. Journal of Marketing, 49(4), 41-50.
Tsinidou, M., Gerogiannis, V., & Fitsilis, P. (2010). Evaluation of the factors that determine quality in higher education: An empirical study. Quality Assurance in Education, 18(3), 227-244.
OpenAI. (2025). ChatGPT (GPT-4o version) [Large language model]. https://chatgpt.comGrammarly. (2025).
Grammarly [Writing assistant software]. grammarly.com
