Analyzing Factors of NUST MS Admission Policy/Process: A Step towards Uniform National Admission Policy for the Universities of Pakistan.

Abstract

Universities introduce the process of selection or screening for the applicants with the objective to select the “BEST” among available. Therefore, this process must be transparent, proficient, balanced, and complete. Studies highlighted a heterogeneous set of variables that individually or collectively can be observed as a process; though, it is almost impossible to define a universal criterion. This study investigated the effectiveness, balance, and completeness of different variables as an admission process, followed by a leading national university (National University of Sciences and Technology (NUST)) at the postgraduate level. The process consists of three variables: i) previous academic record of an applicant (ACAD) ii) marks obtained in graduate record examinations (general) or graduate assessment test (general) (GAT) (a test conducted by the national testing service of Pakistan for Higher Education Commission of Pakistan) and iii) interview (INT) conducted by the concerned school/institution/center of NUST. Moreover, the current weightages of ACAD, GAT, and INT in the merit calculation are 25%, 50%, and 25%, respectively. Since this is an empirical analysis, therefore, an archival student’s admission data, spanning over seven years has been used in this study. The information concerning these mentioned variables of 13094 applicants has been provided by the ICT directorate of NUST. Based on the literature review, the span and size of the sample used for analysis are sufficiently large to derive significant conclusions regarding the process. Descriptive and inferential analysis has been used to observe general trends of variables and for comparison of the performance of admitted against not admitted students. Moreover, multiple linear regression (MLR) & binary logistic regression (BLR) models have been used to develop predictive models for merit (being continuous variable) and status, i.e. admitted vs not admitted (being categorical variable). The results showed that admitted students significantly differ in performance relative to not admitted students, primarily influenced by the INT scores with a marginal difference between GAT and ACAD scores. The results of the process of development of predictive models showed that the linear method is not suitable for this purpose due to the lack of a linear relationship between dependent and independent variables. Therefore, binary logistic regression considering the status of an applicant is a suitable alternative. Results showed that the three variables (ACAD, GAT, and INT) are not balanced (as compared to subjective weights assigned to them) and complete (lack in predictive ability). Therefore, there is a need for revision of weightages and inclusion of other relevant factors like popularity of a program, financial status of the applicant, place of residence, hostel facility, etc., These results provide useful insight for the choice of variables to be observed as a process of admission for postgraduate students, not only for NUST but for other national and international universities. Further research considering other factors and case studies for different universities can pave the way towards the uniform national admission process especially at the postgraduate level.