Prediction of academic success in manufacturing engineering technology at Texas A & M University

Abstract

The manufacturing program of studies within the Engineering Technology Department at Texas A&M University has regularly attracted numerous students. Administrators are required to be selective in the admission of students entering the manufacturing program. The objectives of this study were to develop predictor formulas needed to aid the screening of potential students entering a program of manufacturing engineering technology. Transcripts were the source of the data. All 89 manufacturing engineering technology graduates during a five year period were the population. Dependent variables were cumulative grade point average (CUMGPA), engineering technology grade point average (ENTCGPA), and manufacturing engineering technology grade point average (MFGGPA). There were 24 independent variables studied. Sixteen independent variables were grade point averages from all courses listed in the first four semesters of the curriculum. The remaining eight independent variables were graduation age, cumulative GPA after completing 30 semester credit hours, and scores on the Scholastic Aptitude Test (SAT). Both t-tests and F-tests were used to determine if data from graduates with complete university records were statistically different than data from graduates with incomplete university records. Pearson product-moment correlations were used to determine relationships between prediction variables and response variables. The stepwise procedure in multiple regression analysis was used to develop the prediction models and corresponding parameter estimates. Cross-validation of the developed prediction equations were conducted by the split-halves technique. The analyses provided multivariate quantitative models to predict CUMGPA, ENTCGPA, and MFGGPA which accounted for 73, 73, and 81 percent of the variance associated with the respective measures of academic achievement. Results indicated that performance in selected preliminary courses increased the explained variance of CUMGPA, ENTCGPA, and MFGGPA while SAT scores had no predictive value. Based on the results of this study, it is recommended that the predictor models be used as a tool in selecting students for the manufacturing engineering technology program at Texas A&M University. Administrators desiring to selectively choose applicants for an educational program should consider using the methods of this study to develop specific equations for their programs. Application data should be made available for analysis in future studies.