A statistical analysis of factors affecting subject performance at the University of Dar es Salaam, Tanzania: a case of Undergraduate Subjects

Research on subject performance was done to establish the relationship between the selected subject performance variables and performance in undergraduate examinations done at the University of Dar es Salaam. Some of the data were collected from the University’s Examinations office, and some were gathered by using questionnaires distributed to instructors who taught the courses which constituted the sample. The response rate was78% which was equivalent to 134 courses. In the analysis, the study has two sections. One of the sections has been devoted to data descriptionand the other for equation modelling/ specification as well as hypotheses testing. In the first section, various statistical techniques were used, such as Box plots, lows smoother (trend) curves, cross tabulations, correlations, and half table matrices. In the second one, the study has examined four phenomena comprising elements that constitute good performance. These are a subject in which at least 10% of students scored A, a subject in which at least 40% of students scored above B+, a subject in which at least 75% of students scored above B, and a subject in which at least 95% of students scored above C. The researcher chose these phenomena top test the consistency of the findings and to weigh the significance of the given variable in various phenomena. Both sections have been built around the study hypotheses. Hypotheses have been tested by first cross tabulating then performing a chi-square test to determine the significance and association of each subject performance variable with various independent variables. The study employed logit regression techniques in analyzing the variables that were significantly associated with good subject performance in order to find out the state of dependency of these variables. Consequently, all variables showed clear and interesting patterns with subject performance in the first section of analysis. In the second one, particularly in testing the significance of the variables, class size, teaching weight, number of tests, number of assignments given, per instructor number of courses, presence of seminars, showed higher consistency, as were significant in at least two scenarios.