Evaluation of Multiple-Choice Questions in the Theoretical Biostatistics Course for Pharmacy Students Using Latent Class Multinomial Logit Model in Sensitivity Analysis of Student Classification

Various studies have evaluated the validity and reliability of test questions, showing that using questions with different difficulty levels improves assessment quality. The aim of the present study was to evaluate the performance of multiple-choices test for accurate evaluation of learning and to identify the strengths and weaknesses of pharmacy students in the theoretical biostatistics course.

In this descriptive-analytical study, the test results of 101 pharmacy students who took the theoretical biostatistics course during the first and second semesters of the academic year 1402-1403 were analyze. In the final exam, 30 multiple-choice questions designed in accordance with the syllabus of the theoretical biostatistics course. Difficulty and discrimination indices calculated for each test question. The latent class multinomial logit model used for educational datamining and sensitivity analysis of student classification into high and low groups.

Students were grouped based on quartiles of their final exam scores, and then the latent class model fitted based on the test questions and students' GPAs, showing complete agreement between grouping based on quartiles and the results of the latent class model. The discrimination index of 15 out of 30 questions was average to high, but the difficulty index of 19 questions was greater than 0.8, indicating that these questions were easy for pharmacy students.

Continuous evaluation and revision of test questions can enhance assessments of student learning and consequently improve the learning and teaching strategies of the biostatistics course for pharmacy students.