Does Repetition of the Same Test Questions in Consecutive Years Affect their Psychometric Indicators? – Five-year Analysis of In-house Exams at Medical University of Warsaw

Abstract

Aim of study:Evaluation of the re-used test questions on the impact of psychometric indicators of test items in examinations in cardiology in Physiotherapy at the Medical University of Warsaw (MUW).Materials and Methods:A case study based on analysis of 132 five-option (528 distractors) multiple-choice questions (MCQs) developed at MUW included in five in-house exams. Questions repeated at least twice during the period considered and constituted 42.4% of all MCQs. Each MCQ was assessed on the basis of the following three indicators: difficulty index (DI), discrimination power (DP), and the number of non-functioning distractors (N-FD). The change in psychometric indicators of test items was assessed using Krippendorff alpha coefficient (αk).Results:Together with each MCQs repetition, a decrease in the number of questions that would maintaining the analogical DI value towards the initial level of easiness was observed. However, the level of DI compliance was significantly higher, even when there were five consecutive repetitions (coefficient αk for the consecutive repetitions was 0.90, 0.85, 0.78 and 0.75). N-FD number in consecutive repetitions remained on a satisfactory level (good and very good compliance), although there was a significant decrease in this range when there were three or more repetitions (coefficient αk was 0.80, 0.69, 0.66 and 0.65, respectively). Whereas the level of similarity as for DP for consecutive repetitions was significantly lower in comparison with those noted for DI and DE (DP coefficient αk was 0.28, 0.23, 0.25 and 0.10, respectively).Conclusions:The observed change in the initial values of psychometric indicators together with consecutive use of the same MCQs confirms the examiners’ concerns as for the progressive wear of the bank of test questions. However, the level of psychometric MCQs values loss, especially in the area of the easiness and the number of non-functioning distractors was not drastic. It appears that the level of MCQs spread among students of consecutive years is not too high, at least within two consecutive years.