Item analysis with precise information: An item information function approach in IRT.

Abstract

Objective: To identify valid and reliable items providing precise information regarding learner ability to be included in the test pool. Study Design: Descriptive Cross-sectional study. Setting: SMBB Medical College Lyari Karachi. Period: December 2024. Methods: Purposive sampling techniques are employed to collect data from 100 MBBS students using physiology SBAs at C1, C2 and C3 level from formative assessment of 2^nd Professional MBBS Gastrointestinal tract module held in a public sector medical institute SMBB Medical College, Karachi. The study employed 2 PL dichotomous IRT model for item analysis using discrimination and difficulty index with subsequent application of item information function (IIF) technique to gauge items that rendered precise information employing statistical software STATA 17. Results: The item analysis using discrimination and difficulty index found only one acceptable and two desirable items at C1 level while rest are redundant.  There is only one desirable item at C2 level while rest are discarded, at C3 level two items are found acceptable, one desirable and rest are discarded. However, subsequent analysis with IIF shows that only one desirable item from C1, C2 and C3 level is appropriate for inclusion in the test pool whilst others need to be revised. Conclusion: The study considers item information function (IIF) as a supporting approach to a traditional item analysis for building a quality test pool.  It is also established that even acceptable items that are possible candidates for inclusion in test pool by item analysis subsequently disqualified by IIF as are failed to submit precise information regarding learner’s ability where only one desirable item from C1, C2 and C3 level appeared to provide precise information hence could be included in the test pool.