PSYCHOMETRIC ANALYSIS OF TEST ITEMS FOR ASSESSING TEACHER KNOWLEDGE IN THE REPUBLIC OF KAZAKHSTAN (BASED ON APPROBATION RESULTS)

Abstract

This article provides a detailed psychometric analysis of mathematics test items used during the approbation of the teacher knowledge assessment procedure (OZP) in Kazakhstan. The research is based on a comparative analysis of three different test versions administered during this piloting phase. To evaluate the quality of the items and identify their strengths and weaknesses, the study employed methods from both classical test theory (CTT) and modern item response theory (IRT), specifically the Rasch model. The purpose of the work was not only to assess the existing evaluation instruments but also to develop practical recommendations for their improvement. The approbation analysis showed that, despite the overall high reliability of the tests (Cronbach's Alpha coefficient exceeding 0,84) and their good ability to differentiate test-takers by proficiency level, a number of problematic items were found in each version. Specifically, items with low or even negative discrimination and ineffective distractors were identified. All issues discovered during the approbation have since been taken into account and rectified in the subsequent work on refining the test item bank. This has contributed to improving the overall accuracy and validity of the assessment instruments