REFINEMENT OF MULTIPLE CHOICE ITEMS USING DIFFICULTY INDEX AND DISCRIMINATION POWER FOR ASSESSING MIDDLE STAGE LEARNERS' ATTAINMENT IN SCIENCE

Abstract

The present research study proposes to analyse multiple-choice items for the development of assessment tools for measuring conceptual learning in science. Initial pool of items were reviewed in a Focused Group Discussion (FGD) with a national level expert group and 44 items were eliminated out of 691 pool of items. Rest 647 items were again reviewed in a Focus Group Discussion (FGD) with state level experts for contextual and language intervention. As a result 30 items of total 647 items were modified. The finalised 647 items were subjected to field try out. The quantitative data were obtained by administration of total 647 items among 516 students from Grades 6, 7 and 8 theme-wise separately. Data obtained were subjected to individual item analysis in which both students' answers and test questions were examined in order to assess the quality and quantity of the items and the test as a whole. Test items were assessed and screened on the basis of values of difficulty indices (Di) and discrimination power (Dp). Total 203 items were finalised for tools for assessing learner's attainment in science (TALS). Test validity was assessed applying Cronbach alpha separately to all tests. Statistical analysis was done using SPSS version 24 (IBM free software trial version). The tool has been developed for the first time for assessing learners’ attainment in science for conceptual understanding among learners of tribal areas of Chhattisgarh. The tool will be utilised for baseline study for measurement of learning gaps and designing the interventions for improvement in learning and teaching of science.