The foundation of a good psychological measurement, such as ability estimation, is to develop an optimal item bank for educational assessments. This research aimed to develop an optimal item bank for the further teacher training programme in Mathematics while detailing all the procedures, including planning, content analysis, and test blueprint going by Bloom's taxonomy of educational objectives. The pre-experimental research design using the one-shot case study type was adopted for the study and complemented by the feature data generation FDG technique. The technique was deployed through a post-hoc simulation using item parameter estimates with real data collected from 35 participants with 68 multiple-choice items, from which 200 items with 535 participants were generated. The developed item bank was validated for ability estimation using the item response theory model with three logistical parameters: item difficulty level, item discriminating power, and pseudo-guessing. The responses to the test resulted in the collection and analysis of dichotomous scored tests for answering the generated research questions. The item parameters and ability estimation were analyzed using the Xcalibre 4.2 programme. Ninety per cent of the items met the criteria of a good item, including having a good difficulty level and functioning properly between the high and low ability participants with the discrimination and guessing parameters. A relatively low Conditional Standard Error of Measurement was recorded as the indices for ascertaining the precision of ability estimation using the developed test based on which recommendations were made. The carefully outlined procedures also ensure replicability within other programmes.

Username
Password