A Rasch Model Distractor Analysis on Ordered Multiple-Choice Items of a Fractional Test
Assessments are not used for evaluating students’ performance only; rather, they should provide information deemed interpretative, diagnostic, highly informative, and potentially prescriptive. To achieve this goal, assessments must link cognition, observation and interpretation. Hence, the purposes of this study were: 1) to explore how to conduct a distractor analysis on ordered multiple-choice items (OMC) developed in accordance with cognitive developmental theory of fractions, under Andrich’s “ordered thresholds” perspective within the framework of Rasch modeling; 2) to compare the psychometric properties of OMC that were given partial credits and those that were scored dichotomously. Although results in general support the proposed developmental levels of distractors, only a few which conformed the principles of ordered thresholds, of appropriate distance between thresholds, and of distractor response curve showing peak along the latent continuum, can reliably distinguish the developmental levels of fractions and hence are suitable for awarding partial credits. However, when OMC were scored (or rescored) properly as polytomous items, not only did test precision increase, the chances for identifying students who are progressing between developmental levels were also increased. Discussions and suggestions for developing OMC and for conducting further research are also provided.
|關鍵詞||分數認知發展、次序選擇題、誘項分析、Rasch模式、cognitive development of fractions、ordered-multiple choice item、distractor analysis、Rasch model|