문항반응이론 기반 컴퓨팅 사고력 평가의 타당성 및 동등화 가능성 검토

초록

Despite the widespread use of algorithmic problem-solving tests based on text-based programming languages (TPL) for assessing computational thinking (CT) in both educational and industrial sectors, empirical validation of such an instrument remains limited. To address this gap, this study analyzes large-scale algorithmic problem-solving assessment data using item response theory to estimate item parameters and examinee abilities, while also establishing a framework for score comparability through equating procedures. The findings confirm that the Rasch model provided reliable estimates of item difficulty and latent abilities, and a grading system developed from these estimates demonstrated significant potential for score comparability. Overall, this study advances evidence-based assessment practices for TPL-based algorithmic problem-solving tests and provides a methodological foundation for enhancing CT-oriented SW education assessments. Furthermore, the present study offers practical implications for developing fair industrial hiring tools and national-level assessment systems.

키워드