Application of Conditional Means for Diagnostic Scoring
Hollis Lai, Mark J. Gierl, Оксана Бабенко · 2015
Abstract. In educational assessment, demand for diagnostic information from test results has prompted the development of model-based diagnostic assessments. To determine student mastery of specific skills, a number of scoring approaches, including subscore reporting and probabilistic scoring solutions, have been developed to score diagnostic assessments. Although each approach has a unique set of limitations, these approaches are, nevertheless, often used in diagnostic scoring, whereas an alternative approach, Complex Sum Scores (CSS), has not received much attention yet. With the process of developing model-based diagnostic assessments becoming increasingly complex, we revisit the CSS and demonstrate two applications of the CSS in the development of diagnostic assessments. Two applications include: (a) illustrating and validating skills within the model, and (b) partial mastery scoring using model-based distractors. By demonstrating the two applications, we aim to show how model-based diagnostic assessments can be developed and scored using the CSS scoring approach, the results of which can be used by teachers to inform teaching and learning.