Refining assessment: Rasch analysis in health professional education and research

Abstract

Educators want to assess learners using assessment processes that provide valid measures of learner ability. An ideal assessment tool would include items that are appropriate for assessing the target attributes. Ideal assessment results would accurately differentiate learners across the spectrum of ability, determine which learners satisfied the required standard and enable comparison between learner cohorts (e.g., across different years). Similar considerations are relevant to researchers who are designing or revising methods used to gather other kinds of assessment data, such as participant responses to surveys or clinical measurements of performance. Analysing assessment scores using Rasch analysis provides information about scores and the nature of each assessment item, and analysis output guides refinement of assessment. However, few health professional educators have published research that includes Rasch modelling methods. It may be that health professional educators find the language used to describe Rasch analysis to be somewhat impenetrable and that this has, to date, limited engagement in exploring applications for Rasch. In this paper, we lay out an overview of the potential benefits of Rasch analysis in health professional education and research.