LAUSR

This issue of Perspectives on Medical Education contains a thought-provoking article by Young, Cummings & StOnge, looking specifically at the stability of difficulty and discrimination indices when calculated for small samples of students. They concluded that difficulty and discrimination scores calculated based upon small samples are not stable across cohorts and therefore should be interpreted with caution [1]. Given our mutual experiences in roles that seek to implement theoretically-based and researchinformed assessment practices, this article prompted us to think about assessment processes as a whole. We believe holistic review of assessment processes and practices is foundational to generating meaningful interpretations of individual metrics, such as difficulty and discrimination indices. In this commentary, we argue that assessment research would benefit from improved integration of theory and practice. The intersection of theory and practice can be tricky ground to navigate. Theoretically, an assessment strategy might be ideal, but impossible or impractical to implement. For example, despite their imperfections, there is theoretical support for the use of (modified) Angoff [2] and Ebel [3] standard setting methods. However, the implementation of these test-based standard setting methods requires someone to advocate for, hold and lead meetings, collect the results of the meeting and have the expertise to use the results to set a cut score. In addition, a team of clinical teachers must attend the meeting, be familiar with the assessment