Utility of factor analysis in optimization of resident assessment and faculty evaluation

Background Increasing focus on more granular assessment in medical education has led to more lengthy instruments, with concern that the increased complexity undermines the utility of these tools. This study evaluated the relative contribution of individual questions in an assessment of resident performance and a faculty performance evaluation by residents. Methods The authors performed factor analysis on the individual items in the resident assessment instrument (3,009 assessments of 71 residents) and faculty evaluations (7,328 evaluations of 61 faculty) collected from 2006 to 2012. Results Factor analysis of the resident assessment tool revealed that 1 component was responsible for 96.6% of the variance. This component encompassed each question from the assessment form, and could also be termed "overall resident competency." Factor analysis of the attending evaluation form revealed 2 unique components, representing "clinical care" and "interpersonal skills," which accounted for 89.9% of variance. Conclusions Three components accounted for 90% to 97% of the observed variance in our analysis. Factor analysis represents a useful strategy for analyzing the utility of data obtained from individual items in the assessment and evaluation instruments.