Fig. 1From: Peer assessment of professionalism in undergraduate medical educationInter-class correlation reliability analyses by group size. Data was binned by the number of assessors and 2-way random inter-class correlations (ICC) calculated per group size for each of the professionalism scales. The number of assessment groups per size category varied and arose from constraints during implementation, including variation in initial working group size and missed assessment submissions. The 95% confidence intervals of each ICC measure were plotted in gray and a reliability co-efficient of 0.7 was considered the threshold for consistent assessmentsBack to article page