Evidence of gender bias in True-False-Abstain medical examinations

Background: There is evidence that males and females differ in their attainment on a variety of assessments in general and in medical education. It has been suggested that the True-False-Abstain (TFA) format with negative marking is biased against females. Methods: Eight years worth of examination data from the first two years of an undergraduate medical curriculum was analysed. 359 courses were evaluated for statistically significant differences between the genders using ANOVA. Logistic regression was used to test if subject area, calendar year or exam format predicted that males or females do better (termed male advantage or female advantage). Results: Statistically significant differences between the genders were found in 111 (31%) of assessments with females doing better than males in 85 and males better in 26. Female advantage was associated with a particular year (2001), the Personal and Professional Development strand of the curriculum, in course assessment and short answer questions. Male advantage was associated with the anatomy and physiology strand of the curriculum and examinations containing TFA formats, where the largest gender difference was noted. Males were 16.7 times more likely than females to do better on an assessment if it contained any questions using the TFA format. Conclusions: Although a range of statistically significant gender differences was found, they were concentrated in TFA and short answer formats. The largest effect was for TFA formats where males were much more likely to do better than females. The gender bias of TFA assessments in medical education is yet another reason why caution should be exercised in their use.