Is a picture worth a thousand words?

Background: It has previously been suggested that the use of illustrations in MCQs may have variable effects on individual items. This study examines the effect of illustrated questions, as opposed to pure text, to discern if any overall bias between the two formats is detectable.

Summary of work: We reviewed 6 Histology MCQ papers from our Medical Junior Cycle. Classical test theory analysis was performed on all MCQs, which were then divided into two groups, those with associated images and those without.

Summary of results: We analysed 195 single best answer MCQs; 100 with associated illustrations, 95 without. The number of students per examination ranged from 277 to 347, with a total of 60,850 student-question interactions. There was no difference in question difficulty between the two groups (0.800 vs. 0.770; p = 0.862, Mann-Whitney-U). The discriminating power of the questions, as measured by point biserial correlation, was also identical (0.315 vs. 0.300; p = 0.939; Independent t-test).

Conclusions: We found no overall bias or effect on either item difficulty or discrimination resulting from the addition of illustrations. We suggest that illustrated questions, as with textual vignettes, may test a range of cognitive levels depending on how they are employed.