Does the pass-mark created using Angoff standard setting correlate with the performance of borderline passing medical students in anatomy examinations at The University of Bristol? Abstracts presented at Joint Meeting of the Anatomical Society & the British Association of Clinical Anatomists Summer Scientific Meeting 2016, Brighton, United Kingdom

Medical students undertake anatomy examinations in year 1 and 2 of their course. The pass-marks for these examinations are generated using standard setting (Angoff technique). This technique requires assessors to consider, for each question, the percentage of borderline passing students (BPS) that will get the answer correct. A moderation is performed to remove bias of outlying assessors. The sum of moderated averaged percentages per question gives the final pass mark. This technique is prospective, so doesn’t consider cohort dependent student performance. We used questions from the anatomy section of the 2015-16 year 2 formative assessment (pass-mark=16.5/30). BPS were defined as +/-5% from the pass mark, (n=9). We looked for correlation (Spearman’s) between the generated moderated pass-mark versus actual performance of BPS. The assessor group were divided by whether they taught anatomy to these students (contact group) or not (non-contact group), and were assessed for correlation (unmoderated and moderated data). Ethical approval was obtained. Results showed moderate positive correlation for the contact group (rs=0.48, p<.007), and weak positive correlation for the non-contact group (rs=0.248, p<.186). Sub-analysis of the moderated data showed that this step improved the correlation for the non-contact group, however it remained weakly positive (rs=0.376, p<.041). (Combined moderated data, moderate positive correlation (rs=0.498, p<.005)). The results suggest Angoff standard setting generates a pass-mark which matches performance of BPS but there must be an awareness of the characteristics of the individual assessors. This study method will be used to continue to analyse the validity of this method for future assessments.