A longitudinal study of how social networks influence similiarity between examinations answers ‐ The hidden medical school

Abstract

Background: Previous studies has showed that prevalence of similarity increase with course years and possible explanation is the increase of social networks of study. Aims: The aim of study is to assess the effect of social networks of study in the course mean and in the similarity of answers. Methods: Using a questionnaire to assess students' working relationships and using a secondary database with all exams answers, we estimate several social network analyses indicators and compare the course mean and exams answers similarity with different indicators. Results: There were detected 135 students' pairs of exams answers similarity. There were positive association between social network indicators in-degree (β=0.067, 95%CI:0.036-0.098) with the course mean. Belonging to the same cluster had higher probability of similarity (OR=2.09, 95%CI:1.08-3.97). When the pair has higher difference in-degree, the probability of similarity is lower comparing with a pair with small difference (OR=0.871, 95%CI:0.794-0.946). When one student is more central and the other student of the pair is less central the more is the probability of similarity (OR=1.0006, 95%CI:1.0000-1.0011). Conclusions: Social networks in our cohort influence both the course mean and similarity of exams answers. To optimize the grades, we should focus on team-based learning and to improve the academic integrity of students we should diminished the disparities within social network of study of students.