The Present Study Investigates the Effects of Seating Arrangement on Academic Outcomes and Prejudice

Traditional desegregation policies were successful in creating mixed ethnic schools, but social networks within the schools are still largely segregated. Possibilities to intervene in students’ relationships are limited, but teachers can determine the seating arrangement. Previous research shows that desk-mates are able to influence each other’s academic achievement. Furthermore, contact with an ethnically different desk-mate has proved to decrease prejudices against the out-group. Thus desk-mate assignment can be a useful tool to achieve different policy goals. We aim to reveal the consequences of different seating arrangement scenarios on prejudice and academic outcomes.

This paper applies agent based modelling. The empirical foundation of the model is the “Wired into Each Other” project, a longitudinal piece of research carried out by HAS “Lendület” RECENS. Our model represents a classroom with one teacher and students. The teacher intervenes in the students’ sitting arrangement. The students are agents who receive a new desk-mate based on the rules defined by the teacher. Further, they update their GPA and level of prejudice based on their desk-mate’s characteristics and the characteristics of the classroom. We run the model for multiple classes with different ethnic compositions and observe how the GPA distribution and level of prejudice changes over time.

The results draw attention to the trade-off between policy goals. For example, seating the students with the highest GPA together with the students with the worst GPA narrows GPA inequalities and reduces prejudice, but does not allow the best students to thrive. In contrast, letting students decide with whom they want to sit allows the best students to perform better, but has little effect on decreasing inequalities and increases prejudice. Further, we show how the GPA and the level of prejudice can further be influenced by different classroom compositions.