I don’t believe it: Correcting students’ misperceptions about peers’ time spent on homework

Abstract

In this master thesis we investigate the effect of providing students with accurate information about their peers’ time spent on homework. We use experimental data collected from two surveys carried out on 10th grade students in Norway. The main survey contained an intervention targeting students who spent below the median time of their class on homework. These students were provided with information about the actual median time spent on homework in their class. The follow-up survey consisted of questions regarding the students’ beliefs. A partial population design was utilized in order to capture any spillover effects, in addition to direct treatment effects. Our main results suggests that the intervention was successful in correcting students’ beliefs. Both the reduced form estimation and the instrumental variable estimation suggested a positive treatment effect across our six outcome variables. We used three different specifications, and while we see some differences between them, the main take-away suggests a positive treatment effect. Our analysis suggests some heterogeneity across students’ attitudes, but the evidence is weak. We also check for heterogeneous effects of treatment and spillover conditional on the students’ centrality in the peer group. We find some initial differences across these subgroups, but the evidence is ambiguous and does not provide any clear insight into this question. We recommend further investigation of the direct behavioral changes of such an intervention, as well as more in-depth investigation of the peer effects.