Mental Models of High School Success -- Supplementary Evidence on Mechanisms

Following two earlier data collections with Danish adolescents at the transition between elementary and secondary school, we conduct a supplementary experiment: Using a within-subject design, we analyze how adolescents infer characteristics of the production function of educational performance from observed gender differences in educational performance. We invite a random 20 percent of the Danish population born in 2009 to participate in an online survey. In the survey, everyone is exposed to two hypothetical scenarios -- one in which girls in a Danish school on average outperform boys in an exam, and one in which boys outperform girls. For each scenario, we elicit i) perceived returns to effort in the exam and ii) perceived returns to intelligence in the exam. Our hypothesis is that the returns to effort (intelligence) are perceived as higher (lower) when girls (boys) outperform boys (girls). We also elicit perceived effort costs among the girls (boys) in the two hypothetical scenarios from female (male) respondents in order to test whether effort costs of one's own gender are perceived as lower (higher) when one's own gender group outperforms the other group.