Extended Data Fig. 9: Parameter estimates and inferential statistics for the entire sample (privileged and marginalized students) in Experiment 5 when accounting for the non-independence due to classroom.
From: Exposure to peers’ pro-diversity attitudes increases inclusion and reduces the achievement gap
Note: We estimated a linear mixed-effects model in which we regressed the outcome variable on the two dummy codes (see Experiment 5 Results) and included a by-classroom random intercept. These results should be interpreted with caution, because 9 (out of 51, that is, 18%) classrooms had 2 or fewer respondents and 19 (37%) had 5 or fewer respondents, leading to substantial volatility in the classroom means (and thus relatively large standard errors of the parameter estimates). With error degrees of freedom around 40, the inferential tests are underpowered. All scales are scored such that higher values indicate more positivity/inclusiveness. “(R)” denotes scales that have been inverted from their original scoring to fit this standard.
