Randomization Method
The study uses a fixed-effects blocked individual randomization design. Randomization cells will be defined by a subset of 5 strata measured at baseline. In order of importance, the strata are: URM-status (binary variable where a student is marked as URM if they identify with any of the following ethnicities - Black, Latino, Mexican, Native American/Pacific Islander, American Indian); gender (binary variable defined as female vs non-female); course section (four values, defined by the section of the course the student is enrolled in); highest level of education attained by a parent (three bins defined as "less than a four-year degree", "four-year college degree", "post-graduate degree"); and international-status (binary variable defined as international vs. domestic). Missing values in each of the demographic strata (URM-status, gender, parental education, and international-status) are treated as their own "bin".
At minimum, the sample will be stratified by URM-status, gender, and course sections. If sample size allows, the sample will be additionally stratified by bins of parental education and by international-status (in that order). Either the latter or both of these strata may be omitted if including them would lead to a large number of randomization cells with less than two observations.
Within each cell, randomization occurs at the student level. Within each cell, every student has a 50% chance of being assigned to the treatment or control group. A detailed explanation of the randomization process is included in the pre-analysis plan.