The Alcohol-Use Disorders Identification Test (AUDIT) is designed to assess hazardous alcohol use and alcohol-related problems. The AUDIT has 10 items (Saunders et al, 1993) and responses are on a likert scale ranging from 0-4. The 10 items are summed for a total score with a possible range of 0-40, with higher scores indicating greater likelihood of hazardous drinking behavior. A total score of 8 or higher was used as a binary variable to identify participants with hazardous drinking (score 8+). Outcome was the proportion of respondents exceeding the clinical cut point.

Binge eating was measured with the Eating Pathology Symptoms Inventory (EPSI; Forbush et al., 2013). The binge eating subscale was used in the study, which includes items on overeating and loss of control eating. The binge eating subscale has 8 items with Likert scale responses from 0=never to 4= very often. Items are summed for a scale score ranging from 0-32. Higher scores indicate more frequent experiences with binge eating behavior.

Stress was measured using the Depression Anxiety Stress Scale-21 (DASS-21), a 21 item measure designed to assess depression, anxiety and stress (Lovibond & Lovibond, 1995). The stress scale score was used to assess stress. Items are on a 4-point Likert scale ranging from 0-3 and can be summed for a scale score ranging from 0-21. Higher scores indicate greater stress.

The Adolescent Reinforcement Survey Schedule - Alcohol Use Version assesses the frequency of past-month engagement in and enjoyment derived from 45 activities (Hallgren et al, 2016). Each question is posed twice - once to assess the frequency and enjoyment of the activity while using alcohol and the once to assess the frequency and enjoyment of the activity while not using alcohol. Items range from 0-4. Frequency and enjoyment items are summed to form respective scores. From these scales, two subscales are created for alcohol-related reinforcement and alcohol-free reinforcement, calculated as the cross product between frequency and enjoyment items for alcohol-related and alcohol-free questions. The two subscales were used to calculate the outcome, the total reinforcement ratio (TRR) between alcohol-related and alcohol-free reinforcement. The ratio has values between 0 and 1, with higher values indicating more relative enjoyment of activities when using alcohol.

Moderator: The Drinking Motives Questionnaire-Revised (DMQ-R) is designed to measure the relative frequency of drinking for four distinct reason motives: enhancement, social, conformity, and coping (Cooper, 1994; Cox & Klinger, 1988). The study used the coping motives scale, which contains 5 items. Items are assessed on a Likert scale ranging from 1-5 and are summed for form a scale score, which higher scores indicating greater endorsement of drinking to cope with stress/distress.

Outcome: The outcome used in the analysis was the Alcohol Use Disorder Identification Test (AUDIT) total score; we originally planned to use AUDIT-Consumption and AUDIT-Problems subscales for two separate analyses and outcomes (in parallel with other moderation analyses); however the models did not converge. Therefore, AUDIT total score was used as the outcome in analyses.

Academic performance was measured using participant self-reported grade point average (GPA). College freshmen first received their GPA in the spring semester of college, which was at the 5 month follow up. Thus, we examined whether there were significant differences in GPA across treatment and control conditions at 5 month adn 17 month follow ups.

The moderator used in analyses was sex. The outcome was the Alcohol Use Disorder Identification Test- Consumption (AUDIT-C) subscale score.

The moderator in the model was sex. The outcome was the Alcohol Use Disorder Identification Test-Problems (AUDIT-P) subscale score

Total of all reporting groups

Sex

Race and ethnicity

Participants could select more than one race or ethnicity

Courses were cluster-randomized to the standard orientation (SO) course condition

The Time Line Follow Back-Computerized (TLFB-C) assessment was used to measure alcohol consumption in the past 30 days (Sobell & Sobell, 2008), or since the prior measurement period. The measure was used to obtain the number of days during which individuals engaged in high-intensity drinking of 8+ drinks for males or 10+ drinks for females per drinking occasion. The number of days participants engaged in high intensity drinking was summed per measurement period, and converted to a rate to reflect the number of high intensity drinking days out of the number of days in the measurement period (high intensity drinking days/days in measurement period). The rate was used because there could be slightly different numbers of days across measurement periods, depending on when participants completed the assessment.

Alcohol-Use Disorders Identification Test- Problem questions (AUDIT-P) are the last 7 items of the full AUDIT that assess increased salience of drinking, morning drinking, guilt after drinking, blackouts, alcohol-related injuries, and drinking that others are concerned about (Saunders et al, 1993; O'Hare & Sherrer, 2005). Responses are on a likert scale ranging from 0-4. The 7 items were summed for a total score on the AUDIT-P, with a possible range of 0-28, with higher scores indicating greater alcohol-related problems.

The Alcohol-Use Disorders Identification Test (AUDIT) is designed to assess hazardous alcohol use and alcohol-related problems. The AUDIT has 10 items (Saunders et al, 1993) and responses are on a likert scale ranging from 0-4. The 10 items are summed for a total score with a possible range of 0-40, with higher scores indicating greater likelihood of hazardous drinking behavior. A total score of 8 or higher was used as a binary variable to identify participants with hazardous drinking (score 8+). Outcome was the proportion of respondents exceeding the clinical cut point.

Depression was measured using the Depression Anxiety Stress Scale (DASS-21), a 21 item measure designed to assess depression, anxiety and stress (Lovibond & Lovibond, 1995). The measure provides scales for depression, anxiety, and stress and conceptualizes the difference between normal and clinical populations as a matter of degree. The depression scale assesses dysphoria, hopelessness, devaluation of life, self-deprecation, lack of interest / involvement, anhedonia and inertia with 7 items. Items are on a 4-point Likert scale ranging from 0-3 and can be summed for a scale score ranging from 0-21. Higher scores indicate greater depression.

Binge eating was measured with the Eating Pathology Symptoms Inventory (EPSI; Forbush et al., 2013). The binge eating subscale was used in the study, which includes items on overeating and loss of control eating. The binge eating subscale has 8 items with Likert scale responses from 0=never to 4= very often. Items are summed for a scale score ranging from 0-32. Higher scores indicate more frequent experiences with binge eating behavior.

Stress was measured using the Depression Anxiety Stress Scale-21 (DASS-21), a 21 item measure designed to assess depression, anxiety and stress (Lovibond & Lovibond, 1995). The stress scale score was used to assess stress. Items are on a 4-point Likert scale ranging from 0-3 and can be summed for a scale score ranging from 0-21. Higher scores indicate greater stress.

Delay discounting was calculated from a computer adjusting delay discounting task that asked participants to choose between smaller immediate rewards and larger, delayed rewards. The reward used in the task was hypothetical money. Mazur's hyperbolic function, V = A/ 1 + kD, was used to estimate each participant's delay discounting rate (i.e., k) for use in analyses. For this formula, V is the discounted value of a delayed reward (i.e., indifference point), A is the reward amount, D is the delay in days, and k represents the estimated delay discounting rate. Greater k values indicated stronger discounting and a preference for immediate monetary rewards. K values across the sample ranged from -11.06 to 2.43. Negative values generally indicate stronger discounting, whereas positive values typically represent less steep discounting.

The Adolescent Reinforcement Survey Schedule - Alcohol Use Version assesses the frequency of past-month engagement in and enjoyment derived from 45 activities (Hallgren et al, 2016). Each question is posed twice - once to assess the frequency and enjoyment of the activity while using alcohol and the once to assess the frequency and enjoyment of the activity while not using alcohol. Items range from 0-4. Frequency and enjoyment items are summed to form respective scores. From these scales, two subscales are created for alcohol-related reinforcement and alcohol-free reinforcement, calculated as the cross product between frequency and enjoyment items for alcohol-related and alcohol-free questions. The two subscales were used to calculate the outcome, the total reinforcement ratio (TRR) between alcohol-related and alcohol-free reinforcement. The ratio has values between 0 and 1, with higher values indicating more relative enjoyment of activities when using alcohol.

Moderator: The Drinking Motives Questionnaire-Revised (DMQ-R) is designed to measure the relative frequency of drinking for four distinct reason motives: enhancement, social, conformity, and coping (Cooper, 1994; Cox & Klinger, 1988). The study used the coping motives scale, which contains 5 items. Items are assessed on a Likert scale ranging from 1-5 and are summed for form a scale score, which higher scores indicating greater endorsement of drinking to cope with stress/distress.

Outcome: The outcome used in the analysis was the Alcohol Use Disorder Identification Test (AUDIT) total score; we originally planned to use AUDIT-Consumption and AUDIT-Problems subscales for two separate analyses and outcomes (in parallel with other moderation analyses); however the models did not converge. Therefore, AUDIT total score was used as the outcome in analyses.

Academic performance was measured using participant self-reported grade point average (GPA). College freshmen first received their GPA in the spring semester of college, which was at the 5 month follow up. Thus, we examined whether there were significant differences in GPA across treatment and control conditions at 5 month adn 17 month follow ups.

The moderator used in analyses was sex. The outcome was the Alcohol Use Disorder Identification Test- Consumption (AUDIT-C) subscale score.

Due to the very small number of individuals reporting sex as intersex, we did not have sufficient statistical power to include this group in the moderating analyses by sex. Therefore, only results for groups of male and female are reported herein.

The moderator in the model was sex. The outcome was the Alcohol Use Disorder Identification Test-Problems (AUDIT-P) subscale score

Due to the very small number of individuals reporting sex as intersex, we did not have sufficient statistical power to include this group in the moderating analyses by sex. Therefore, only results for groups of male and female are reported herein.

Courses were cluster-randomized to the standard orientation (SO) course condition