To assess the concordance of the retrospective SSAGA interview peer drinking responses for ages 12-17 with current peer drinking reported in C-SSAGA assessments, we compared the first adult SSAGA response and the maximum value from all C-SSAGA assessments among individuals with at least one adult and one child questionnaire. For the 996 participants with both adult and child interviews, 73% of peer drinking assignments had the same dichotomous variable (none/few vs most/all best friends). This concordance demonstrates that our retrospective approach of using the first SSAGA interview when available is a reasonable strategy to assess peer drinking across adolescence. It also shows that for the 12% of participants without a single adult SSAGA assessment, using the maximum value from CSSAGA assessments reasonably estimates the proportion of best friends drinking from ages 12-17.
