The extent to which youth attend Tier 2 evidence-based intervention is an important dimension of implementation. This study examined attendance patterns of 369 middle schoolers involved in a randomized trial testing the impact of Coping Power, an evidence-based Tier 2 preventive intervention. We conducted latent profile analysis to examine student attendance at the 25 Coping Power sessions and found three attendance patterns: 69.9% had "high and stable" attendance, 19.5% of youth had "moderate and modestly" declining attendance, and 10.6% had "poor and sharply declining" attendance. We then examined whether students of a particular gender and race or in single-gender/race intervention groups were more likely to demonstrate certain attendance patterns and whether there were mean differences across attendance patterns on student behavioral risk, affect, and group engagement; group characteristics (e.g., group behavioral norms); and individual contacts with the group leader. Analyses indicated students with the "poor and sharply declining" attendance pattern had higher early-session negative affect than students with the other two attendance patterns and were less likely to be in gender-balanced groups than students with "moderate and modestly declining" attendance. Students with "moderate and modestly declining" attendance spent more time in contacts with group leaders than students with "high and stable" attendance. Students with high attendance were in groups with the highest early-session group attendance rates. Implications of these findings for adaptation and tailoring of the Tier 2 Coping Power program are discussed. [This paper will be published in "Journal of Applied School Psychology."]