The NR3C1 moderation of intervention effects was only observed in AAs, but not in EAs. One possible explanation relates to population stratification in genome structure and allele frequencies, which differs between ethnic groups (Cardon and Palmer 2003). The susceptibility locus identified among AAs (rs12655166) may be a proxy for an unmeasured variant that is in high LD with rs12655166 in the AAs but not in the EAs. Therefore, it is possible that we did not effectively measure and test the NR3C1 variant responsible for differential susceptibility to intervention among EAs. Relatedly, different from Albert et al. (2015a, 2015b) who found a SNP (rs10482672) moderated intervention effects among EAs on a static outcome (any externalizing psychopathology at age 25), the current study found another SNP (rs12655166) moderated intervention effects among AAs on a dynamic and different outcome (slope of alcohol abuse) in a different developmental period (i.e., adolescence). The lower levels and growth of alcohol use among AAs than for EAs, which could be attributed to differential community characteristics, peer and family processes (e.g., peer substance use, family support; Wallace 1999;
