To make sure that the addressed associations were specific for regular use, rather than for substance use in general, analyses were repeated comparing regular compound screening assay users to experimental or less regular users. At age 15–18, regular alcohol and cannabis use were reported by, respectively, 12.2% and 6.3% of the adolescents. Boys were more likely than girls to be regular users of alcohol (χ2 (2 df, N = 1192) = 16.16, p < .01) and

cannabis (χ2 (2 df, N = 1192) = 23.82, p < .001). Mean scores or percentages of the variables used are shown in Table 1. For descriptive purposes, we presented the mean of the unstandardized scores. Genotype frequencies of DRD2 and DRD4 are depicted in Table 2. Allele frequencies were calculated UMI-77 and analyzed for deviations from Hardy–Weinberg equilibrium (HWE) using χ2-tests. No deviations from HWE were detected (p = 0.31 for DRD2 and p = 0.94

for DRD4). Because of the very small number of regular alcohol and cannabis users with two copies of the genetic risk markers DRD2 A1 and DRD4 7R, subsequent analyses were performed comparing the individuals carrying at least one genetic risk factor with individuals carrying no genetic risk factor. This has also been done in many previous studies ( Conner et al., 2010, Conner et al., 2005, Sakai et al., 2007 and van der Zwaluw et al., 2009). The univariate analyses (not depicted in a Table) showed that the A1 allele of the DRD2 TaqIA polymorphism had no direct effect on regular alcohol (OR = 0.98, 95%CI = 0.57–1.70,

p = 0.95) or cannabis use (OR = 0.91, 95%CI = 0.52–1.61, p = 0.75). Similarly, L-DRD4 was not significantly related to regular alcohol (OR = 0.65, 95%CI = 0.37–1.11, Calpain p = 0.11) or cannabis use (OR = 0.79, 95%CI = 0.44–1.41, p = 0.43). DRD2 by parenting measure interactions did not yield any significant associations, indicating that rejection, overprotection, and emotional warmth did not modify the effect of the A1 allele of the DRD2 TaqIA polymorphism on regular alcohol or cannabis use (see Table 3). DRD4 by parenting measure interactions resulted in a significant interaction between DRD4 and emotional warmth. Regression analyses separate for S-DRD4 and L-DRD4 individuals indicated that a higher level of emotional warmth was associated with regular alcohol (versus irregular) consumption in carriers of the L-DRD4 (OR = 1.62, 95%CI = 1.12–2.33, p = 0.01). In S-DRD4 individuals, our findings pointed in the direction of an inverse association between emotional warmth and regular alcohol consumption, though this was not significant at p < 0.05 (OR = 0.84, 95%CI = 0.68–1.03, p = 0.09). Because adjusting for parental substance use might have ruled out part of the variance explained by genetic factors, analyses were repeated without adjusting for parental substance use. These analyses yielded comparable results.