Table A1.

Depression (BDI), anxiety (BAI), ethnicity, sex and methylation and attachment

A. We note that all analyses contain a variable labeled Asian American versus others. The term “others” refers to all non‐Asian‐American participants. This variable was added because the only differences detected in genotype or in SLC6A4_1pc methylation between ethnic groups was between Asian Americans and Euro‐Americans (χ2 = 22.606, P < 0.001, F = 8.031, P = 0.006, respectively). We added all remaining ethnic groups to the “other” category out of an abundance of caution. Although this is primarily a female study, some differences in SLC6A4_1pc methylation by sex were detected (F 1,98 = 11.686, P = 0.001). Accordingly all SLC6A4_1pc methylation analyses added sex to the analysis. Although no differences in sex or ethnicity were detected in SLC6A4_2pc methylation we controlled for both variables, as well.

B. Neither anxiety nor depression was related to categorical attachment classification (four way [F, D, E, U] or two way [U vs. F]), SLC6A4_1pc or SLC6A4_2pc methylation, sex or ethnicity. We initially detected a significant difference in depression between organized versus unresolved participants (F 1,99 = 4.470), P = 0.037), but significance disappeared once we controlled for SES. Low‐SES participants did have higher levels of depression than mid/upper SES participants (F 1,98 = 12.90, P = 0.001) but it should be remembered that overall this was not a highly depressed sample (93% scored as having no or mild depression). Nevertheless, out of an abundance of caution, we ran the univariate ANOVA models with depression and found depression insignificant.

C. Continuous UL, trauma and loss or trauma was unrelated to anxiety. UL (w/o trauma) (F 1,78 = 0.667, P = 0.417) and UT (F 1,99 = 0.713, P = 0.749) was unassociated with depression. UL (including those participants with loss and trauma) and UL and trauma was initially associated with depression (F 1,97 = 5.824, P = 0.02); F 1,99 = 6.305, P = 0.014 respectively), but after SES was controlled for or added to the model significance disappeared. Moreover, the association between depression and UL, and loss or trauma, was due to one outlier. Removing that subject eliminated the association. In addition, we ran the models (UL and loss and trauma) with depression and found depression was insignificant. Accordingly, depression was removed from all analyses.

BAI, Beck Anxiety Inventory; BDI, Beck Depression Inventory; SES, socioeconomic status; UL, unresolved loss; UT, unresolved trauma.