Table 2.
Sex-differences latent growth curve model results for adult antisocial behavior (AAB) symptoms
| Unstandardized variance components | ||||
|---|---|---|---|---|
| A | C | E | Factors | |
| Men | ||||
| Factors | ||||
| Intercept (e.g. ai) | 0.163 (84%)* | 0.000 (0%) | 0.031 (16%) | — |
| (0.081–0.222) | (0.000–0.077) | |||
| Linear slope (e.g. as)∼ | 0.0004 (18%) | 0.0001 (6%) | 0.0018 (76%)* | — |
| (0.000–0.002) | (0.000–0.001) | (0.0001–0.0035) | ||
| Genetic/environmental correlations (e.g. rA) | −1.0 | 1.0 | −0.54 | |
| Residuals | ||||
| Time 1 (e.g. a1) | 0.009 (3%) | 0.000 (0%) | 0.107 (35%)* | 62% |
| (0.000–0.055) | (0.065–0.151) | |||
| Time 2 (e.g. a2) | 0.011 (3%) | 0.023 (7%) | 0.128 (40%)* | 50% |
| (0.000–0.068) | (0.000–0.061) | (0.099–0.162) | ||
| Time 3 (e.g. a3) | 0.000 (0%) | 0.008 (4%) | 0.073 (36%)* | 60% |
| (0.000–0.037) | (0.033–0.118) | |||
| Women | ||||
| Factors | ||||
| Intercept (e.g. ai) | 0.088 (77%)* | 0.005 (5%) | 0.021 (18%) | — |
| (0.030–0.121) | (0.000–0.057) | (0.000–0.045) | ||
| Linear slope (e.g. as)∼ | 0.0006 (38%) | 0.0002 (11%) | 0.0008 (51%) | — |
| (0.000–0.001) | (0.000–0.001) | (0.000–0.002) | ||
| Genetic/environmental correlations (e.g. rA) | −0.46 | −1.0 | −0.65 | |
| Residuals | ||||
| Time 1 (e.g. a1) | 0.000 (0%) | 0.000 (0%) | 0.076 (41%)* | 59% |
| (0.056–0.098) | ||||
| Time 2 (e.g. a2) | 0.000 (0%) | 0.024 (12%) | 0.084 (41%)* | 47% |
| (0.000–0.041) | (0.069–0.102) | |||
| Time 3 (e.g. a3) | 0.000 (0%) | 0.000 (0%) | 0.058 (45%)* | 55% |
| (0.032–0.084) | ||||
A, C and E represent genetic, shared and non-shared environmental influences respectively. The intercept factor is composed of the variance in AAB that is common or stable across time. The slope factor captures systematic, linear change in AAB over time. Unstandardized estimates are presented, followed by the proportion of variance accounted for (the latter is in parentheses). Both factors were decomposed into their genetic and environmental components, and therefore each row sums to 100% of the variance within that factor. Genetic and environmental correlations between factors are also indicated (of note, these are presented for informational purposes but are not particularly meaningful in the present study, as influences on the slope (i.e. E) and intercept (i.e. A) seemed to be fundamentally distinct). The residual estimates index the variance remaining in AAB at each assessment after that contributed by the factors is accounted for; accordingly, the variance contributed by the factors is necessary for the rows to sum to 100%. For example, the factors (both the intercept and the slope) contributed 62% of the total variance in men's AAB at time 1, while the residuals contributed 38% (i.e. 3% of which is genetic and 35% of which is non-shared environmental in origin). For those estimates greater than zero, 95% confidence intervals are presented below the variance estimates in parentheses (note that these cannot be smaller than zero, as variances are necessarily positive). Statistically significant variance components (p<0.05) are bolded and indicated with an asterisk. We also explicitly evaluated the significance of the linear slope overall. A tilda (∼) indicates that the slope contributes a significant proportion of variance (p≤0.05) to the growth model (i.e. constraining all slope parameters to be zero resulted in a significant decrease in fit).