Skip to main content
. Author manuscript; available in PMC: 2015 Jul 1.
Published in final edited form as: Alcohol Clin Exp Res. 2014 Jun 13;38(7):2024–2033. doi: 10.1111/acer.12446

Table 6.

Parameter Estimates from Model 1a

Estimate 95%CI p
Autoregressive
  ASB12→ASB13b 0.787 0.681, 0.893 <0.001
  ASB13→ASB15 0.751 0.688, 0.814 <0.001
  ASB15→ASB17 0.680 0.609, 0.751 <0.001
  ALC12→ALC13b 0.673 0.595, 0.751 <0.001
  ALC13→ALC15 0.496 0.453, 0.539 <0.001
  ALC15→ALC17 0.508 0.461, 0.555 <0.001
Cross-lagged
  ALC12→ASB13 −0.059 −0.177, 0.059 0.319
  ALC13→ASB15 0.049 −0.018, 0.116 0.155
  ALC15→ASB17 0.116 0.043, 0.189 0.002
  ASB12→ALC13 0.063 −0.033, 0.159 0.204
  ASB13→ALC15 0.274 0.221, 0.327 <0.001
  ASB15→ALC17 0.124 0.067, 0.181 <0.001
Correlationc
  ASB12-ALC12 0.673 0.622, 0.724 <0.001
Residual correlationc
  ASB13-ALC13 0.239 0.116, 0.362 <0.001
  ASB15-ALC15 0.455 0.384, 0.526 <0.001
  ASB18-ALC17 0.648 0.579, 0.717 <0.001
Fit indices χ2 = 11.421, df = 5,425, p < 0.001
CFI = 0.966, TLI = 0.967
RMSEA [90%CI] = 0.016 [0.016, 0.017]

CFI, comparative fit index; RMSEA, root mean square error of approximation; TLI, Tucker-Lewis index.

a

In Model 1, all parameters were constrained to be equal between the male and female subgroups.

b

ASB and ALC are factors for antisocial behavior and alcohol use, respectively.

c

Correlations were estimated at the initial age (age 12), and residual correlations were estimated for the rest of the ages because ALC and ASB factors at ages 13 to 17 were endogenous factors (i.e., dependent latent variables).