TABLE III.

Model-Fitting Analyses Within Each Wave for Reactive and Proactive Aggression, at Ages 9–10 and 11–14 Years

	Overall fit				Model difference test			Parameter estimates (95% CI)
	−2LL	df	AIC	BIC	Δχ2	Δdf	P	A2	C2	E2
Wave-1a								Reactive aggression
Saturated model	3,328.19	1,222	884.19	−2,268.39
Males ≠ Females	3,341.43	1,231	879.43	−2,290.74	13.24	9	.15
Males = Females	3,344.95	1,234	876.95	−2,298.63	16.76	12	.16	.26 (.04–.49)	.27 (.08–.45)	.46 (.39–.55)
								Proactive aggression
Saturated model	3,357.73	1,223	911.73	−2,256.84
Males ≠ Females	3,367.16	1,231	905.16	−2,277.87	9.42	8	.31
Males = Females	3,372.58	1,234	904.58	−2,284.82	14.84	11	.20	.32 (.09–.55)	.21 (.02–.39)	.47 (.40–.56)
Wave-2								Reactive aggression
Saturated model	2,348.28	857	634.28	−1,435.99
Males ≠ Females	2,363.36	866	631.36	−1,455.86	15.08	9	.09
Males = Females	2,365.93	869	627.93	−1,463.71	17.65	12	.13	.43 (.16–.65)	.15 (.00–.36)	.42 (.34–.52)
								Proactive aggression
Saturated model	2,374.81	858	658.81	−1,425.77
Males ≠ Females	2,384.56	866	652.56	−1,445.26	9.75	8	.28
Males = Females	2,386.42	869	648.42	−1,453.46	11.61	11	.40	.48 (.21–.65)	.08 (.00–.30)	.44 (.35–.54)

−2LL, −2(log-likelihood); AIC, Akaike’s Information Criterion; BIC, Bayesian Information Criterion; Δχ², difference in log-likelihoods between nested models; Δdf, change in degrees of freedom.

^aUnivariate results on reactive and proactive aggression, Wave-1 have previously been reported on these data (Baker et al., 2008).