Table 3.

Model Fit Comparisons for the Moderation Models.

Model	−2lnL (df)	χ2Δ (df)	p	AIC
Full Model	2706.76 (984)	--	--	--
Drop Sex Moderator	2711.06 (990)	4.3 (6)	.64	−7.70
Drop Puberty Moderator	2713.90 (990)	7.14 (6)	.30	−4.86
Drop Sex × Puberty Moderator	2710.13 (987)	3.37 (3)	.33	−2.63
Drop Moderators on A, C or E:
Drop all moderators on A	2711.25 (987)	4.99 (3)	.17	−1.01
Drop all moderators on C	2706.76 (987)	0 (3)	1.00	−6.00
Drop all moderators on E	2707.39 (987)	0.63 (3)	.89	−5.37
Drop all moderators on C and E	2710.39 (990)	0.63 (6)	.99	−11.37
Drop all Moderators on A, C and E	2713.99 (993)	7.23 (9)	.61	−10.77

Note. χ²_Δ = chi-square change; Full Model = model with paths and all three moderators. Each nested model is compared to the full model when calculating the χ²_Δ and degrees of freedom. The best-fitting model is noted by bolded text; unstandardized estimates (with 95% confidence intervals in parentheses) from this model are as follows: a = .69 (.54, .84), c = .00 (−.33, .33), e = .68 (.63, .74), β_XS = .08 (−.09, .34), β_YS = 0, β_ZS = 0, β_XD = .05 (−.16, .27), β_YD = 0, β_ZD = 0, β_XSxD = −.30 (−.58, −.02), β_YSxD = 0, β_ZSxD = 0,