Table 2.

Changes in Model Fit After Dropping Individual Paths via the Path-Dropping Process for the Model Shown in Figure 2, Resulting in the Model Shown in Figure 3

Path set	Change in fit after dropping paths
	ΔAIC	Cumulative no. paths dropped (Δdf)	p
Reading: within-trait	−3.91	4	.39
IQ: within-trait	−9.76	11	.35
Reading → IQ: cross-lagged	−12.57	17	.21
IQ → reading: cross-lagged	−24.89	26	.40
IQ ↔ reading: covariance	−26.16	27	.42

Note. The saturated model included 45 paths; the final model included 18.

AIC = Akaike information criterion; IQ = intelligence; Reading = reading ability. Reference model AIC = 5905.05 on 27 df.