SDIT was normalised by a logarithmic transformation (e.g., Alexander & Mackenzie, 1992; Crawford, Deary, Allan, & Gustafsson, 1998; Deary, McCrimmon, & Bradshaw, 1997).
There were significant sex differences in SDIT and all IQ composite scores.
MLEs of the phenotypic correlations for MZ and DZ groups for SDIT and FIQ are presented in Table 1 -- these are based on data where multivariate outlying twin pairs have been removed.
The MLE of the phenotypic correlation between SDIT and FIQ was -.35 (N=738), indicating that a faster IT was associated with a higher IQ score.
Low factor loadings in the full AE model were tested for significance and dropped where appropriate; only the unique environment path between SDIT and FIQ could be dropped.
The first additive genetic factor accounted for 36% of variance in SDIT and 32% of variance in FIQ, such that genetic influences that decreased SDIT increased FIQ.
A common genetic factor mediated the relationship among SDIT (36%), PIQ (30%), and VIQ (18%).
It may surprise some readers that we report IT values as high as 774 ms, within the transformed SDIT distribution such high ITs were not outlying.