Do Other Cognitive Functions Account for Expressive Language Discrepancies?
An Expressive Language Composite (ELC) was computed, equal to the mean of expressive language z scores for children from both samples with at least 2 expressive language scores. ANCOVAS predicting the ELC were conducted controlling for cognitive scores. We first tested whether RM could account for expressive delays in IDMs in a subsample of twins. Twin IDMs did perform more poorly than controls on RM (F1,413 = 5.306, P = .02) but, the effect of GD on the ELC remained significant (F2,412 = 6.417, P = .01) once RM was controlled. Using the combined samples, we then tested whether NVIQ (48 months, QLSCD; 60 months, QNTS) and short-term memory (60 months) accounted for expressive delays in IDMs. Groups did differ on NVIQ (F1,2310 = 3.733, P = .05), but not on short-term memory (F1,2166 = 2.856, P = .09). However, after controlling for both NVIQ and short-term memory, IDMs still scored lower than controls on the ELC (F3,2136 = 5.950, P = .01). Do Genes and Maternal Education Moderate the Ef
