Testing the generalized slowing hypothesis in specific language impairment.

This study investigated the proposition that children with specific language impairment (SLI) show a generalized slowing of response time (RT) across tasks compared to chronological-age (CA) peers. Three different theoretical models consistent with the hypothesis of generalized slowing--a proportional, linear, and nonlinear model--were examined using regression analyses of group RT data. Each model was an excellent fit with the RT data. The most parsimonious model indicated that the SLI group was proportionally slower than the CA group. Mean RTs of the SLI group were about one fifth slower across tasks than the CA group's mean RTs. Less slowing was evident for a subgroup of young children with expressive SLI than for children with mixed (expressive and receptive) SLI. Although the mean RT data reflected many individual SLI children's RT performance, not all SLI children showed generalized slowing.