Results Groups
Children were recruited into the LLI group if they had a diagnosis of LLI and to the TD group if their teacher had no concerns over their development. However, group membership did not remain stable after a criterion was applied for inclusion in the LLI group, defined as having z scores at least as low as −1 on more than one of the standardized language tasks. Three children recruited with a diagnosis of LLI did not meet criterion for impairment, and one child recruited into the TD group did. This resulted in 43 children being included in the TD group (mean age: 8 years 1 month), 23 of whom were boys and 20 of whom were girls, and 25 children in the LLI group (mean age: 8 years 10 months), 21 of whom were boys and four of whom were girls. As the four children who switched group membership may have shown slightly different profiles compared with their fellow group members, all analyses were rerun excluding these four participants. There was no effect on the pattern of results reported. Table 1 gives z scores for each standardized task for both groups and the results of fully factorial univariate analyses of covariance run on each test; main effects of Group (TD and LLI) and interactions between Group and Chronological Age (CA; in months) are given. As expected, either a main effect of Group or an interaction between Group and CA is seen for each standardized test of language but not for the Pattern Construction subscale from the BAS-II, supporting criterion for LLI group membership. Whenever analyses are performed with age (chronological or language) as a variable, age in months has been rescaled such that 0 is the age of the youngest child in the LLI group. This transformation allows for easier interpretation of group effects as groups are compared at the youngest age at which both are represented.