Brief Report: Predicting Inner Speech Use Amongst Children with Autism Spectrum Disorder (ASD): The Roles of Verbal Ability and Cognitive Profile

It has been suggested that using such a subtraction procedure to derive the size of the PSE does not take into account baseline differences in short-term memory at different ages/VMA levels (Jarrold and Tam 2009). This potential problem can be corrected by employing a “PSE proportion score”, derived by dividing the number of rows correctly recalled in the phonological condition by the number of rows correctly recalled in the control condition. In the current sample, the correlation between the basic PSE size score and the PSE proportion score was .93. Furthermore, we can confirm that results from the regression analyses were essentially identical when a PSE proportion score was used instead of a PSE size score. As such, we simply report the results from the latter analyses in which PSE size was included as the dependent variable.

Lidstone (personal communication) questioned whether, despite the fact that absolute VMA was a better predictor of PSE Size than Cognitive Profile amongst the whole sample, Cognitive Profile might accurately predict the degree of verbal mediation shown by participants with VMAs over 7 years only. However, we can confirm that a regression analysis conducted only amongst participants with VMAs over 7 years yielded essentially identical results to those regressions conducted amongst the whole sample. Full details of this latter regression analysis are available from the first author upon request.