Is the Ability to Integrate Parts into Wholes Affected in Autism Spectrum Disorder?

Since the ASD sample size was 20, there was a risk concern that the findings might have been pulled by an outlier and/or the use of covariates. We tightened our cutpoint from 2.5 to 2.0 SD from the mean of the ST2–ST1 RT difference score. One participant’s SD was 2.1; we labeled him a potential outlier. We retested the strength of the correlation between SRS Motivation Score and the ST2–ST1 RT difference score remained between .51 and 5.3 (p < .05) with or without the potential outlier and/or the covariates. We used the comparison of two correlations with one variable in common from the same sample (Meng et al. 1992) to assure the reader that the ST2–ST1 RT difference score was in fact more strongly correlated with the Motivation subscale than either (a) the Overall SRS score (z = −2.369. p = .01); (b) the Mannerisms subscale, highlighted because it is a measure of non-social repetitive behavior (z = −2.734, p = .01); and (c) the Awareness Score, which was the next strongest correlation (z = −2.00, p < .05). Finally, in terms of the ST2–ST1 RT difference score, when all 20 ASD participants were included, the group effect was F(1, 38) = 16.29, p < .001. The group effect remained significant (p < .05) without the potential outlier and/or the covariates.

One of the reviewers suggested an alternative interpretation of our data which we think is intriguing: “Individuals with ASD who find integration of information difficult may find they ‘get things wrong’ socially and so lose motivation to try”.