Temperamental Differences between Autistic and Typical Children : A Meta-Analysis

Background:

Although autism is no longer considered a rare condition, challenges remain in the diagnosis process due to an important clinical heterogeneity in the phenotype. Some authors have suggested that temperament may be a useful construct for understanding early developmental differences in infants at risk for ASD.

Objectives: The present meta-analysis aimed to determine if ASD children differed from normative children on three temperamental factors (Negative Affectivity, Surgency and Effortful Control).

Methods:

A comprehensive literature search was conducted to identify published articles and unpublished dissertations using six electronic databases: ProQuest, Google Scholar, Medline, PsycInfo, Cinhal, Embase, using the following keywords : autis* OR ASD OR “autism spectrum disorder” OR Asperger OR TED OR PDD AND temperament*. Literature search covered a period of time from January 1^st, 1970 to July1^st, 2017. To be selected for this meta-analysis, a study had to (a) include a sample of autistic children aged between 2 and 16 years old, which was compared to a normative group ; (b) assess child temperament with validated temperament questionnaires. Studies also had to meet minimal criteria regarding participant ascertainment, such as having a confirmed diagnosis of autism spectrum disorder derived from a standard method (e.g. ADOS) or by a mental health professional. Studies including high-risk siblings (not diagnosed with ASD) were not retained, in order to limit potential bias and heterogeneity among studies. Analyses were performed using the Comprehensive Meta-analysis software 2.0 (Borenstein, Hedges, Higgins, & Rothstein, 2005). Based on Borenstein and colleagues (2009) recommendation, when original studies compare two groups and differences between their means are available, we used Cohen’s d to transform and combine every available statistics.

Results: Twenty-one studies were included in this meta-analysis. Results showed that autistic children and normative children differed significantly on the Negative Affectivity (d= -.36; k=19; 95%CI: -.57 to -.15; Q= 127.14; p < .001) and the Effortful Control factors (d= -.65; k=20; 95%CI: -.93 -.36; Q= 244.74; p < .001). No differences emerged for the Surgency factor (d = -.24; k=20; 95%CI: -.54 to .54; Q= 246.53; p < .001). Further analyses showed that these associations are moderated by age, country and publication year.

Conclusions:

This meta-analysis confirms that autistic children show both higher levels of Negative Affectivity (greater temperamental reactivity) and lower levels of Effortful Control (less capacities to regulate this reactivity), in comparison to normative children. Temperament could thus be an important variable to consider and document to help identify children who are at a higher risk of having ASD.