Are “Strange Stories” Strange to a Non-Clinical Population? Relationships Between AQ and Physical and Mental State Reasoning

Background:  Performance associated with autism spectrum disorders has been demonstrated in individuals with high levels of autistic traits but without a diagnosis of ASD.  These include lexical effects (Stewart & Ota, 2008) and compensation for co-articulation (Yu, 2010) in speech perception; cognitive flexibility in set-shifting (Goken et al., 2014); and social cognition (Goken et al., 2014; Sasson et al., 2012).  One of many lines of evidence indicating that individuals with ASD have a specific deficit in reasoning about the mental states of others comes from the “strange stories” test (Happé, 1994) which reveal that children and adolescents with ASD diagnoses score particularly poorly on stories that require reference to mental states compared with stories which require causal reasoning about physical states only. It is not known whether this pattern of results extends to those members of the general population who score highly on autistic traits as measured by instruments such as the Autism Spectrum Quotient (AQ; Baron-Cohen et al., 2001).

Objectives: We hypothesized that participants without an ASD diagnosis who nevertheless scored highly on the AQ would perform less well on the mental-state stories than participants with low AQ scores, and that there would little or no relationship between AQ and performance on the mental state stories.

Methods: The AQ was completed online by a sample of 177 adult participants, all of whom reported no previous or current ASD diagnosis. Participants also completed both physical and mental-state strange stories), in which individuals with ASD typically perform poorly on the mental-state stories. 

Results: Pearson’s correlation coefficients revealed that AQ was negatively associated to a significant extent with performance on both mental and physical state stories. A stepwise multiple regression analysis showed that – once educational level and verbal ability were accounted for – ability to answer mental-state stories accounted for significant variance in AQ scores once variance associated with ability to answer physical state stories was accounted for, but the converse was not the case: Once variation associated with ability to answer mental-state stories was accounted for, performance on physical state stories no longer accounted for significant variance in AQ scores. 

Conclusions: These results show the utility of running behavioral studies in a general population with the advantages this gives of easy access to large numbers of participants, and with AQ scores as a proxy for ASD-like traits. The results also indicate that – contrary to initial suppositions – ASD-like traits in a general, non-clinical, population (as measured by the AQ) are associated not only with difficulties in reasoning about mental states but also with difficulties in reasoning about physical states when these are presented in a verbal, “strange story” format. Mental-state story scores did, however, account for a unique proportion of the variance in AQ suggesting that individuals who score highly both have a general difficulty with verbally-presented reasoning tasks and a specific difficulty with such tasks when they require explicit reference to the mental states of others.