A-ToM: A New Measure of Theory of Mind in Adults

Background:  A Theory of Mind (ToM) deficit suggests an inability to explain, predict, empathise with and understand behaviours, intents and emotions of others.  Although ToM is believed to be delayed or deficient in children with ASD, adult data highlight considerable variability in the degree of deficit (Baron-Cohen, 2001).  One of the most widely used ToM tests is Happé’s Strange Stories test (Happé, 1994).  The original set of items comprised 24 pencil-and-paper vignettes about everyday situations that require participants to identify motivations that may underpin utterances that are not literally true.  These vignettes require participants to recognise expressions of sarcasm, a white lie, a figure of speech, and so on.  It has been argued, however, that the variability in ToM performance among adults with ASD results from such measurement tools allowing adults to “hack out” a strategy that provides an appropriate response (Frith et al., 1991) rather than indicating ToM variability per se.

Objectives:  In line with this proposition, we developed a tool that does not allow participants time to apply analytic reasoning skills to determine the appropriate response.  This tool (Adult Theory of Mind: A-ToM) comprises acted-out “strange stories” (i.e., films) from Happé’s test using visual and auditory contextual cues to assess ToM.  The A-Tom simulates real life situations in which people are required to make decisions based on ambiguous social cues and subtle social information presented in a short film.  We also examined the psychometric properties of this measurement tool.

Methods: Participants viewed films based on Happé’s (1994) Strange Story tasks and were required to provide a written response to the question associated with each script within one minute.  Participants were required to interpret the events in the films as they unfolded, just as they would in real life (cf. having an opportunity to read, reread, and think about the story as can happen with a pencil-and-paper scenario), thereby limiting the opportunity for “hacking out” correct responses.  We examined internal consistency, and test-retest reliability after a 2 week interval.  Concurrent validation was investigated by correlating A-ToM performance with performance on Happé’s Strange Stories Test, and the Reading the Minds in Films (Golan, Baron-Cohen, Hill & Golan, 2006) for a sample of (i) non-ASD adults (N = 100) and (ii) individuals with an ASD (N= 50).

Results: The final version included only those items characterised by reasonable internal consistency and test-retest stability.  The validity of the instrument was suggested by the following pattern of findings: (i) the correlation between A-ToM performance and performance on Happé’s Strange Stories and the Reading the Minds in Films tests.  (ii) the different performance levels for the A-ToM compared with existing pencil-and-paper ToM scales.  (iii) performance differences between ASD and non-ASD participants on the social versus physical items.

Conclusions: This study introduces a psychometrically sound tool for assessing ToM among adults with ASD.  The tool involves a naturalistic setting and demands real-time responsiveness, reducing the capacity for participants to “hack out” solutions.