Stage 2 registered report: investigating a preference for certainty in conversation among autistic adults

Social Anxiety in Autistic People: Does the Clark and Wells Model fit?

PURPOSE

Cognitive behaviour therapy based on the Clark and Wells (1995) model is a first-line treatment for neurotypical people seeking support for social anxiety. While autistic people frequently report high social anxiety, it is unclear how appropriate the model is for this population.

METHODS

Over 300 autistic and non-autistic adults completed an online survey measuring key variables of the Clark and Wells model (socially-related negative thoughts, safety behaviours, self-focused attention). Using multiple regression and structural equation modelling, we assessed whether these variables accounted for the link between autism and social fears.

RESULTS

In multiple regression, autistic people experienced greater social fears than expected based on Clark and Wells variables, and safety behaviours were less predictive of social fears in autistic people. In structural equation modelling, Clark and Wells variables only mediated half the link between autistic traits and social fears. In exploratory analysis, we found that distress relating to uncertainty was an additional variable that needed to be taken into consideration in the relationship between autistic traits and social fears.

CONCLUSION

The Clark and Wells variables were relevant in autism, but did not fully explain elevated social fears in autistic people, which suggests that other factors are also important in accounting for social anxiety in autistic people. This means that therapy informed by the model may not be optimal for autistic people. We recommend further research developing adapted therapy for social anxiety in autistic people.

Development and validation of the conversation questionnaire: A psychometric measure of communication challenges generated from the self-reports of autistic people

Existing measures of communication challenges in autism are based on diagnostic criteria and research/clinical observations of autistic people, rather than what autistic people themselves identify as difficulties. In this study, the Conversation Questionnaire (CQ) was developed based on community engagement with autistic people to identify what they find challenging about conversation. This new tool was then administered online to autistic, dyslexic and neurotypical people (N = 312) in a validation phase of the study. Item-response theory modelling indicated that a two-dimensional structure accounted for response patterns. These dimensions reflected difficulties knowing what to say (15 items) and engaging in behaviours possibly disruptive to neurotypical conversation (21 items). The dimensions showed good internal consistency and convergent and discriminant validity, and could distinguish between autistic and neurotypical people (d = 1.59 and d = 2.07 respectively). The CQ might help contribute to diagnostic assessment for autism in adults as part of a holistic assessment. The questionnaire might also be useful with other neurodiverse groups, and provide a tool for clinicians and researchers to identify individuals' strengths and difficulties in conversation (e.g., as part of interventions in speech and language therapy).