Can autistic traits present in the normal population determine social communication abilities? An investigation into correlations between AQ score and performance in communication tasks testing audience design and lexical alignment.
Rosalind Baring's dissertation 2009.pdf.doc (727.5Kb)
Baring, Rosalind M B
MetadataShow full item record
This study explored the concept of the Autistic Spectrum (Wing (1988)), and the theory that autistic traits of varying severity can be identified and measured in the general population using Baron-Cohen, Wheelwright, Skinner, Martin and Clubley’s (2001) Autism-Spectrum Quotient. It was hypothesised that variation in the levels of autistic traits present in the normal population might determine social communication abilities. This was tested using two referential communication tasks. The first task was designed to assess participants’ aptitude for audience design by manipulating shared versus privileged information. The second task was a picture naming exercise designed to assess participants’ tendencies to lexically align with interlocutors. The results showed that despite no significant result being found between audience design and AQ in Task 1, Task 2 showed that those scoring higher on the AQ were significantly more likely to lexically align with their interlocutor that those scoring lower (Pearson’s r=-0.536, p<.05). The lack of significant results in task 1 was attributed to methodological issues- and the concept not dismissed. The significant results found in Task 2 were proposed to be attributed to the impairment in Theory of Mind found in autistics also being responsible for the variation in communication abilities in the general population. Such a deficit would impair a speaker’s ability to put themselves ‘in the shoes of their interlocutor’ in order to facilitate communication, and also have adverse effects on an individual’s awareness of social politeness implications of lexical dis-alignment.