mentalising (although the correlation with the Eyes task approaches significance). This
lack of association is particularly interesting when set alongside the finding that the autism
group and the control group differed significantly on coherence and reflective function
scores. Two interpretations are possible. It may be that autistic symptomatology and
mentalising ability are continuously correlated across all people with and without autism,
but the measures used are better suited to picking up gross between-group differences
rather than subtle within-group variations; this is discussed in more detail below. This is
consistent with the observation that all of the correlations between security measures
(reflective function, coherence of mind and coherence of transcript) and participant
characteristics (ADOS scores, IQ and mentalising) were in the predicted direction but few
were significant or even approached significance. The alternative conclusion is that there
may be gross group differences between people with autism (low reflective functioning
and mentalising, high autistic symptomatology) and people without autism, rather than
each factor existing on a continuum. This latter idea does not seem to be consistent with
existing findings about attachment and mentalising in typical development, but
circumstances may be different for people with autism, particularly given the suggestion
that they may form secure attachments in a different manner from typically-developing
children and adults. Although attachment and mentalising have been explored together in
typical samples, attachment-based mentalising measures like reflective function have not
been compared to more traditional theory of mind measures. Overall, however, the general
lack of association between attachment measures and these other variables
(symptomatology, theory of mind and IQ) should make us more confident that the AAI
was genuinely tapping into attachment representations in the participants, rather than
acting as a proxy for other abilities or deficits.
Limitations
22