Interest, topic and attention: a transdiagnostic approach to child conversation

Once children reach school age, social conversation becomes the key means of maintaining peer relations. However, our understanding of the underpinnings of child conversational difficulties has often focussed on diagnostic categories and group averages, obscuring substantial within-group variability. In this talk, I argue for a shift from categorical comparisons to mechanistic, dimensional explanations of conversational differences.

First, I present two pre-registered experimental studies (N = 204, half autistic) testing whether intense interests causally shape conversational topic management in 7–11-year-olds. Using a novel conversation paradigm, children were exposed to visual distractors via screenshare. In Study 1 we manipulated whether the visual distractors corresponded to children’s high-interest domains or were of generic interest. In Study 2 we manipulated whether the conversation topic was of high versus generic interest.

Across both these studies we found that high-interest content systematically altered topic management. In Study 1 autistic and neurotypical children were less likely to maintain topic in the presence of high-interest distractors. In Study 2 they were less likely to shift topic (here: when a shift was pragmatically warranted e.g. to comment on a warning) during high versus generic interest conversation topics. Regression analyses conflating across diagnosis found that parent-rated interest intensity accounted for substantial and significant unique variance across participants on these tasks, even when controlling for autistic traits, cognitive flexibility, and core language ability.

Second, parent-report data from 109 autistic children revealed two distinct but strongly correlated conversational styles: verbosity and reticence. ADHD traits and interest-related preoccupations were associated with both conversational styles. However, core language difficulties specifically predicted reticence.

Finally, I discuss implications for practice through a feasibility evaluation of the Conversation for Social Interaction - a new universal classroom programme targeting conversational skills in 7–11-year-olds. Teachers reported good acceptability and emphasised the growing need for structured support for social communication in primary classrooms. Pre versus post-programme comparisons for researcher-to-child assessments indicate improvements in relevant responding, conversational turn-taking, and reduced reticence (although there was no control group for this pilot). Children’s self-reports indicated high acceptability, with 87% reporting that CoSI had helped them with real-life conversations, often supported by concrete examples.

Together, these findings show that a needs-based, transdiagnostic approach can advance explanatory models of child conversation while offering a practical solution for the diverse support needs observed in mainstream