Assessment of Pragmatic Language in High-Functioning Autism

Background:

Although individuals with high-functioning autism (HFA) often have average syntactic and lexical skills, deficits in pragmatic language are a hallmark of the disorder (Tager-Flusberg et al., 1990). Difficulties with pragmatic language, including reduced turn-taking, reduced range of speech acts, and inadequate management of turns and topics in conversation have long been reported in the literature. Although such pragmatic difficulties are well documented, there are few assessment instruments designed to characterize these pragmatic language acts.  One such instrument developed to quantify pragmatic language deficits is the Yale in vivo Pragmatic Protocol (YiPP; Schoen, Paul & Volkmar, in prep). This semi-structured conversational tool was designed to elicit pragmatic behaviors in children with HFA.

Objectives:

To examine YiPP performance by children who have HFA and compare their performance to a typically developing (TD) control group.

Methods:

Children and adolescents age 9 - 17 years were divided into four groups based on diagnosis (HFA vs. typically developing; TD) and age (younger (Y) = children 9-12, and older (O)  = children 13-17). Groups (HFA-Y vs. TD-Y and HFA-O vs. TD-O) were matched on chronological age and IQ.  Each participant completed a 20-minute YiPP conversation that was video recorded. Each YiPP item was given an error code based on the appropriateness of the child’s response. Error codes for each behavior were divided into five domains of pragmatic language (discourse management, DM; communicative function, CF; conversational repair, CR; presupposition, P; register variation, RV). The number of utterances produced by each participant (#child) during the YiPP was tallied to capture amount of language each participant used.

Results:

ANOVA with LSD post hoc testing was used to examine differences in error scores among the four groups. Differences between groups were observed within the DM, CR, P and RV pragmatic domains. No differences were noted on the CF domain.

An independent samples t-test yielded a significant difference between the HFA and TD groups on overall number of participant utterances (t(136)= -2.403, p < 0.02). Pearson correlation coefficients were calculated for the relationship between number of utterances and pragmatic error scores for participants with HFA. A negative correlation was found between # utterances and DM (r= -.420, p < 0.01); #child and CF (r= -.349, p < 0.01); and #child and P (r= -.342; p < 0.01).

Conclusions:  

Children with HFA produced responses during the YiPP that were less pragmatically correct relative to age-matched TD peers. Differences were primarily observed in discourse management, conversational repair and presupposition domains. Additionally, children with HFA produced fewer utterances than did their TD peers. Among HFA participants, a greater number of child utterances was associated with lower error scores in DM, CF, and P domains. These results confirm pragmatic language difficulties and lower speaking rates among children with HFA relative to TD peers. This pattern of results illustrates the difficulties that children with HFA have in maintaining conversation and responding to a partner’s conversational bids and has implications for the development of social-communicative interventions.