Individual Submission Summary
Share...

Direct link:

Poster #50 - The Influence of Conversational Partner on Expressive Language in Children and Adolescents with ASD

Thu, March 21, 4:00 to 5:15pm, Baltimore Convention Center, Floor: Level 1, Exhibit Hall B

Integrative Statement

Expressive language (EL) is one of the most heterogeneous characteristics of ASD (Kjelgaard & Tager-Flusberg, 2001). The collection of natural language samples is an excellent way to examine it in a more naturalistic way (Tager-Flusberg et al., 2009) and is especially useful for the understudied end of the spectrum of minimally and low verbal individuals making up about 30% of the population (Kim et al., 2014).
Although numerous studies have used language samples to examine EL in autism, few have evaluated the role of sampling context. The social demand-related aspects of sampling context might determine whether the elicited language is representative of actual ability, particularly for children with ASD. Only one study compared the role of conversational partner (parent vs. examiner) on EL in ASD, but it focused solely on 4-year-olds (Kover et al., 2014).
We aimed to examine the effects of conversational partner (parent and examiner) and setting (home and lab) on the EL of children and adolescents with autism who varied in age and language ability, but with the majority classified as minimally verbal.
Methods:
27 (23 females) children and adolescents with ASD (Mean age=11;6) were included in this study. Standardized measures of language, communication, and nonverbal IQ were used: VABS, SCQ, and Leiter, and diagnosis was confirmed with the ADOS (Table1).
The Eliciting Language Sample for Analysis (ELSA) protocol, developed by our group for use with individuals with a wide range of age and language ability, was used to collect EL samples. ELSAs were collected from children and adolescents with ASD (N=27) by their parents in their home and by an examiner in the lab. Instructions for nine different activities and a standard set of props were used for the samples. Fidelity of administration of the activities was coded. The ELSAs were transcribed following standard procedures in SALT. Frequency of speech utterances and conversational turns per minute, mean length of utterance in words, and number of different words produced by the target child were coded.
Results:
Parent-elicited (M=27.83min;SD=12.58) and examiner-elicited samples (M=22.88;SD=5.18) did not differ in duration; however, examiners’ administration fidelity (activities attempted out of 9; M=8.30;SD=1.07) was significantly higher than parents’ (M=6.38;SD=2.78; Z=-3.11,p<.05).
For the EL measures, frequency of speech utterances per minute extracted from the parent-elicited sample (M=7.66;SD=5.41) was significantly higher than from the examiner-elicited sample (M=4.49;SD=3.83; Z=-3.87,p<.01). The same pattern was observed for conversational turns and number of different words per minute, but not for mean length of utterance (Figure1).
Conclusions:
Although parents were not as successful as examiners at administering the ELSA activities, they collected samples of comparable duration. The EL they elicited from their children and adolescents was qualitatively and quantitatively different from that elicited by examiners. Participants produced utterances and took turns more often when with their parents, and their EL was characterized by a more diverse lexicon. This is noteworthy considering that the majority of participants (21/27) were classified as minimally verbal. These findings demonstrate the importance of context, conversational partner, in particular, in the evaluation of EL in autism.

Authors