Select Page

Source URL: Click here to view this document

Intro: This study investigated the effectiveness of expository discourse language sample analysis in differentiating between typically developing and language impaired adolescents.

Background: Language sampling has historically been a standard of speech-language assessment.  However, many language elicitation tasks were conversational in nature which tends not to produce as much complex language from adolescents. While most morphology is acquired by early elementary school, previous research (Nippold et al, 2005) indicated that the use of compound and complex sentences continued to develop through adolescence into adulthood, highlighting the need for a language elicitation task appropriate to adolescents and adults. 

Hypothesis: The authors investigated whether language sampling analysis of conversational versus expository discourse would discriminate adolescents that were typically developing (TLD) versus those with specific language impairment (SLI). Previous research indicated that measures of language complexity sensitive to growth into adulthood included mean length of t-units (one main clause and a subordinate clause), subordination of clauses and clausal density. 

Methodology and Participants: For the study, Nippold et. al. used a sample of 444 eighth graders attending American public schools who had participated in a large study of language impairment as kindergartners. As a result, they had already been classified in kindergarten as TLD, SLI or non-specific language impairment (NLI) at that time. The examiners tested the students using subtests of the CELF-3 as a test of syntax and subtests of the WISC third edition as a measure of general intelligence. For the language elicitation task, students were engaged first in conversation and then were asked to describe how a favorite game or sport worked. The authors then analyzed the language samples for mean length of T-unit (one main and one subordinate clause), use of subordinate clauses and clausal density (average number of clauses per t-unit). 

Conclusion: Results of the study revealed that while the language produced for the conversational task did not differentiate between the groups, the measure of Mean Length T-unit from the expository task did. The authors speculated that by adolescence, conversation may not stress the linguistic system enough to demonstrate differences in language ability. Conversational tasks may not elicit language complex enough in order to differentiate a typically developing adolescent from one with a language disorder. 

Relevance to the field: This study further emphasizes the effectiveness of language sampling analysis in assessment but also underlines the need to elicit high level language. Expository tasks also will produce natural language that is more similar to the type of language skills adolescents need for school and work. It is important to recognize the need for this type of language sampling in understanding how language impairment affects young people in their day to day lives in order to plan effective intervention. 


Nippold, M. A., Mansfield, T. C., Billow, J. L., & Tomblin, J. B. (2008). Expository discourse in   adolescents with language impairments: Examining syntactic development. American Journal   of Speech-Language Pathology. 17(4):356-366.