The Observation of Spontaneous Expressive Language: A New Measure for Spontaneous and Expressive Language of Children with Autism Spectrum Disorders and Communication Disorders

Background: An assessment of language level and delay is one of the crucial parts of identifying and describing Autism Spectrum Disorders (ASD).   Language impairments (e.g. echolalia, pronoun reversal, and odd intonation) have been found to be one of the strongest indicators of ASD.  For the assessment of spoken language in ASD, transcriptions or formal linguistic analyses (e.g. CHILDES, SALT) can be used, which can be time consuming and highly technical.  However, there have not been instruments that measure spontaneous expressive language (defined as language not directly elicited as part of the test) of children with ASD in a more naturalistic setting.

Objectives: The main objective of the present study is to evaluate a newly developed assessment tool, the Observation of Spontaneous Expressive Language (OSEL) that measures children’s social use of spontaneous expressive language in natural contexts. 

Methods: Data were obtained from 160 typically developing children between 2 to 5 years who were administered a set of language measures, such as Preschool Language Scale (PLS) and the OSEL.  During the OSEL, children’s use of spontaneous expressive language in natural social contexts (e.g. playing with a variety of toys, interacting with an examiner during imaginative play, telling stories from a cartoon strip) were coded.  Caregivers were administered a set of parent questionnaires, such as the Vineland Adaptive Behavior Scales (VABS).  The first set of analyses focused on inter-rater and test-retest reliabilities using weighted kappa.  Internal consistency was assessed using Cronbach’s alpha.  Concurrent and convergent validity were also assessed through calculating correlations between the OSEL scores and other measures, as well as age and verbal/nonverbal IQ scores.  Another focus of analyses was to derive age equivalents based on item means and standard deviations 

Results: Inter-rater and test-retest item reliabilities ranged from kappa of 0.58 to 0.77 and from 0.49 to 0.78 respectively. Cronbach’s alpha for internal consistency was 0.94.  Most OSEL items showed moderate associations with age (65% showing r >0.3).  Correlations between the OSEL items and expressive communication scores in the PLS and communication domain scores in the VABS were moderate (for both, more than 50% showing r >0.3). Correlations between the OSEL items and verbal and nonverbal IQ scores were generally low (90% showing r <0.3 for verbal IQ, 85% showing r <0.3 for nonverbal IQ).  As expected, item means for grammatical uses for older cohorts were higher than for younger cohorts.  In contrast, the opposite trend emerged for ungrammatical uses. 

Conclusions: Results indicate moderate to strong inter-rater and test-retest reliabilities and strong internal consistency for the OSEL.  Generally, age showed moderate associations with the OSEL items scores.  The OSEL scores were also moderately associated with other language measures including the PLS and VABS.  Verbal and nonverbal IQ scores were fairly independent of the OSEL scores.  Age equivalents are in the process of being derived.  We expect that the further assessment of the OSEL in children with ASD and other communication disorders can address the potential utility of quantified language profiles obtained from the OSEL in intervention, genetic and neuroimaging research.