|
Measuring the Effectiveness of Behavioral Interventions for Children with Autism |
Tuesday, May 30, 2006 |
9:00 AM–10:20 AM |
Chicago A-F |
Area: AUT; Domain: Theory |
Chair: Gerald E. Harris (Texas Young Autism Project) |
Discussant: Gerald E. Harris (Texas Young Autism Project) |
CE Instructor: Gerald E. Harris, Ph.D. |
Abstract: Making behavioral treatment available to more children with autism entails convincing others, including health service providers, parents, insurance companies and public policy makers, that this treatment is effective. Unfortunately, all too often behaviorists are relying on assessment instruments with unknown or unproven reliability and validity for this special population. The 3 presentations in this symposium present data that significantly advances the psychometric knowledge for some of the most widely used measures of intelligence, language and behavior problems in the autistic population. Data was collected from comprehensive assessments of over 100 children diagnosed with autistic disorder as they participated in behavioral treatment programs. Sample sizes for the data analytic procedures are thus larger than usually seen in this area. The first presentation examines the most popular comprehensive intelligence test, the WPPSI-III, in terms of its applicability for children with autism. The second presentation looks at construct validity for several widely used language measures. What do they really measure, and what is the overlap in what they do measure? The third presentation investigates the utility of an efficient behavior report instrument for this special population. Together, these presentations provide insight into our ability to demonstrate the effectiveness of behavioral interventions. |
|
An Examination of the Use of the WPPSI-III Intelligence Test with Children with Autism. |
WENDY J. NEELY (Texas Young Autism Project), Allison Serra Tetreault (Texas Young Autism Project), Ehsan Bayat (Texas Young Autism Project), Gerald E. Harris (Texas Young Autism Project) |
Abstract: Assessment of cognitive abilities of children with autism is crucial to designing and evaluating behavioral interventions. Little is known about the psychometrics of the most widely used intelligence test, the Wechsler Preschool and Primary Scale of Intelligence -III, for this population. Wechsler published a study in the WPPSI-III Technical Manual (The Psychological Corporation, 2002) addressing the utility of the WPPSI-III for this special population. However, several significant methodological problems are noted in that study, including a very small sample (n = 21), and restrictions of age and I.Q., as well as unknown test administration and scoring procedures for the data provided by an independent third party examiner. In the present study, data from standard initial administrations of the WPPSI-III to a much larger group of children (n = 91) diagnosed with autism was analyzed and the results compared to the findings from the Wechsler study. Significant differences were found in means and distributions of subtest and composite area standard scores. Scores for lower functioning (I.Q. < 60) children with autism, in particular, were very different. Full results and implications for the use of the WPPSI-III for children with autism in clinical and research settings are discussed. |
|
Language Skills of Children with ASD: Construct Validity of Commonly Used Language Tests. |
CATRIONA BORG-HANSEN (Texas Young Autism Project), Gerald E. Harris (Texas Young Autism Project) |
Abstract: Language skills of children with an Autism Spectrum Disorder (ASD) are a central focus for both clinicians and researchers. When diagnostically assessing or when testing for effectiveness of interventions or treatments, common practice is to use only one of several tests available, assuming that the test is indicative of the participants’ overall language skills. The current study looked at construct validity of the most frequently used language measures: Reynell Developmental Language Scales, Preschool Language Scale- IV, Vineland Adaptive Behavior Scales, and General Language Index-WPPSI-III. One-hundred-one children diagnosed with ASD between the ages of 15 months and 13 years were assessed with multiple language measures as well as an intelligence measure. Participants’ mental age ranged from 10-53 months and I.Q. from 15-100. Partial correlations across all tests controlled for age equivalent scores on the intelligence test. Surprisingly, correlations across tests were generally low or nonsignificant. The overlap of variance accounted for beyond general intelligence ranged from r2 =.07-.69, indicating that the tests are measuring different constructs 31%-93% of the time. According to the results of this study, researchers, diagnosticians, and treatment providers need to use several assessments to accurately measure language skills in children with ASD. |
|
The Utility of the CBCL as a Screening Tool in Identifying Children with Autism. |
ALLISON SERRA TETREAULT (Texas Young Autism Project), Lauren Harrington (Texas Young Autism Project), Gerald E. Harris (Texas Young Autism Project) |
Abstract: As a behavior oriented measure, the Child Behavior Checklist (CBCL) has potential for increasing behavioral awareness and focus in clinical activities associated with autism. This investigation examines the CBCL to determine if the measure is useful as a screening tool in the identification of children with autism. The CBCL has been widely used to assess patterns of behavior problems in children based on parental, caregiver or teacher report. Problem behaviors are clustered into seven syndromes and five DSM-Oriented Scales, which group items based on their relationship to criteria for DSM diagnoses such as Autistic Disorder and Asperger’s Syndrome. Although the items in the DSM-Oriented Scales are related to criteria for clinical diagnoses, these scales do not correspond precisely to diagnostic criteria. The current study investigated the validity of the Pervasive Developmental Problems (PDP) DSM-Oriented Scale in 120 children previously diagnosed with autism. Results reveal a pattern of very elevated t-scores on the PDP scale in addition to clinical elevations on the Withdrawn Syndrome and the Attention Deficit Disorder Syndrome. A proposed CBCL profile for autism is presented and implications for clinical practice and future research are discussed. |
|
|