Evidence-based practice for homeschooled children with ASD
Simmons, Christina Anne
MetadataShow full item record
A sample of 114 parents of children with autism spectrum disorder (61 homeschool; 53 traditional school) from across the U.S. completed an online survey on their educational experiences. A multivariate analysis of covariance (MANCOVA), controlling for gender and diagnosis, indicated a statistically significant main effect between groups on educational experiences, Wilks’s Ʌ = .40, F(5, 65) = 18.84, p < .001, η2p = .60. Educational group had a statistically significant effect on satisfaction with educational placement, interactions with education professionals, child’s social-emotional functioning, child’s safety, and stress placed on the family. Overall, homeschool programming was described as having significantly fewer aspects of evidence-based practice than traditional school programming with minimal differences between groups in the extracurricular/social activities provided. Study 2 evaluated the effect of a self-management intervention implemented by three parents of children with autism spectrum disorder in the home environment, including participant self-monitoring and self-evaluation, on on-task behavior during independent work completion. Results indicate that parents implemented intervention procedures with high fidelity (M = 98.92%), intervention led to increased on-task behavior across participants (NAP = 1.0 for all participants, M increase from baseline = 60.67%; Cohen’s d = 3.46), results maintained post-intervention, parents and children rated procedures as high in social validity, and observation via live video technology resulted in high correspondence between parent and child ratings (M = 90.03%). This study extends the use of live video technology for parents to collect data during intervention. Visual analysis methods incorporated masked visual analysis to control for Type I error and to increase the internal validity and scientific credibility of this single-case design study. The blind data analyst correctly identified baseline and treatment segments for each participant, thus results represented a significant intervention effect on on-task behavior (p = .03). Masked visual analysis methods control for Type I error in single-case research and provide a summary statistic that is easily interpreted across fields.