We explored possible cognitive, behavioral, emotional, and physiological risk markers for sleep disturbance in children with autism spectrum disorders. Data from 1,583 children in the Autism Treatment Network were analyzed. Approximately 45 potential predictors were analyzed using hierarchical regression modeling. As medication could confound findings, it was included in the analyses as a covariate. Results revealed that anxiety, autism symptom severity, sensory sensitivities, and GI problems were associated with sleep disturbance. IQ positively predicted sleep disturbance, and children with Asperger's Disorder were more vulnerable than others. The amount of variance in sleep outcomes explained by predictor variables was modest (i.e., R[superscript 2] from 0.104 to 0.201). Predictor variables were evaluated in the context of a bidirectional theoretical framework.