In light of the steady rise in the prevalence of students with autism, this study examined the definition of autism published by state education agencies (SEAs), as well as SEA-indicated evaluation procedures for determining student qualification for autism. We compared components of each SEA definition to aspects of autism from two authoritative sources: Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR) and Individuals with Disabilities Education Improvement Act (IDEA-2004). We also compared SEA-indicated evaluation procedures across SEAs to evaluation procedures noted in IDEA-2004. Results indicated that many more SEA definitions incorporate IDEA-2004 features than DSM-IV-TR features. However, despite similar foundations, SEA definitions of autism displayed considerable variability. Evaluation procedures were found to vary even more across SEAs. Moreover, within any particular SEA there often was little concordance between the definition (what autism is) and evaluation procedures (how autism is recognized). Recommendations for state and federal policy changes are discussed.