Background: School-based interventions can result in profound social, behavioral, and academic improvement for students with autism, but teachers rarely implement them. It is important to understand why this occurs and use this information to increase the use of evidence-based practices. Toward this goal, 2 proof-of-construct studies demonstrate the theoretical and methodological advantages of measuring behavioral intentions to use specific practices. Methods: Two observational studies enrolled public school teachers who work with students with autism. The studies measure the strength of teachers' intentions to use each of 4 different evidence-based practices, assess variability in intentions, and test whether intentions predict future teacher behavior. Results: In adjusted analyses, intentions to use a specific evidence-based practice were strongly associated with subsequent use (adjusted odds ratio = 5.2). The proportion of teachers who reported strong intentions varied from a low of 33% to a high of 66%, depending on the practice. Conclusions: The findings illustrate that the strength of intentions, which varies depending on the specific practice, can predict implementation. More generally, the studies demonstrate how measures of intention can aid efforts to identify implementation barriers. The approach taken can be applied to study implementation of any practices designed to improve student health.