Early Childhood Special Education teachers select practices to use to promote social communication development for their students with autism spectrum disorder. Understanding what evidence-based practices teachers select and why can inform the development of dissemination and implementation supports at the critical Adoption-Decision stage of implementation. The researchers used discrete choice analysis to examine and test the effects of Early Childhood Special Education teachers’ beliefs (self-efficacy, attitude, subjective norms; Theory of Planned Behavior) about key evidence-based practices on their practice selections. To assess malleable determinants of practice selection, this study addressed these aims: (1) assess Early Childhood Special Education teachers’ beliefs (attitudes, subjective norms, and self-efficacy) about key evidence-based practices, (2) describe which evidence-based practices teachers selected, and (3) examine whether teachers’ beliefs predicted practice selections. A web-based survey gathered responses from 222 Early Childhood Special Education teachers. Participants had the most favorable beliefs about naturalistic intervention and the least about discrete trial teaching and scripting. These beliefs aligned with the frequencies with which these practices were selected, and teachers’ beliefs predicted which practices they selected. Together, these findings suggest that beliefs serve as determinants of Early Childhood Special Education teachers’ practice selections that can be acted on to tailor pre-implementation supports and improve implementation from the start. Lay abstract Preschool special educators’ are more likely to choose an educational practice to teach a young child with autism a social communication skill if they have positive beliefs about it. We asked preschool special educators to read a description of an autistic student and their social communication goal and imagine they were the student’s teacher. We then asked them to pick one of five practices to teach the student. We also asked them questions to understand their attitudes about, confidence in their ability to use, and their perception of their coworkers’ support of each practice. There are many research-based practices that a teacher could use to help children learn, and preschool teachers often make these decisions for their students. Teachers’ beliefs varied in how supportive they were of each practice, and research shows people are more likely to do something that their beliefs support. In this study, they had more supportive beliefs and were more likely to use some practices, like naturalistic intervention, than other practices, like discrete trial teaching. By knowing this, researchers can help teachers use practices that their beliefs support and help change teachers’ beliefs to be supportive of a practice they may need to use.