In order to implement more science, technology, engineering, and mathematics (STEM) into K12 classrooms, it is important to find out whether teachers are comfortable with this pedagogy. To determine teachers' current self-efficacy of STEM pedagogy, teachers in a southern state in the United States were asked to enlighten researchers into this phenomenon. Participants were K12 teachers (n = 119) from a public school district undertaking a district-wide STEM initiative. A measure of STEM teacher self-efficacy and a demographic questionnaire were administered online to participants. STEM teacher self-efficacy data were analyzed, along with demographic data, using descriptive discriminant analysis (DDA) and canonical correlation analysis (CCA). Results suggest some demographic variables are more predictive of STEM self-efficacy (gender, grade level taught, feelings of administrative support, and professional development sessions attended) than others (whether or not gifted courses are taught, age, and length of teaching experience. This data should be used by school administrators that seek to begin or improve STEM pedagogy in their schools.