Previous research has indicated that laypeople, students and legal professionals often hold flawed beliefs about memory, and the present study sought to extend this area of research to the teaching profession. Are teachers’ beliefs about learning in line with the scientific consensus? A set of vignettes with contrasting options for classroom practice were presented to trainee (n = 77) and in-service (n = 44) teachers, and in each case a 7-point Likert scale prompted them to predict which would be the best course of action in terms of student outcomes. As hypothesized, responses were often out of line with research on ‘desirable difficulties’ in memory and learning such as retrieval practice, spacing, and interleaving, with choices indicating a lack of awareness of these evidence-based approaches, although they were more accurate than previous studies of students. Surprisingly, accuracy of response did not correlate with the duration of a teacher’s classroom experience; trainee teachers outscored in-service teachers in certain areas, suggesting that recent familiarity with technical literature on learning could be advantageous.