Behavior analysts typically teach conditional discriminations in a mixed-trial format but may switch to a blocked-trial format for learners displaying limited acquisition. No known research has shown that mixed-trial methods are more effective or efficient than blocked-trial methods for teaching discriminations, so it is not clear why this format has been adopted as the "first-line" intervention. We compared blocked and mixed-trial formats for teaching novel auditory-visual discriminations to three adults with intellectual and developmental disabilities (IDD). Results show blocked-trial methods resulted in faster skill acquisition in all cases, suggesting this format may be a preferable starting point for instruction.