Purpose: To address whether increasingly authentic instructional formats are more effective in improving preclerkship medical students' performance. Method: From 2007 to 2009, the authors conducted a prospective, randomized, crossover study with second-year medical students in a clinical reasoning course at the Uniformed Services University of the Health Sciences. The authors randomly assigned students to one of three cohorts and used instructional formats of differing authenticity (paper, DVD, standardized patient) to teach three subject areas (abdominal pain, anemia, polyuria). Each cohort received one instructional format for each subject area. The authors collected outcome measures (objective structured clinical exam, video quiz, and essay exam scores) at the end of each academic year. They stratified the students into tertiles by first-year grade point average to investigate the impact of instructional formats on learners of different abilities. Results: Outcomes for students in the top tertile improved with increased authenticity of the instructional format compared with outcomes for students in the middle and bottom tertiles (0.188 versus -0.038 and -0.201, P = .001 and .027, respectively). However, outcomes for students in the bottom tertile decreased when students were given only the paper case, compared with the middle and top tertiles (-0.374 versus 0.043 and 0.023, respectively, P = .001), but subsequently improved with more authentic instructional formats. Conclusions: The authors could not demonstrate that increased authenticity of the instructional format resulted in improved learner performance. However, they believe that there may be some benefit to tailoring preclerkship clinical education based on students' ability.