This study investigates how a natural experiment occurring in the teaching of principles of microeconomics allows a test of the Dunn and Dunn learning styles model (Dunn & Griggs, 2000). The material for the first exam, based on essential definitions and theoretical foundations, was taught in a conventional, inductive style, more compatible with analytic learners. The second and third examinations were based on applications of those foundations and rely on a more deductive teaching method, more with compatible global learners. This dichotomy provides an opportunity to measure the relative productivity of two alternative teaching methods to student achievement while accounting for variables representing students' different learning style preferences. Separate OLS regressions were estimated to measure the influence of student learning style preference variables for the segments of the course utilizing presumptively inductive and deductive teaching methods. Although there was little evidence that students with analytic learning style preferences performed better in the more inductive segment of the course, there was persuasive evidence that students with global learning style preferences performed better in the deductive segment of the course. The authors consider how these results reflect on validity of the Dunn and Dunn learning style model and possible adjustments in teaching strategies. (Contains 4 tables.)