Contextual dependence has been hypothesized to influence classroom test performance such that taking a test away from the lecture room should lead to lower test scores (Abernethy, 1940). We studied the performances of students who took typical college tests in rooms different from the lecture rooms and made comparisons to classmates who remained in the lecture rooms. No statistically reliable effects were found in 21 such comparisons in seven courses. Although contextual dependence has been produced under laboratory control, college classes induce students to decontextualize information. The theoretical utility of contextual associations is based on simpler, more tightly controlled conditions, and generalization to representative situations is an empirical matter.