Three studies replicated a classroom experiment on single-exposure musical conditioning of consumer choice (Gorn, 1982), testing whether simultaneous exposure to liked (vs. disliked) music and a pen image induced preferences for the shown (vs. a different) pen. Experiments 1 and 2 employed the original music, Experiment 3 used contemporary music. Experiments 2 and 3 employed hypothesis-blind experimenters. All studies incorporated post-experimental inquiries exploring demand artifacts. Experiments 1 and 2 (original music; N = 158, N = 190) showed no evidence for musical conditioning, and were qualified (conclusive) replication failures. Experiment 3 (contemporary music; N = 91) reproduced original effects, but with significantly smaller effect size. Moreover, it had limited power and showed extreme scores in one experimental group. Aggregated, the three studies produced a null effect. Exploration of demand artifacts suggests they are unlikely to have produced the original results. © 2014.