Framing effects on retail store choice decisions were investigated in four experiments. Subjects preferred the store that guaranteed (a sure option) good prices (experiment 1), product availability (experiment 2), or a rebate (experiment 3) when consequences were framed in terms of gains; subjects preferred the risky option when consequences were framed in terms of losses. Consistent with fuzzy-trace theory, framing effects were reduced when the expected values of options were disparate in a direction that disfavored sure gain or probabilistic loss options (experiment 2) and when the perceived costs of committing a judgmental error were high (experiment 3). Experiment 4 shows that the moderating effects of disparate expected values and costs of judgment errors generalize to within-subject designs.