The effects of decision framing on choice of risky vs certain options

Abstract

Tversky and Kahneman (1981, Science, 211, 453–458) reported that the choice between two decision alternatives of equal expected value, differing only in the degree of risk, was significantly influenced by the "framing" of the alternative decision outcomes. When alternative outcomes were phrased "positively" in terms of lives saved, subjects preferred the risk-averse alternative. When outcomes were phrased "negatively" in terms of lives lost, the risk-seeking option was preferred. Experiment 1 investigated the robustness of the framing effect and its remediation by formal training. Forty-five MBA students responded to a decision problem before and after training in decision theory. Results differed from those published by Tversky and Kahneman (1981) in two respects. First, although results in the positive framing condition were similar to Tversky and Kahneman's results, no preference for the risky option in the negative framing condition was observed. Consequently, no framing effect was found even on the pretest. Second, many students reported being indifferent to the two options. No significant changes occurred after training. Experiment 2 replicated these results on a nonbusiness sample. These findings suggest that (a) unidentified factors affect whether a framing effect is observed, and (b) the bias known as the framing effect may not be as robust as has been believed.