Gain/Loss Framing Effects

Risky and Riskless Contexts

Tverksy and Kahneman describe this aspect of their work as research on the cognitive and psychophysical determinants in risky and riskless contexts. For these authors, framing refers to the cognitive point at which decision problems can be described (framed) in multiple ways, giving rise to different preferences being elicited that are dependent on the frame.

They further argue that framing effects can help explain some of the anomalies found in consumer behavior. Other researchers have extended their point to medical decision making in that caution needs to be used in deciding how decision problems are presented to patients.

Early Research in Framing

Attention to the use of data in decision making was brought into the medical-decision-making arena in a scientific article by Barbara J. McNeil, R. Weichselbaum, and S. G. Pauker appearing in the New England Journal of Medicine in 1978 on the fallacy of 5-year survival in lung cancer. McNeil and colleagues focused attention on the 5-year survival data in lung cancer. This article focused attention on the importance of choosing therapies not only on the basis of objective measures of survival but also on the basis of patient attitudes. However, while McNeil and colleagues derived their data from existing data on 5-year survival from the published medical literature, they did not present graphical displays of 5-year survival curves to study participants. Rather, McNeil and colleagues presented data derived from 5-year survival for lung cancer in terms of cumulative probabilities and life-expectancy data in this study.

In a subsequent article published in the New England Journal of Medicine in 1979, McNeil, Pauker, H. C. Sox, and Tversky asked study participants to imagine that they had lung cancer and 523 [Page 524]to choose between two therapies on the basis of either cumulative probabilities or life-expectancy data. In this study, different groups of respondents received input data that differed in the following ways: whether or not the treatments were identified (as surgery and radiation therapy) and whether the outcomes were framed in terms of the probability of living or the probability of dying. The authors found that the attractiveness of surgery as a treatment choice, relative to radiation therapy, was substantially greater (a) when the treatments were identified rather than unidentified, (b) when the information consisted of life expectancy rather than cumulative probability, and (c) when the problem was framed in terms of the probability of living (survival frame) rather than in terms of the probability of dying (mortality frame). The authors in their conclusion suggest that an awareness of such influences among physicians and patients could help reduce bias and improve the quality of medical decision making.

...