A New Paradigm For Studying The Economic And Behavioral Consequences Of Framing Health-Related Decisions

Traditional attribute framing effects occur when the same object is evaluated differently depending on whether a particular attribute is labeled or framed in positive or negative terms. For example, in one of our earlier studies, “80% lean ground beef” was evaluated more favorably and was “worth” 8 cents more per pound than “20% fat ground beef.” In the present study of health-related judgments and decisions, attribute framing effects were extended to situations where consumers had to infer framing valence depending on whether one’s health status was described in comparison to a lower standard or a higher standard of reference. For example, a person’s health status was rated higher when the same level of vitamin intake was stated in terms of its distance above an established low-protection level compared to when it was stated in terms of its distance below an established high-protection level. 1.0 Introduction versky and Kahneman’s (1981) classic ―Asian disease problem‖ sparked considerable interest in ―information framing‖ effects—how the same information can lead to different decisions depending on the way that information is labeled or framed. In the case of the Asian disease problem, respondents were more apt to make risky decisions when options were described in terms of potential losses than when they were described in terms of potential gains. This is known as the ―Risky Choice Framing Effect.‖ A somewhat simpler framing effect that doesn’t necessarily involve the element of risk is known as the ―Attribute-Framing Effect.‖ An Attribute-Framing Effect is found when the evaluation of a given object or product differs depending on whether a key attribute is described in positive terms or negative terms. Because a single attribute within any given context is the subject of the framing manipulation, attribute-framing experiments provide researchers with a very general and, indeed, flexible paradigm to test how the positive/negative valence of available information influences one's judgment processes. Experimental studies of attribute framing often involve consumer judgments, including health-related judgments. For instance, Levin and Gaeth (1988) showed that perceptions of the quality and healthiness of ground beef depended on whether the beef is labeled as ―75% lean‖ or ―25% fat.‖ Levin, Gaeth, Schreiber and Lauriola (2002) added an analysis of the economic impact of attribute framing by showing that consumers were willing to spend an average of 8.2 cents more for a one-pound package of ―80% lean ground beef‖ than for a one-pound package of ―20% fat ground beef.‖ Attribute framing sometimes involves the evaluation of risky behaviors. For instance, Van Schie & Van der Pligt, (1995) showed that emphasizing "potential positive outcomes" rather than ___________________ Readers with comments or questions are encouraged to contact the authors via email. "negative ones" in a risky option led to increased preference for this option. This has important implications for T International Business & Economics Research Journal Volume 2, Number 9 26 health-related decisions. McNeil, Pauker, Sox and Tversky (1982) showed that a medical treatment was more apt to be selected when it was described in terms of success rate rather than failure rate. After an extensive literature review, Levin, Schneider & Gaeth (1998) came to the conclusion that the processes accounting for attribute framing effects are associative processes which produce more favorable evaluations when the framed attribute primes positive associations rather than negative ones. We refer to this process and its expected effects as the "standard" attribute framing effect. In the present study we provide a different and, perhaps more general, form of attribute-framing, where a given attribute is first described relative to a reference point, and then the framing valence is inferred based on whether the distance from the reference point is viewed as a positive or a negative. We define the positive/negative conditions in terms of valence consistent shifts associated with gains and losses relative to higher or lower standards of risk and protection. Thus, describing one's health status as a given level above a lower standard of risk is expected to produce a consistent negative shift in valuation compared to describing one’s health status as a given level below a higher standard of risk which is expected to produce a consistent positive evaluation shift. 2.0 Aims and Scope In the present study, we first test the so-called "general" attribute-framing effect by comparing valence consistent shifts across variations in standard of reference. Thus, for example, we expect positive consistent shifts when evaluating one's exposure to risk as being below a higher standard of risk. Conversely, we expect negative consistent shifts when evaluating one's exposure to risk as being above a lower standard of risk. Second, we expand on previous research by comparing this "general" attribute-framing effect in two different health domains: either evaluating one's exposure to risk factors such as high cholesterol or judging one's exposure to protective factors such as vitamin intake. We administered four different versions of the attribute-framing problem by providing different frames of reference for a health-risk and health-protection scenario. (A, B, C, and D in the figure). We define each condition as representing positive or negative framing, depending on whether that condition represents a particular value as being below a stated high-risk or high-protection level, or as being above a stated low-risk or low-protection level. This represents an extension of earlier definitions of attribute framing. Condition A describes one's blood cholesterol level as 4/10 above the low-risk level. We define this condition as negative framing because one's blood cholesterol level is described as an increase in risk relative to a low-risk reference point, which is apt to be viewed as bad or negative. Condition B describes one's blood cholesterol level as 6/10 below the high-risk level. We define this condition as positive framing because one's blood cholesterol level is described as a loss of risk relative to a high-risk reference point, which is apt to be viewed as good or positive. We thus expect risk to be rated higher in Condition A than in Condition B. Condition C describes one's Vitamin E consumption level as 4/10 above the low-protection level. We define this condition as positive framing because one's vitamin consumption level is described as a gain of protection in comparison to a low-protection reference point. Condition D describes one's Vitamin E consumption Level as 6/10 below the high-protection level. We define this condition as negative framing because one's Vitamin Consumption level is described as a loss of protection in comparison to a high-protection level. We thus expect protection level to be rated higher in Condition C than in Condition D. International Business & Economics Research Journal Volume 2, Number 9