Does It Matter If It Involves My Group? How the Importance of Collective-Esteem Influences a Group- Based Framing Task

Studies that have addressed questions concerning when framing effects are likely to occur have produced mixed results. In this article we examine how differences in personality factors influence a group–based framing task. Specifically, when high collective importance individuals evaluated a framing task involving their group no framing effects were observed. Different results were obtained for low collective individuals. When low collective individuals evaluated a task involving their group, the typical framing effects were found. These results are interpreted in light of the dual analytic-holistic analysis. Most people agree that preference for one option over another should remain constant regardless of whether the options are presented positively, in terms of gains, or negatively, in terms of losses. Yet, as a large body of research demonstrates, people do tend to change their preferences based on how the alternatives are presented. This change, or preference reversal, has become known as framing. The framing phenomenon, as derived from prospect theory (Kahneman &Tversky, 1979), has become the most widely tested example of irrational decision–making. For over two decades investigations in numerous research domains have tested and applied this phenomenon—proof of its viability and its ability to stimulate interest. Most of the framing tasks investigated follow the standard risky-choice format similar to the Asian disease problem (see Appendix 1). However, not all research has been supportive and the framing effect seems to be plagued with a number of inconsistencies (see Kuhberger, 1998 and Levin, Schneider & Gaeth, 1998 for reviews). For example, Fagley and Miller (1990) investigated how personal factors of risk–taking predisposition and sex would affect the strength of framing. Their overall findings suggested a weaker framing effect for males and an unclear relationship for risk–taking predisposition which seemed to be obscured by the situation or “arena” of presentation. Following these findings, Schneider (1992) proposed that differences in aspiration accounted for the lack of consistency between framing arenas. Further research investigating the framing phenomenon demonstrated that person factors such as need for cognition (Chatterjee, Heath, Milberg & France, 2000; Smith & Levin, 1996; Zhang & Buda, 1999), perceptions of threat (Highhouse & Yuce, 1996; Highhouse & Paese, 1996), individual perceptions (Frisch, 1993) and personality traits including intuition (Levin, Gaeth & Schreiber, 2002) can all effect the likelihood of framing. Similar to person factors, a number of contextual factors have also been shown to affect the potential for framing. Examples include statistical versus medical context (Bless, Betsch & Franzen 1998), decision time (Takemura, 1992), target of the task (Levin&Chapman, 1990, Wang, 2001) and presentation format (Fagley & Miller, 1997). Attesting to much of this inconsistency, Kuhberger (1998) found in a meta–analytic review of framing research that the overall effect for framing was relatively small (d = 0.33). The strongest framing effect seemed to be found with scenarios that most closely followed the Asian disease problem (see Appendix 1) and, not surprisingly, some types of framing scenarios seemed to produce no framing effects at all. DUAL–PROCESS APPROACH In recent decades, researchers investigating various psychological phenomena have developed numerous dual–process models (e.g., Chaiken, 1987; Epstein, Lipson, Holstein&Huh, 1992; Johnson– Laird & Byrne, 1993; Petty & Cacioppo 1986; Stanovich & West, 2000). The focus of this research has spanned across both social and cognitive psychology. Because of the extensive history of the dual–processing approach, it has been the foundation for a large number of studies that have demonstrated its potency. Consequently, it remains one of the most investigated and important frameworks in social psychology (see Chaiken & Trope, 1999 for review). While most dual–process models focus on specific topics and have theoretically identifying characteristics, they nonetheless maintain a central theme. They posit that two distinct processing modes or systems exist. One system requires relatively more effort and performs a more comprehensive analysis of the information. While the other system requires relatively less effort and, consequently, performs a less extensive analysis of the information. ANALYTIC/HOLISTIC DISTINCTION Recently, McElroy and Seta (2003, 2004) proposed a dual–process account to explain some of the discrepancies found within the framing literature. According to this view, individuals process decision tasks using both analytic and holistic processing styles. In a typical decision task, either the analytic or holistic processing style can be particularly influential depending upon personality, relevance and hemispheric factors. Consistent with prior dual–process theories (e.g., Chaiken, 1987; Epstein, Lipson, Holstein & Huh, 1992; Johnson–Laird & Byrne, 1993; Petty & Cacioppo 1986; Stanovich & West, 2000), the analytic/holistic model proposes two systems of processing that operate under different principles. Although most individuals typically operate under the less effortful–holistic processing style, several factors can either induce or predispose individuals to process a task using the more effortful–analytic processing style. When the task is of sufficiently high relevance, the more effortful, analytic system is induced whereas if the task is of low personal relevance, the less effortful–holistic system is induced. Further, both systems have a personality component that can predispose individuals to one or the other systems of processing. Demonstrating these points, McElroy and Seta (2003) found that when participants were given the traditional Asian disease problem (see Appendix 1), typical framing effects occurred only when the task was presented as being of relatively low personal relevance or when individuals had a personality predisposition favoring holistic thought (McElroy & Seta, 2003). The analytic/holistic account differs from prior dual–processing accounts in two fundamental ways. First, the analytic/holistic model suggests that, in addition to effort and personality factors, the functional specializations of the respective hemispheres provide a representative basis for the two system of processing (McElroy & Seta, 2004). The left hemisphere represents the analytic system whereas the right hemisphere represents the holistic system. The analytic/holistic view differs further from other dual–process models in the functionality of the two systems. The analytic system focuses on breaking information down into distinct and unconnected elements, then combining this information and focusing on quantitative weights for solutions. The holistic system differs from other accounts primarily because of “contextual referencing”— the holistic system is especially sensitive to cues within the surrounding context. The contextual cues then act to elicit a framework within which the problem can be evaluated. This process is largely derived from work investigating language comprehension (e.g., Brownell, Pincus, Blum, Rehak, & Winner, 1997). According to the dual analytic–holistic approach then, if a person identifies the target of a framing task as an important identifying aspect of their self–construct, then related information should be of especially strong personal relevance. Consequently, individuals should process the task with the more effortful–analytic processing style and the predicted framing effects should not occur. Conversely, if the target of the task is not particularly tied to their self–construct, then personal relevance should be low and the less effortful–holistic style should be selected. This should result in special sensitivity to contextual cues (i.e., the frame) and commonly found framing effects should occur. One way that individuals differ in their self– identifying attachments is in their identity orientation toward groups.