|
Article Excerpt
Dual process models conceptualize two systems of processing that are activated when presented with a decision task, the quick and affective System 1 and the deliberative and rule-based System 2. In this article, we explore whether the affective component of System 1 has the potential to interfere with the information integration component of System 2
by utilizing everyday consumer decision-making situations that require the integration of provided information to make optimal choices. We posit that if the provided information has an affective System 1 element, then the affective reaction serves as an input to the System 2 process of information integration. Such an affective input has the potential to cause improper information integration resulting in a biased mental representation, which in turn leads to suboptimal choices. Across three experiments, we test the interfering role of affect in information integration. Experiment 1 establishes the mediating role of the affective System 1 generating a preference for the suboptimal option and rules out the alternate account of analytical skills. Experiments 2 and 3 provide converging evidence for the proposed account that System 1 interferes with System 2 and argue against the alternate account of System 1 directly influencing choice.
Key words: system 1 and system 2; dual process models; choice; affect; irrationality; two-part tariff; misattribution; information integration; stimulus-induced affect; affect and cognition; behavioral decision theory
History: This paper was received April 5, 2005, and was with the authors 13 months for 2 revisions; processed by Ravi Dhar.
1. Introduction
Consider a consumer contemplating the purchase of a toaster oven at a store. The consumer carefully compares the various models and attempts to weigh the risks and benefits of consumption, to balance the appeals of ergonomics and features against the costs and comparison points. Now consider the same consumer at dinner--when the dessert tray rolls around, one look and he immediately succumbs to the lure of the cheesecake and rationalizes the choice with notions of deservedness or the promise to spend an extra 15 minutes on the treadmill. Situations abound in everyday decision making when it is assumed that affective and deliberative systems are set in motion simultaneously and work independently to provide their respective feedback and influence the decision outcomes. Extant findings have amply demonstrated such influences on consumer decisions (Hoch and Lowenstein 1991, Hsee and Kunreuther 2000, Luce 1998, Luce et al. 1999, Pham et al. 2001, Shiv and Fedorikhin 1999). The affective system is considered to have the ability to bias judgments and lead to suboptimal decisions, and thus a common recommendation is to rely on the deliberative system. Adam Smith (1759, Part 1, [section]2) long ago warned of "dangerous passions," suggesting that affect was capable of undermining normative decision-making ability. Although the deliberative system's responses are considered slower, they are believed to remain unbiased by the affective cues in the decision environment. However, in this work we show that relying on the deliberative system does not necessarily lead to an optimal choice, since affective reactions have the potential to interfere with the ongoing processes of the deliberative system and thus bias the output. Specifically, we posit that along with providing a direct decision output, affect also has the potential to interfere with the working of the deliberative system.
In this article, we utilize the nomenclature of Systems 1 and 2 suggested by decision theorists (Kahneman and Frederick 2002, Stanovich and West 2002, Sloman 2002). An affective response is part of System 1 processing (the fast, affectively charged, and automatic system) and a deliberative response belongs to System 2 processing (the slower, rule-based, and deliberatively controlled system). Therefore, System 1 being intuitive and faster provides a quick output, while System 2 being slow and rule based takes longer to generate an output.
Systems 1 and 2 are conceptualized as two distinct entities that are set in motion simultaneously and work independently to provide their respective feedback when faced with a decision task. Conceptualizing these two systems as independent modules, however, leaves a crucial question unanswered. Can the quick and affectively charged System 1 influence the slower and rule-based System 2? In other words, instead of directly influencing the decision, can System 1 influence the decision by interfering with the working of System 2? The objective of this article is to demonstrate the interference of the quick and affective System 1 in the working of the rule based System 2. To this end, we utilize a specific aspect of System 1, affect, and show how it interferes with a specific aspect of System 2, information integration.
An inquiry into the interference of System 1 into System 2 has both managerial and theoretical relevance. From a managerial standpoint, many decisions require the consumer to make a choice after comparing and contrasting the presented information across two or more options (Dhar and Simonson 1999). An optimal choice in such instances entails an unbiased System 2 response. However, if one piece of information produces a System 1 response (e.g., an affective reaction), then it could lead to a suboptimal choice by interfering with System 2 processing (e.g., information integration). By utilizing common choice tasks such as choosing between bank accounts and cell phone plans, this article highlights the relevance of such System 1 interference with System 2 for everyday consumer decision making.
From a theoretical standpoint, a finding of System 1 interfering with System 2 would be relevant to the domain of decision-making models. Many decision-making models (e.g., Fishbein and Ajzen 1974, Kahneman and Tversky 1979, Von Neumann and Morgenstern 1944) often propose a decision calculus requiring deliberative, rule-based System 2 processing to integrate various pieces of information. As marketing researchers have extensively employed these models of information integration in making predictions, an examination of how System I can potentially interfere with the processing of System 2 would provide valuable insights in predicting consumer preferences and choices. In addition, this would be among the first attempts to gather empirical evidence testing the conjecture that System 1, along with a direct response, may also interfere with System 2 processes.
The article is structured in the following manner. We first briefly review literature, reporting findings on System 1 and System 2 processes that are relevant to this work, and present our conceptualization. Second, we describe the general format of the stimulus used across all experiments. Third, across three experiments we test for the proposed "interference" account and rule out alternate accounts. Finally, we explicate the theoretical and managerial relevance of the findings.
2. The Two Systems of Processing
The notion of two parallel systems providing feedback to a decision task has gained significance recently (Kahneman 2003, Stanovich 2004, Stanovich and West 2002, Sloman 2002). Across psychological and neurological domains, researchers have provided evidence for two systems competing for control in inference making and action (Goel et al. 2000). System I tries to do this by employing affectively charged, automatic, rapid, parallel, heuristic-based processes that are relatively undemanding of cognitive capacity (Kahneman 2003, Stanovich 2004, Sloman 2002). System I includes fast and instinctive processes formed by associative learning methods that operate largely beyond awareness. On the other hand, System 2 is controlled, rule-based, slow, serial, and deliberative and is constrained by working memory capacity. Unlike System 1, System 2 is capable of constructing mental representations and simulations of the future based on hypothetical or abstract thought (Evans 2003). Stanovich (2004) states that while System 1 seems to be hard wired, System 2 operations are cognitively more effortful and take longer to run to completion. Research has also documented that people hold intuitive or System i beliefs with greater confidence since they can be arrived at faster and more easily. Only in the face of strong arguments are individuals willing to exchange an intuitive belief for a nonintuitive one (Simmons and Nelson 2006).
2.1. Interactions Between Systems 1 and 2
An area recently gaining attention is the notion of System 1 interacting with System 2. System 1 forms a bulk of brain processes and takes place rapidly and outside conscious awareness. Theorists have recently posited that not only do System 1 processes directly trigger responses of their own, they may also provide input to System 2 processes, "... thus biasing analytical processing by the nature of the cognitive representations given to it" (Stanovich 2004, p. 52). In the following paragraphs, we examine evidence from two streams of literature that provide support to the notion of an interaction between Systems 1 and 2.
Affect and Reasoning. Blanchette and Richards (2004) provide a demonstration of how System 1 (affect) influences System 2 (logical reasoning). They gave participants a conditional reasoning task (i.e., if p, then q) and manipulated the affective nature of the statements used as p and q. They asked participants to draw inferences based on these statements. Results showed that participants were more likely to draw invalid or wrong inferences in response to affective compared to neutral statements. The authors suggest that this might occur because prior to drawing any inference people represent...
|
