Implicit Automatic Low Effort Rapid High Capacity The Default Process Holistic, Perceptual Non-Verbal Modular Cognition Associative Domain Specific Contextualized Pragmatic Parallel Stereotypical Independent of General Intelligence Independent of Working Memory
Conscious Explicit Controlled High Effort Slow Low Capacity An Intervening Process Analytic, Reflective Linked to Language Fluid Intelligence Rule-Based Domain-General Abstract Logical Sequential Egalitarian
Linked to General Intelligence
Limited by Working Memory Capacity
ǂ Summarized from (Evans, 2008)
Figure 6 lists common information processing characteristics referenced from social and cognitive psychology, behavioral economics, and philosophy literatures, and categorizes them under two information processing routes, generically referred to as Systems 1 and 2 (for a review, see Evans, 2008). Standard dual-process theory suggests that System 1 and 2 processes operate with a default-interventionist relationship. Less effortful, System 1 processes are activated first, and are intervened upon by System 2 processes under certain conditions. If we consider consumer choice as a generic form of individual choice, then there is a substantial body of literature on non-effortful, System 1 effects on judgment and choice that are potential influences on the attraction effect. Indeed, results from the moderators discussed thus far suggest that less
effortful, more automatic, processes are involved in the creation of the attraction effect. For example, stronger attraction effects are observed with decision makers who are:
More likely to use in intuition (Mao & Oppewal, 2009, 2012)
Cognitively depleted or resource constrained (Pocheptsova et al., 2009)
Forced to use a perceptual mode of stimulus processing (Hamilton et al., 2007)
Reliant on brand loyalty and popularity (Mishra et al., 1993), and country of origin (Chuang & Yen, 2007) as heuristic cues.
In addition, we would expect an attenuated attraction effect when decision makers use more conscientious, deliberate, System 2 process. And indeed prior research supports that the attraction effect is substantially reduced or mitigated when decision makers:
Operate under a more cautious, conservative, prevention mode of decision making (Mourali et al., 2007; Malkoc et al., 2008)
Are forced to deliberate their decision (Hamilton et al., 2007) Exhibit high levels of task involvement, (Mishra et al., 1993).
Based on the premise that System 1 processes tend to accentuate, and System 2 processes mitigate the attraction effect, future research can do well to identify new moderators in context manipulations that trigger shifts from one system of thinking to another.
If the attraction effect does stem from a System 1 process, we must then ask ourselves what processes and mechanisms could the decoy trigger to create increased preference for the Target. A potentially fruitful avenue to explore is how increased processing fluency, driven by the presence of the decoy, can impart positive affect onto the Target.
Fluency
Fluency refers to a general ease or difficulty that accompanies the reaction to a stimulus. (Dis)fluency has been shown to trigger changes in judgment and decision making, and has been shown to fully explain the strength of other decision context effects, such as the compromise effect (Novemsky, Dhar, Schwarz, & Simonson, 2007). As applied to the attraction effect, fluency may be the vehicle through which positive affect is created and then imparted on to the Target. Theoretically, the presence of the decoy can increase processing fluency through similarity and subjective dominance. A subjective dominance relationship creates a salient, readily available easy justification between options (Novemsky et al., 2007). Similarity (between T and D) may operate as repeated exposure of a singular option type, facilitating conceptual processing of both T and D (Zajonc, 1968). In addition, similarity between T and D may also make the Target more cognitively accessible and easier to recall from memory. From an empirical standpoint, there is already physiologically evidence of decision fluency in the presence of a decoy (Hedgecock & Rao, 2009).
Increased processing fluency could then be translated into various forms of positive affect thereby increasing preference for the target. For example, increased processing fluency has been shown to:
Increase general liking (Bornstein, 1989; Schwarz, 2004; Zajonc, 1968)
Improve judgments of truth and trustworthiness (for a review, see Schwarz, 2004) Increase feelings of familiarity (e.g. Whittlesea et al., 1990)
Be interpreted as increased prevalence or popularity4
Any one of these fluency-induced positive affects can help increase preference. But why would this positive affect be associated with the Target and not the Competitor?
First, in the context of {T,C,D}, fluency would be created through similarity and subjective dominance only when evaluating {T,D}, but not {T,C} or {C,D}. In fact, evaluations of {T,C} and {C,D} are physiologically (Hedgecock & Rao, 2009) and cognitively (Luce, Bettman & Payne, 2001) disfluent. Second, attribution of positive affect generally occurs through simultaneous activation. If positive affect is experienced while evaluating the Target, consumers will, by default, assume the Target to be the source of that positive affect. This ‘aboutness principle’ of inference asserts consumers naturally assume the thoughts that come to mind while thinking about a concept are relevant to that concept (Higgins, 1998; for a review see Schwarz, 2004).
Thus, the positive sense of ease that comes with fluency is likely attributed to and interpreted as liking or preference for the Target. Under this framework, there is no range- or frequency-based valuation or analysis to arrive at a preference. Of the existing explanations, affect-as-information is most similar to subjective dominance explanations. However the major difference between the two types of explanations is that affect-as-information is reliant upon experienced positive affect, not justification, as the source of preference.
The affect-as-information explanation can be tested by manipulating processing fluency and then measuring preference for the Target. Decreasing Target fluency is expected to decrease liking, familiarity, and/or trustworthiness, and should decrease preference for the Target and
4 This belief operates through the availability heuristic where the more fluently content is processed, the more
dampen or neutralize the attraction effect. Increasing the relative fluency of the Competitor should also weaken the attraction effect. Empirically, Competitor fluency can be increased by subconsciously priming decision makers with a stimulus that is closely related, but not comparable, to the Competitor. Target processing can be made more disfluent by presenting the Target and Decoy pair in a more difficult to interpret font, color or scale. Neither of these manipulations should affect the thought content of the decision set, but should impact processing fluency to reduce the attraction effect.
Summary
Research on the attraction effect can benefit from the application to a generic dual- processing framework such as those found in cognitive and social psychology, philosophy, and behavioral economics. Current attraction effect research primarily operates under the premise that the decision maker uses only conscientious, deliberate, System 2 thought processes. Sparse consideration is given to the role that unconscious, rapid, automatic, System 1 thought processes may play in the formation of the attraction effect. Normatively perfect choices are rarely made because consumers are routinely, subconsciously, influenced by the decision context and the information, associations, and emotions that the choice context may activate.
There is still much research to be conducted on the attraction effect. If consumer preference is indeed constructed and not revealed, a diverse pool of metacognitive influences must also be considered. Moods, emotions, affect states, fluency, and individual differences should all be examined as potential explanatory variables of the attraction effect. Context effects such as the compromise and deferral effects have been shown to be mediated through processes chosen based on changes in decision fluency, and there is little reason to believe attraction effect
implications of fluency states maps well to current findings on the attraction effect, and hold great potential in helping to explain the automatic and affectual components of the attraction effect.