Fluency, Familiarity, and Risk Perception

Low processing fluency fosters the impression that a stimulus is unfamiliar, which in turn results in perceptions of higher risk, independent of whether the risk is desirable or undesirable. In Studies 1 and 2, ostensible food additives were rated as more harmful when their names were difficult to pronounce than when their names were easy to pronounce; mediation analyses indicated that this effect was mediated by the perceived novelty of the substance. In Study 3, amusement-park rides were rated as more likely to make one sick (an undesirable risk) and also as more exciting and adventurous (a desirable risk) when their names were difficult to pronounce than when their names were easy to pronounce. Guided by expected-utility theory, researchers have often assumed that risk perception is an analytic procedure: people presumably assess the severity and likelihood of possible outcomes and integrate this information through an expectation-based calculus (for a review, see Harless & Camerer, 1994). However, the observation that subjective perceptions of risk are insensitive to changes in probability (e.g., Kahneman & Ritov, 1994; Kahneman, Ritov, & Schkade, 1999) challenged this assumption. Accordingly, recent research conceptualized lay risk judgment as an intuitive rather than an analytic process and emphasized the role of feelings—like worry, fear, dread, and anxiety—in risk perception (for reviews, see Loewenstein, Weber, Hsee, & Welch, 2001; Slovic, Finucane, Peters, & MacGregor, 2004). The present research contributes to this ‘‘risk as feeling’’ approach (Loewenstein et al., 2001) by exploring the role of a previously neglected experiential variable, namely, the fluency with which a stimulus can be processed. As reviewed below, fluently processed stimuli are judged as more familiar and elicit a more positive affective response than disfluently processed stimuli (for a review, see Schwarz, 2004). Both familiarity and affect are assumed to be involved in intuitive judgments of risk. In an influential series of studies, Zajonc (1968) observed that liking for a neutral stimulus increases with the frequency of exposure. He suggested that people prefer previously seen, familiar stimuli over novel ones because novel stimuli are associated with uncertainty, whereas familiar stimuli are considered safe, at least in the absence of negative memories (Zajonc, 1980, 1998). Consistent with the hypothesized familiarity-safety association, the perceived risks of technologies (e.g., Richardson, Sorenson, & Soderstrom, 1987), investments (e.g., Weber, Siebenmorgen, & Weber, 2005), and leisure activities (e.g., Zuckerman, 1979) have been found to decrease as their familiarity increases. Unfortunately, the role of familiarity, per se, is difficult to isolate in such studies, and the observed effects may instead reflect differences in knowledge and previous experience or desensitization to the threat. To avoid these ambiguities, the present studies presented only novel stimuli and manipulated their perceived familiarity through manipulations of processing fluency. Because familiar material is easier to process than novel material, people (erroneously) infer familiarity from ease of processing, even when fluent processing is merely due to presentation variables like exposure duration, high figure-ground contrast, or an easy-to-read print font (for reviews, see Alter & Oppenheimer, 2008; Reber, Schwarz, & Winkielman, 2004). In cognitive research, this fluencyfamiliarity link gives rise to erroneous recognition judgments for perceptually easy-to-process stimuli (e.g., Whittlesea, Jacoby, & Girard, 1990) and to strong feelings of knowing (e.g., Koriat & LevySadot, 2001). In social psychological research, fluent processing of a statement results in the impression that one has heard it before, suggesting that the opinion is popular (e.g., Weaver, Garcia, Schwarz, & Miller, 2007) and increasing the likelihood that the statement is accepted as true (e.g., Reber & Schwarz, 1999). If the Address correspondence to Hyunjin Song or Norbert Schwarz, Department of Psychology, University of Michigan, 530 Church St., Ann Arbor, MI 48109-1043; e-mail: [email protected] or [email protected]. PSYCHOLOGICAL SCIENCE Volume 20—Number 2 135 Copyright r 2009 Association for Psychological Science apparent familiarity of a stimulus plays a prominent role in intuitive judgments of risk, novel stimuli should therefore be perceived as less risky when they are easy rather than difficult to process. Numerous variables, from figure-ground contrast and the readability of print fonts to the ease with which a name can be pronounced, can influence processing fluency and have been found to exert comparable effects on a wide variety of judgments (for reviews, see Alter & Oppenheimer, 2008; Reber et al., 2004). The present studies used ease of pronunciation as a fluency manipulation. We found that ostensible food additives were rated as more harmful (Study 1) when their names were difficult rather than easy to pronounce and that this effect was mediated by the perceived familiarity of the stimuli (Study 2). However, previous research also showed that high fluency is experienced as pleasant and elicits a low-level affective response, as indicated by increased activation of zygomaticus major, the muscle involved in smiling (Winkielman & Cacioppo, 2001). This positive affective response may further contribute to evaluations of fluently processed stimuli as less risky. We addressed this possibility by assessing the impact of processing fluency on evaluations of risks with a positive or negative connotation. As indicated by a large body of research, positive affect elicits more favorable evaluations than negative affect (see Schwarz & Clore, 2007, for a review). If fluency effects on intuitive judgments of risk are driven by the affect associated with fluent processing, low processing fluency should therefore result in more negative evaluations, and high processing fluency should result in more positive evaluations, independent of the positive or negative connotations of a given risk. We found no support for this prediction (Study 3); instead, stimuli with difficult-to-pronounce names were rated as more risky, independent of valence.