الان چه حسی داری؟درباره اعوجاج ادراکی از تغییرات شدید اخیر در خشم

Previous research on retrospective biases in emotion has been largely concerned with mistakes that are made when people are asked to recall temporally distant affective experiences (e.g. those that occurred weeks or months ago). However, far less is known about people's abilities to accurately track extremely recent shifts in affective experience. Across three experiments, we show that people consistently distort perception of a very recent change in anger after being reminded of a historical act of revenge (i.e. the assassination of Osama bin Laden). Consistent with the implications of the “revenge paradox” (Carlsmith, Wilson, & Gilbert, 2008) these reminders made participants more angry. However, participants believed that this act of revenge had made them less angry—the exact opposite of what happened—provided that their psychological allegiance to the ingroup had been primed. We discuss the implications of our findings in previous research on the interconnections between emotional experience and social categorization processes (Mackie, Maimer, & Smith, 2009), as well as the role of revenge in protecting the interests of the ingroup (Fehr & Gachter, 2002).

Research and theory on affective forecasting (Wilson, Wheatley, Meyers, Gilbert, & Axsom, 2000) have shown that people commit a multitude of errors when they are asked to predict their own emotions. In a study reported by Gilbert, Pinel, Wilson, Blumberg, and Wheatley (1998), for example, researchers asked participants to predict how they would feel in the aftermath of a romantic breakup. Although participants correctly predicted that the breakup would have a negative impact, their actual reactions were neither as intense, nor as long lasting, as anticipated. Interestingly, people often make these and other types of errors in relatively familiar domains, in which they have a great deal of personal experience. Consider the aforementioned study on romantic breakups by Gilbert et al. (1998). In that study, many participants (who were in their late teens or early twenties) were likely to have already experienced the pain of a romantic breakup. Intuitively, one might imagine that such experience might lead people to realize that their mental models about emotional breakups were wrong, thus allowing them to make more accurate predictions in the future. However, even if participants had gained such experience, the data provided little indication that they had been able to learn from it. Ayton, Pott, and Elwakili (2007) provided more direct evidence of people's failure to learn from personal experience, when they asked participants to imagine how they would feel after failing a future driving test. Results showed that “experienced” individuals (i.e. those who had previously failed a driving test) were no more accurate in their predictions than those who had never experienced this event. The affective forecasting literature thus suggests that people make the same types of mistakes over and over again. This state of affairs calls to mind the perspective of Marcus Cicero, who wrote, “Any man can make mistakes, but only an idiot persists in his error.” With all due respect to Mr. Cicero, we would not go so far to suggest that persistent error in affective forecasting represents a lack of intelligence per se. Rather, such mistakes can often reflect foundational limitations in human information processing (cf. Wilson & Brekke, 1994). Moreover, people can sometimes be accurate in their affective forecasts (Wilson & Gilbert, 2003) and there are some conditions in which people can learn from their own previous emotional history (Brown & McConnell, 2011). Nevertheless, given that people often make the same mistakes (even in familiar domains), this raises an important question: what makes it so hard to learn from personal experience? Currently, the most popular explanation is that such errors are attributable to long-term memory biases. In particular, just as people often manifest errors when they attempt to reconstruct the past (Schacter, 2001) this penchant for error has been shown to extend to cases in which people recall their own emotional histories (for reviews, see Levine et al., 2009 and Robinson and Clore, 2002). For example, suppose that people were, in fact, capable of correctly recalling that the pain of their romantic breakups was much shorter than originally anticipated. If so, they could use these accurate memories as a learning experience, increasing the chance that they would make better predictions in the future. However, given that we often distort memories of our own past, this could perpetuate errors when we attempt to forecast the future. Stated another way, people are, in a sense, “learning from the past”. However, the problem lies in the fact that they are remembering the wrong past, one that reflects a memory-based distortion of what really happened. On the assessment of extremely recent changes in mood In nearly all of the research on retrospective memories of emotion, there is considerable delay between (a) when the event originally occurred and (b) when retrospective retrieval occurs, often on the order of days, weeks, or even months (cf. Levine et al., 2009). Given the well-known limitations of long-term memory (see above), it is not surprising to see that errors in emotional memories are so common. However, is it possible that people might show greater accuracy under shorter time frames? Few studies have examined short-term biases in affective retrospection, but Van Boven and Robinson (2012) represents one notable exception. When male and female participants were asked to recall the intensity of a very recent affective experience (i.e. that happened twenty minutes earlier), researchers observed stereotypic biases in the judged intensity of affect, such that (a) men, compared to women, judged their anger as more intense, and (b) women, compared to men, judged their sadness as more intense. Notably, these biases were only evident when, at the time of recall, participants were placed under cognitive load, or when the gender stereotype had been experimentally primed. Given the short time period involved in the Van Boven and Robinson (2012) paradigm, it seems unlikely that the observed errors were due to biases in long term memory. Rather, it seems more likely that their results reflected a bias in how participants interpreted their own emotional experience. In particular, given that emotional experience is often somewhat ambiguous (cf. Bem, 1972), inferences about previous emotional change may involve some degree of social construction ( Martin & Tesser, 1992), in which people rely on the implications of any relevant mental constructs that might be accessible at the time of judgment (cf. Wyer & Srull, 1989). The plausibility of this explanation is bolstered further by the fact that these biases emerged when resources were scarce, or when the relevant (gender-based) expectations had recently been primed, which are precisely the kinds of conditions under which heuristic, accessibility-driven processing is most likely to occur ( Higgins & Bargh, 1987). On the possibility of extremely short term biases in emotional valence Although Van Boven and Robinson's (2012) findings are important, they demonstrated these biases in the realm of intensity: “how much” of a given emotion participants had felt. Precise recollection of how much emotion one felt in the past (e.g. “exactly how sad were you after watching that movie?”) may require cognitive effort, something that people may not have the ability/motivation to exert (Wilson & Brekke, 1994). Relevant, too, is the fact that many of the affective forecasting errors demonstrated in the literature bear on people's inabilities to predict emotional intensity (cf. Wilson & Gilbert, 2003). In general, then, accurate appraisal of emotional intensity may simply be a relatively difficult thing to do. However, what about valence? Suppose that a recent event put you in a bad mood. When reflecting upon this event, would you be accurate in assessing this change in affect? Intuition suggests that, at least for the basic issue of valence—did X make you feel better, or worse?—recollection should be accurate. Moreover, such ability should be especially good if the event in question had happened just moments ago. In this article, we offer a counterintuitive prediction, in that people can make a mistake of valence, even for extremely recent events. To our knowledge, ours is the first program of research to address this kind of error. The goal of this article was to demonstrate the existence of these biases as well as provide some insight into the mechanisms by which they occur. The present research Our curiosity in these matters was stimulated by a line of research by Carlsmith, Wilson, and Gilbert (2008). As these researchers demonstrate, people expect that revenge feels good (“revenge is sweet”). However, this is a domain in which people make a blatant affective forecasting error. In particular, revenge actually makes people feel worse, not better. According to Carlsmith et al. (2008), this is because retribution against the transgressor can trigger thoughts about the misdeeds for which that person is being punished in the first place. This, in turn, can make people feel even worse than they otherwise would. In other words, “revenge can prolong peoples' hedonic reactions to a transgression because punishing others can cause people to continue to think about (rather than to forget) those whom they have punished” ( Carlsmith et al., 2008, p. 1324). It is important to emphasize, however, that the Carlsmith et al. (2008) model stipulates that such negative affect is stimulated by the thoughts about the original transgression, not by the act of revenge itself. In this article, we were interested in an issue that was not addressed by Carlsmith et al. (2008): would people recognize that a revenge-related event had just made them feel worse? Aside from the intuitive sense that people are able to figure this out, research suggests that people often notice when expectations have been violated (Roese & Sherman, 2007). Hence, people might be able to ascertain that revenge had made them angry, precisely because this change was unexpected. However, such optimism might be tempered by a long line of research showing that self introspection is more complex, and more prone to error, than one might imagine ( Bem, 1972, Chaiken and Baldwin, 1981, Dijksterhuis et al., 2006, Salancik and Conway, 1975, Stanovich, 2013, Wilson and Schooler, 2008 and Wilson and Stone, 1985). Hence, although one could make a case for optimism in terms of retrospective accuracy in our paradigm, this literature led us to take a more pessimistic position. We predicted that participants, despite the seemingly “optimal” conditions for accuracy, would mistakenly conclude that exposure to an act of revenge made them feel better, even though they actually felt worse. On the relevance of anger to the revenge paradox In the introduction of their paper, Carlsmith et al. (2008) place considerable emphasis on cultural expectations as to the cathartic potential of revenge, especially as it pertains to anger. In particular, they suggest that “there is widespread acceptance of the notion that aggression, either directed against the target of one's ire or displaced to some other object, relieves the tension, and thus the anger, that had been pent up inside” (Carlsmith et al., 2008, p. 1316). Hence, even though Carlsmith et al. (2008) sometimes frame the revenge paradox in global terms (e.g. as a general expectation to “feel better”), these considerations suggests that these expectations are rooted in an expected reduction in negative affect—and anger in particular—rather than an increase in positive affect. As for the actual consequences of revenge, the Carlsmith et al. (2008) model suggests, too, that anger may often be relevant. There is a wealth of evidence suggesting that people feel anger when they perceive that others have violated social norm(s), especially when those actions harm others (e.g. Alicke, 2000, Goldberg et al., 1999, Russell and Giner-Sorolla, 2011, Schlenker, 1997, Tetlock et al., 2007 and Weiner, 1995). This is not to say that perception of wrongdoing cannot trigger other types of negative feelings. However, across all of the studies presented in this article, the effects obtained with anger were, by far, the most stable. The stability of our findings with respect to anger is consistent with the fact that intentional social norm violations, especially when such actions harm the self and/or ingroup members, should consistently elicit increased feelings of anger. Ironically, then, the type of emotion for which people might most consistently expect hedonic benefit—a cathartic-like reduction of anger—represents the very type of emotion domain along which people are actually likely to feel worse. Summary of research goals The present research adds to existing knowledge about retrospective biases in emotion (Levine et al., 2009) with particular emphasis on the possibility that people might show lack of insight into extremely recent changes in mood. In principle, this research could have been conducted in several different substantive domains. However, in order to generate a persuasive set of findings, we choose one domain, and continue this focus across a programmatic line of studies. The affective dynamics associated with revenge (Carlsmith et al., 2008) represented an excellent match to our objectives, for several reasons. For one thing, we wished to “set the bar” low to facilitate the chance that participants might maintain some degree of retrospective accuracy. To our knowledge, researchers have identified only one domain in which people hold incorrect expectations about valence, and that is revenge. Moreover, the present research allowed us to show retrospective biases with a rather familiar type of emotion. It would be one thing to show retrospective biases for relatively subtle/esoteric emotional experiences, such as ennui or Schadenfreude. It is quite another thing, however, to demonstrate bias for anger, arguably one of the least subtle of all human emotions. Finally, apart from retrospective assessment, our research provides more insight into the dynamics of revenge, a topic of scholarly interest in its own right (de Quervain et al., 2004). Hence, while our main goal was to learn more about retrospective biases, a secondary but still important goal was to learn more about revenge. As we shall show, this endeavor builds upon but also adds to previous research on the interconnections between emotional experience and social categorization/social identity processes (Mackie, Maimer, & Smith, 2009) and highlights the role of revenge in protecting the interests of the ingroup (cf. Fehr & Gächter, 2002).1 Preliminary Study One The killing of Osama bin Laden was widely perceived by many Americans as justifiable payback for his role in 9/11 attacks (Bowden, 2012). Moreover, although President Obama did not actually use the word “revenge” when announcing bin Laden's death, his description leaves little doubt that bin Laden's death was retaliation for bin Laden's earlier attacks: “On nights like this one, we can say to those families who lost loved ones to al Qaeda's terror: justice has been done”. It is safe to say, then, that thoughts about bin Laden are associated with people's memories of the 9/11 attacks, and vice versa. Given that Americans are still likely to associate a negative emotion with the 9/11 attacks (Lambert et al., 2010), the Carlsmith et al. (2008) model suggests that reminders about the killing of Osama bin Laden's death would actually make people feel worse, not better. We conducted an initial preliminary study to validate this assumption. Using a sample of American residents drawn from the MTurk website (N = 232), participants were randomly assigned to read a short historical account of the Olympic Games devoid of any political content or, alternatively, a passage of similar length which described the main events surrounding the capture and death of Osama bin Laden (e.g. “Osama bin Laden, the mastermind of the most devastating attack on American soil in modern times and the most hunted man in the world, was killed in a firefight with United States forces in Pakistan.”). Afterwards, all participants completed a standard mood inventory, where they rated their ambient mood along a scale that ranged from 0 to 6. Our primary interest in this study was in constructing a composite measure of anger, based on an average of angry, mad, irritated, and irate (alpha > .80). Consistent with the framework offered by Carlsmith et al. (2008), we found greater anger among participants who were reminded (vs. not reminded) of the revenge killing of bin Laden, Ms = 2.38 vs. 1.51, F (1, 230) = 38.38, p < .001, ηp2 = .14. 2