When Dreaming Is Believing; The (Motivated) Interpretation of (PDF)

ATTITUDES AND SOCIAL COGNITION When Dreaming Is Believing: The (Motivated) Interpretation of Dreams Carey K. Morewedge Carnegie Mellon University Michael I. Norton Harvard University This research investigated laypeople’s interpretation of their dreams. Participants from both Eastern and Western cultures believed that dreams contain hidden truths (Study 1) and considered dreams to provide more meaningful information about the world than similar waking thoughts (Studies 2 and 3). The meaningfulness attributed to specific dreams, however, was moderated by the extent to which the content of those dreams accorded with participants’ preexisting beliefs—from the theories they endorsed to attitudes toward acquaintances, relationships with friends, and faith in God (Studies 3–6). Finally, dream content influenced judgment: Participants reported greater affection for a friend after considering a dream in which a friend protected rather than betrayed them (Study 5) and were equally reluctant to fly after dreaming or learning of a plane crash (Studies 2 and 3). Together, these results suggest that people engage in motivated interpretation of their dreams and that these interpretations impact their everyday lives. Keywords: anchoring, attribution, dreams, motivated reasoning, unconscious thought Yet it is in our idleness, in our dreams, that the submerged truth sometimes comes to the top.—Virginia Woolf, A Room of One’s Own A dream which is not interpreted is like a letter which is not read.— The Talmud Each morning, many people glance suspiciously across the bed at the person who broke their heart moments before with imagined infidelities, or fight the urge to change travel plans despite having foreseen their death in a fiery plane crash. The bitter thoughts and strong emotions that such experiences evoke attest to the potency of information “revealed” in dreams. In the present investigation, we examined laypersons’ beliefs about the importance of dreams, the manner in which they interpret dreams, and the consequences of these interpretations on their beliefs and behavior. Psycholo- gists’ interpretations of the meaning of dreams range widely, from a view of dreams as the by-product of increased activity in brain regions engaged during sleep (Muzur, Pace-Schott, & Hobson, 2002) to a view of dreams as “the royal road to the unconscious,” which reveal hidden truths (Freud, 1900/1953; Jung, 1974; Weg- ner, Wenzlaff, & Kozak, 2004). We suggest that—despite dis- agreement among scientists—laypeople endorse the latter perspec- tive, holding a general belief that dreams provide meaningful insight into both themselves and their world. In addition, we propose that people’s interpretation of the meaningfulness of any specific dream is impacted, by the extent to which that dream accords with their beliefs and desires when awake, and that these interpretations subsequently influence the impact of dreams on their diurnal (i.e., waking) lives. THE INTERPRETATION OF DREAMS Aside from the ubiquity of psychoanalytic theories in art, liter- ature, film, and other media (Baumeister, 2005), why might people endorse the Freudian theory that dream content is meaningful, believing that their dreams provide special insight?1 First, events in dreams often feature familiar people and locations, and thus can be difficult to distinguish from events that occur while awake (Johnson, Foley, Suengas, & Raye, 1988; Mazzoni & Loftus, 1996). We suggest, however, that people do not merely lend the same amount of credence to thoughts that occur in dreams as thoughts that occur while awake but actually treat the content of their dreams as more meaningful than the content of similar waking thoughts. More specifically, a decreased ability to trace dream content to an external source may lead people to give 1 We use Bering’s (2003) conception of meaning throughout this arti- cle—that people believe that the mental and physical events they consider to be meaningful have some reason or purpose, even if the reason for or purpose of the event is unknown to them. Carey K. Morewedge, Department of Social and Decision Sciences, Carnegie Mellon University; Michael I. Norton, Harvard Business School, Harvard University. We thank Paul Litvak, Benoıˆt Monin and Dan Wegner for their helpful comments, and Jennifer Bartels, Arudra Burra, Jiyhe Chong, Larissa Chopyk, Amy Cuddy, Alex Davis, Leah Feola, Bobby Jones, Reetika Khera, Matthew Killingsworth, Amit Kumar, Rebecca Levine, Sara Rabi- novich, Mindi Rock, Todd Rogers, Shimon Saphire-Bernstein, Sarah Sears, Jill Swencionis, and Alicia Warlick for their assistance in the execution of the experiments. Correspondence concerning this article should be addressed to Carey K. Morewedge, Department of Social and Decision Sciences, Carnegie Mel- lon University, 5000 Forbes Avenue, 208 Porter Hall, Pittsburgh, PA 15213. E-mail: [email protected] Journal of Personality and Social Psychology, 2009, Vol. 96, No. 2, 249–264 © 2009 American Psychological Association 0022-3514/09/$12.00 DOI: 10.1037/a0013264 249  greater weight to that seemingly random information and increase the likelihood that it will impact subsequent judgments and behav- ior. Ironically, then, although the content of dreams often appears to be produced purely by random associations (Muzur et al., 2002), which might make one expect that information to seem less mean- ingful, it may be the apparent randomness of those associations that makes people believe their dreams. This prediction is grounded in two classic lines of research in psychology: research exploring anchoring effects and research exploring attribution. Countless studies exploring numerical an- choring effects have demonstrated the remarkable tendency for random and irrelevant information to exert undue influence on subsequent judgment (Tversky & Kahneman, 1974). In standard experimenter-provided anchoring experiments, numerical esti- mates are biased in the direction of a number provided in a comparative judgment made before the estimate. For example, participants first asked whether the average price of a German car was greater or less than 20,000 Deutsche Marks (about $15,019 U.S.) for instance, they estimated the average price of a German car to be lower than did participants who first asked whether the average price of a German car was greater or less than 40,000 Deutsche Marks (about $30,039 U.S.) (Mussweiler & Strack, 2000). Because people display a general tendency to believe and seek validation for the content of their thoughts (Griffin & Ross, 1991; Klayman & Ha, 1987; Pronin, Gilovich, & Ross, 2004; Trope & Liberman, 1996), it is not surprising that, given some information, people first treat that information as valid and only subsequently attempt to adjust or correct those beliefs to reflect the actual veracity of that information (Gilbert, 1991). In some sense, the importance initially ascribed to one’s thoughts can thus be attributed to a simple counterfactual: Why would I have thought (or dreamt) this if it were not meaningful? Indeed, implausible information (e.g., “Did Gandhi live to be greater or less than 1,000,000 years old?”) exerts a similar impact (Ariely, Loewen- stein, & Prelec, 2003; Mussweiler & Strack, 2000), attesting to the fact that even one’s most outlandish thoughts can influence judg- ment in this manner. The meaning attributed to one’s thoughts does not, however, explain why thoughts that occur in dreams would be accorded more weight than thoughts that occur while awake. Certainly, unconscious thought influences judgment and behavior (Aarts & Dijkskterhuis, 1999; Bargh & Chartrand, 1999; Dijksterhuis, Char- trand, & Aarts, 2007; Greenwald, Poehlman, Uhlmann, & Banaji, in press), but it may be the manner of their generation rather than their concealed nature that lends them such importance. As Freud (1955/1910) first noted, unconscious thoughts appear powerful precisely because they seem immune to other influences to which conscious thoughts are susceptible (see Spence & Holland, 1962). We suggest that unconscious thoughts, such as dreams, exert a stronger influence on judgment than similar conscious information because they appear to be internally generated and are therefore less likely to trigger correction processes. Indeed, self-generated thoughts (i.e., those for which an external source is not evident) can exert a particularly powerful influence on judgment (Epley & Gilovich, 2001, 2006; Slamecka & Graf, 1978). A lack of aware- ness of primes, for example, can increase the magnitude of mere exposure effects (Bornstein, 1989; Bornstein & D’Agostino, 1994). Most important for our account, when people become aware that their thoughts may have been externally generated, as when primed supraliminally, they attempt to correct for those external influences (Schwarz & Bless, 1992; Wegener & Petty, 1995, 1997; Wilson & Brekke, 1994). In short, unconscious thoughts such as dreams should be more likely to influence judg- ment than conscious thoughts with similar content because of the tendency to correct for the apparent influence of external sources on the latter form of thinking. Our assertion that people are more likely to perceive dreams as internally generated than waking thoughts—a necessary precursor to according dreams more meaning—is grounded in research ex- amining the attribution of attitudes and beliefs to other actors. The fundamental distinction in the attribution literature is whether behavior is attributed either to an actor’s situation or to an actor’s disposition (Bem, 1972; Ross & Nisbett, 1991), and the extent to which observers believe a behavior reveals something meaningful about the actor is largely a function of whether that behavior can be attributed to the actor’s situation. We suggest that a similar process underlies people’s interpretations of the meaning of their conscious thoughts and dreams. When thoughts are easily attrib- uted to an external source, they are unlikely to be perceived as internally generated and consequently are considered less impor- tant and are less likely to influence the thinker’s judgment and behavior. In Schachter and Singer’s (1962) classic study, for example, participants who were not informed that their arousal was externally generated were most likely to attribute that arousal to their irritation with a confederate, whereas participants who were informed that their arousal was externally generated (and could attribute that emotion to the shot they were administered) were less likely to behave in a manner suggesting that they were irritated. It is thus the more tenuous link between the external world and unconscious thoughts that rise to the level of consciousness such as dreams than between the external world and similar conscious thoughts that make unconscious thoughts more likely to be per- ceived as internally generated and interpreted as more meaningful to the thinker. Dreams are perhaps the form of unconscious thought best suited to test this prediction, as dreams are uncon- scious thoughts that are ambiguously linked to the external world, but at the same time are unconscious thoughts to which the thinker has some access. Compare interpsychic information that appears in a dream to the same information appearing in a thought during the day: Imagine a woman who has either a waking thought or a dream of her husband being unfaithful. If the thought occurred during the day, then the thinker could easily attribute that thought to the fact that she just received an e-mail from her husband, for example, or may even have seen a coworker to whom she is sure her husband would be attracted. In either case, the waking thought could easily be attributed to an external stimulus in the immediate environment. Although the thought may still be upsetting, it is more likely that the thinker might “correct” for the external source of her suspi- cions than confront her husband. Should that thought of infidelity occur in a dream, however, the connection to the external stimuli that may have prompted it is less evident. By definition, sleep involves a decreased awareness of external events, and thus the thoughts that occur during sleep (i.e., dreams) generally lack an immediate external cause to which they may be attributed. Of course, dreams are sometimes attributed to external causes (e.g., a loud ringing in a dream may be readily attributed to the sound of one’s alarm clock ringing in the morn- ing), but their internal generation in the absence of immediate 250 MOREWEDGE AND NORTON  external stimuli makes them less likely to be attributed to external causes and more likely to be interpreted. As it is thus more difficult to attribute the suspicious thought in her dream to an external source, the wife may be less likely to correct for this suspicion and therefore be more likely to confront her husband. The increased meaning attributed to dreams compared with similar thoughts is not limited to interpsychic dreams of infidelity. Our account suggests that whenever people are less able to at- tribute some thought to an external source, that thought will be seen as more meaningful. Indeed, even a quick survey through one’s own dreams reveals the enormous variability of dream content, from mundane dreams about one’s daily life, to commu- nications with deceased loved ones and deities, to dreams about future real-world events such as one’s death. We suggest that across this wide range of both inter- and intrapsychic content, dreams are not only unlikely to be dismissed but also likely to be considered more meaningful than conscious thoughts containing similar information, and are therefore more likely to influence attitudes and behavior. MOTIVATED INTERPRETATIONS Of course, people do not entirely fail to correct for their dream content: Not everyone who dreams of a plane crash cancels his or her flight. Rather, we suggest that people are less likely to correct for the possibility of external influence when ascribing meaning to their dreams than to similar conscious thoughts, and are therefore more likely to be influenced by their dreams. What might prompt individuals to correct for the impact of the thoughts and images that arise in their dreams? Imagine the different emotions experi- enced after a dream of an intimate encounter with one’s own significant other compared with a dream of an intimate encounter with the significant other of a close friend. Both dreams should be considered more meaningful and be more influential than similar waking thoughts. Factors that impact the influence of conscious thoughts on judgment and behavior, however, should impact the influence of those dreams as well. In waking life, people exhibit a motivated interpretation of thoughts and information, ascribing more meaning to thoughts and information that accord with their existing beliefs and desires. Like conscious thoughts, undesirable or negative dreams could thus be interpreted defensively, in a manner that allows dreamers to main- tain a positive view of the self (Kruglanski, 1989; Kunda, 1990; Murray, 1938; Pyszczynski & Greenberg, 1987; Taylor & Brown, 1988). This account would predict that greater meaning would be ascribed to the dream of an intimate encounter with one’s own significant other than with the significant other of a close friend, as the latter would have more disturbing implications. Alternatively, dreamers may be motivated to interpret dreams as providing un- filtered insight into their unconscious beliefs and desires (Molden & Higgins, 2005), which would suggest that both desirable and undesirable dreams should be considered meaningful.2 When such conflicts arise, we suggest that people exhibit the former kind of motivated interpretation, viewing the ambiguous images and thoughts that arise in dreams in a manner that bolsters their diurnal beliefs and desires (Ditto & Lopez, 1992; Gilovich, 1983; Hastorf & Cantril, 1954; Lord, Ross, & Lepper, 1979; Swann, Stein-Seroussi, & Giesler, 1992). Certainly, perceivers take a motivated approach when interpreting their environment, thoughts, and behavior. Basic visual perception is affected by perceivers’ motivations, such that ambiguous images are inter- preted in a manner that foretells desirable rather than undesirable future experiences (Balcetis & Dunning, 2006). Explicit self- evaluations tend to be peculiarly charitable in ambiguous domains, in which there is ample room for favorable interpretation (Dun- ning, Meyerowitz, & Holzberg, 1989; Kruger & Dunning, 1999). Most important, just as people view tests that reflect favorably on them to be more “truthful” than tests that do not (Ditto, Munro, Apanovitch, Scepansky, & Lockhart, 2003; Ditto, Scepansky, Mu- nro, Apanovitch, & Lockhart, 1998), dreamers may consider dreams reflecting their existing beliefs and desires to be more “truthful” than dreams that do not. In short, dreaming may be believing—in that people are likely to see meaning in their dreams—but the weight accorded to particular dreams may be moderated by the extent to which dreams are in accordance with dreamers’ agendas once awake. OVERVIEW We report the results of six studies in which we examined two hypotheses related to the motivated interpretation of dreams. First, we propose that people perceive dream content to be particularly meaningful and to provide insight into their diurnal lives. In Study 1, we tested the validity of this assertion by examining laypersons’ endorsement of four prominent theories of dreaming, including the Freudian view that dreams contain hidden meaning. In Study 2, we further tested whether lay perceivers consider dreams to be more meaningful than similar conscious thoughts by comparing the impact of dreamed events with the impact of conscious thoughts on intentions to engage in a future behavior. In Study 3, we explored whether the belief that dreams are more meaningful than similar waking thoughts is restricted to individuals who endorse the Freudian view that dreams contain hidden meaning or whether the belief that dreams are more meaningful than thoughts is widely held. Second, although dreams should be seen as more meaningful than similar thoughts, we propose that perceivers take a motivated approach to the interpretation of their dreams, engaging in correc- tion processes when such correction is self-serving. In Study 4, participants recalled actual dreams involving an acquaintance. We examined the extent to which the meaning attributed to those real dreams could be predicted by the correspondence between the positivity of their dreams and the positivity of their attitudes toward those acquaintances. In Study 5, we explored the impact of dream interpretations on perceivers’ diurnal lives by examining how interpretations of a dream about a friend’s commendable or deplorable behavior impacted attitudes toward that friend. Finally, in Study 6, we examined two motivated components of dream interpretation by testing whether the meaningfulness attributed to dreams about communications from deities depended both on the (religious) beliefs of the perceiver and the desirability of the message communicated. 2 Interestingly, Freud embraced both notions: The idea that people contort reality to preserve and buttress their beliefs and opinions can be traced to Freud’s (1894/1962) taxonomy of defense mechanisms, yet he also argued that dreams provide privileged insight into the unconscious (Freud, 1900/1953). 251 THE (MOTIVATED) INTERPRETATION OF DREAMS  STUDIES 1, 2, AND 3: DREAMING IS BELIEVING To test our first hypothesis, we examined the extent to which dreams are considered meaningful sources of information in sev- eral different ways. In Study 1, we assessed how three distinct groups of participants—students from the United States, South Korea, and India—endorsed four prominent theories of dreaming. We predicted that participants from all three cultures would be more likely to endorse a theory suggesting that dream content has meaning than theories suggesting otherwise. Participants in Study 2 imagined that a frightening event (a plane crash) either occurred in a dream, occurred in a diurnal conscious thought, actually occurred, or was deemed likely to occur by federal authorities. We predicted that participants would report being more affected by the dream than by the conscious thought and used the meaningfulness attributed to the other sources of information as benchmarks by which to further assess the meaningfulness attributed to the dream. Finally, in Study 3, we explored whether the meaning attributed to dream content varies as a function of the theory of dreams one endorses. We expected that all participants would consider a dream more meaningful than a similar conscious thought, whereas the extent to which participants considered the dream more meaning- ful than other sources of information would vary in accordance with the theory of dreaming that they endorsed. Study 1 Participants rated the extent to which they endorsed four prom- inent theories of dreaming. We expected participants to consider dreams meaningful and rate the Freudian view of dreams—that dreams reveal hidden truths about the self—more highly than a theory that dream content is a by-product of unrelated brain activity as well as two other theories suggesting that the function of dreams is important but that dream content results from the filtration of external stimuli. Most important, we attempted to assess the universality of the belief that dream content is mean- ingful by surveying populations from three cultures: the United States, South Korea, and India (Markus & Kitayama, 1991; Nis- bett, Peng, Choi, & Norenzayan, 2001). Method Participants United States student sample. Fifty undergraduate and gradu- ate students at the Massachusetts Institute of Technology (MIT) in Cambridge, Massachusetts (31 women; Mage � 27.3, SD � 12.0) participated in a short survey in a campus student center in ex- change for candy. South Korean student sample. Fifty-seven students in a psy- chology class at Korea University in Seoul, Korea (22 women; Mage � 21.8, SD � 2.5) completed the survey in a packet of unrelated surveys as part of a requirement for an introductory psychology course. The survey was translated into Korean by a Korean research assistant unaware of the hypothesis and checked for accuracy by an independent translator. Indian student sample. Forty-two undergraduate students at the Delhi School of Economics and master’s students at Jawa- harlal Nehru University in New Delhi, India (9 women; Mage � 21.6, SD � 3.2) participated in a short online survey in ex- change for participation in a lottery. The survey was adminis- tered in English. Procedure In a within-subjects design, each participant assessed four prom- inent theories of dreams, described in nontechnical language: Dreams provide “useful insights into how to solve problems” (Cartwright, 1974; Cavallero & Foulkes, 1993; Wagner, Gais, Haider, Verleger, & Born, 2004); dreams are a by-product of unrelated brain activity that occur “when the brain tries to interpret random impulses from the pons as sensory input, producing vivid hallucinations” (Hobson & McCarley, 1977; Muzur et al., 2002); dreams reveal hidden truths “when emotions buried in the uncon- scious surface in disguised form” (Freud 1953/1900; Wegner et al., 2004); and dreams assist learning as they “throw out unwanted information to prevent information from becoming jumbled” (Crick & Mitchison, 1983; Maquet, 2001; Revonsuo, 2000; Stick- gold, 2005).3 The exact wording of each theory appears in the Appendix; the theories were not labeled. Participants reported the extent to which they agreed with each theory on four identical 7-point scales ranging from 1 (do not agree at all) to 7 (agree completely). Finally, participants circled the theory they consid- ered most true. Results United States Student Sample As expected, participants were more likely to endorse the Freud- ian theory that dreams reveal hidden truths (M � 4.84, SD � 1.70) than the problem-solving (M � 3.40, SD � 1.81), by-product (M � 3.98, SD � 1.56), or learning theories (M � 3.78, SD � 1.93); support for the Freudian theory was significantly greater than for each of the other theories, all ts(49) � 2.54, all ps � .01. In addition, the majority of participants (56%) selected the Freud- ian theory as most true, far more than any of the other options: problem solving (8%), by-product (18%), or learning (18%), �2(3, N � 50) � 26.96, p � .001 (see Figure 1). South Korean Student Sample As before, the Freudian theory that dreams reveal hidden truths (M � 5.28, SD � 1.07) was endorsed more highly than each of the other theories: problem solving (M � 3.53, SD � 1.44), by- product (M � 4.60, SD � 1.39), or learning theories (M � 3.05, SD � 1.22), all ts(56) � 3.28, all ps � .001. In this sample, nearly two thirds of participants (64.9%) thought the Freudian theory most true, far more than any of the other options: problem solving (3.5%), by-product (29.8%), or learning (1.8%), �2(3, N � 57) � 59.70, p � .001 (see Figure 1). Indian Student Sample Once again, the Freudian theory that dreams reveal hidden truths (M � 5.29, SD � 1.60) was endorsed more highly than each of the 3 Italics in quotes were added for emphasis. 252 MOREWEDGE AND NORTON  other theories: problem solving (M � 3.33, SD � 1.72), by- product (M � 3.90, SD � 1.54), or learning theories (M � 2.90, SD � 1.41), all ts(41) � 3.78, all ps � .001. In this sample, nearly three quarters of participants (73.8%) thought the Freudian theory most true, far more than any of the other options: problem solving (11.9%), by-product (9.5%), or learning (4.8%), �2(3, N � 42) � 53.81, p � .001 (see Figure 1). Discussion As predicted, the theory of dreams that emphasized their mean- ing (i.e., the Freudian view that dreams contain hidden truths) was most strongly endorsed, whether assessed with scale ratings or a forced-choice measure. Participants were thus more likely to en- dorse a theory of dreams suggesting that dreams reveal meaningful internally generated information that bubbles to the surface of consciousness than two theories suggesting that dreams’ function is meaningful, but dream content merely reflects the day’s events (the learning and problem-solving theories), and a theory suggest- ing that neither dreams’ function nor content are meaningful (the by-product theory). These results held across three different cul- tures, in samples of college students in South Korea, India, and the United States. Although researchers still debate the function of dreams and dream content’s meaning, laypeople around the world appear to believe that dreams serve an important function and have meaning, revealing hidden truths. Study 2 Participants in our first study were most likely to endorse a theory of dreams suggesting that dreams contain meaningful in- formation. In our next study, we examined the extent to which people attribute meaning to the information revealed in dreams. In an ancillary survey of commuters in Boston, Massachussets (N � 81; 46 women; Mage � 24.0, SD � 8.1), the majority of partici- pants reported that dreams impacted their everyday behavior— influencing both their social relations (67%) and decision making (52%)—perhaps due to their reported belief that dreams foretell the future (68%) and that at least one of their dreams had come true (63%). These results lend further support to our assertion that lay perceivers consider dreams to provide important insight into their diurnal lives. We sought to establish in Study 2 the degree of importance lay perceivers grant to dreams by comparing the impact of dreamed events with the impact of both imagined and real-world events on their proclivity to engage in a behavior. As outlined earlier, we suggest that dreams should be ascribed more importance than similar conscious thoughts. Therefore, we expected participants to report being more affected by events occurring in dreams than by the same event occurring in a waking thought. To benchmark the importance of dreams in comparison to other forms of information, we compared the impact of dreams with the impact of similar real-world events. Method Participants One hundred eighty-two commuters at South Station in Boston (95 women; Mage � 35.9, SD � 16.0) volunteered to participate. Procedure In a between-subjects design, participants completing a survey on “air traffic safety” were asked to imagine one of four scenarios: that the night before one of their scheduled airline trips either (a) the United States Department of Homeland Security (n.d.) issued a warning by raising the national threat level to “Orange,” indi- cating a high risk of terrorist attack; (b) they thought consciously about their plane crashing on the flight they planned to take; (c) they dreamt about a plane crash on the flight they planned to take; or that (d) a real plane crash occurred on the route they planned to take. Participants then reported how anxious they would feel if they were scheduled to fly that day and how likely they would be to avoid flying on 5-point scales ranging from 0 (not at all anxious/likely) to 4 (extremely anxious/likely). A composite mea- sure of these two items was created, as they were highly correlated, r(180) � .69, p � .001. Results As expected, participants were differently impacted by the type of information they considered, F(3, 178) � 4.15, p � .007, �p 2 � .07. Participants were more likely to report that a dream of a plane crash would affect their travel plans than a conscious thought of a crash or a warning from the government, F(1, 178) � 9.16, p � .003, r � .32; and, F(1, 178) � 3.54, p � .05, r � .19, respec- tively. Even an actual plane crash did not exceed a dream of a Figure 1. Participants in the United States, South Korea, and India were more likely to prefer the Freudian theory of dream content than three other prominent theories of dream content (Study 1). 253 THE (MOTIVATED) INTERPRETATION OF DREAMS  plane crash in its impact on the likelihood that they would engage in air travel (F � 1; see Figure 2).4 Discussion Dreams appear to be potent sources of information. Participants considered a dream of a plane crash to be more unsettling than an identical waking thought and a federal warning indicating a “high risk of a terrorist attack,” and as unsettling as an actual crash. While these reports reflect the impact of information on feelings about flying rather than actual behavior, results from our ancillary survey—in which the majority of participants reported that dreams influence their everyday lives—suggest a general willingness to heed advice distilled from dream content. In short, the results of the first two studies suggest that dreamed events, even when unpleasant, are perceived to be meaningful sources of information, to be more meaningful than similar conscious thoughts, and can even be perceived to provide information as important as similar real-world events. Study 3 Endorsement of the Freudian theory of dreams appears most prevalent across cultures, yet a portion of each of the populations surveyed endorsed one of the other three prominent theories of dreaming (Study 1). It is thus important to assess whether the meaningfulness attributed to information appearing in dreams in Study 2 reflected only the opinions of those endorsing the Freudian view or whether dreams are generally considered meaningful sources of information. In Study 3, we assessed whether belief in the meaningfulness of information appearing in dreams varied with regard to the theory of dreams that participants endorsed. Our account predicts a greater impact of dreams than similar thoughts on judgment and behavior, irrespective of the theory of dreams that people believe most true. As those theories may influence the extent to which dreams are considered important sources of information, however, we also expected participants who believe dream content is generated for external reasons or for no reason (i.e., reflecting one’s current problems, distilling the day’s events, or completely random) to consider dreams less meaningful than participants who believe dreams are generated for internal reasons (i.e., revealing hidden truths about the self). Method Participants Three hundred forty-one pedestrians in Cambridge, Massachus- sets (205 women; Mage � 26.6, SD � 13.0) participated in ex- change for candy. Procedure In a within-subjects design, participants first rated four theories of dreams (see the Appendix) on scales identical to those described in Study 1 and reported which theory was their favorite and then imagined each of the following three scenarios: that the night before one of their scheduled airline trips, either (a) they con- sciously thought about their plane crashing on the flight they planned to take; (b) they dreamt about a plane crash on a flight they planned to take; or that (c) a real plane crash occurred on the route they planned to take. Participants then ranked the events, from the event (1) most likely to make them cancel or miss their flight to the event (3) least likely to make them cancel or miss their flight. Results As in Study 1, the Freudian view theory that dreams reveal hidden truths (M � 5.17, SD � 1.60) was endorsed more highly by participants than each of the other theories: problem solving (M � 3.99, SD � 1.75), by-product (M � 4.26, SD � 1.57), or learning theories (M � 3.58, SD � 1.79), all ts(340) � 8.03, all ps � .001, all rs � .40. It was also most often selected as the favorite theory (49.3%), more often than the problem-solving (12.3%), by-product (22.0%), or learning theories (16.4%), �2(3, N � 341) � 113.53, p � .001. To test whether beliefs about the source of dream content influences the meaningfulness ascribed to information distilled from dreams, we divided participants into two groups—those who most preferred the Freudian theory of dream content (i.e., Freud- ians) and those who most preferred one of the three alternative theories (i.e., non-Freudians). First, we performed Friedman chi- square tests within each group of participants to test whether the 4 Considered separately, participants’ reports of the anxiety they would feel and how likely they would avoid flying were similarly influenced by the type of information they considered, F(3, 178) � 4.38, p � .005, �p 2 � .07; and, F(3, 178) � 2.81, p � .04, �p 2 � .05. Reported feelings and behavior were influenced to a greater extent by a dream of a plane crash (Manxiety � 2.16, SD � 1.38; Mavoid flying � 1.25, SD � 1.56) than by the conscious thought of a crash (Manxiety � 1.26, SD � 1.20; Mavoid flying � 0.56, SD � 0.93), Fs(1, 178) � 6.19, ps � .01, rs � .26, respectively. Of interest is that dreams made participants more anxious but no more likely to avoid flying than a warning from the government (Manxiety � 1.48, SD � 1.37; Mavoid flying � .94, SD � 1.30), Fanxiety(1, 178) � 5.77, p � .02, r � .24; and, Favoid flying(1, 178) � 1.25, p � .27. Even an actual plane crash did not exceed a dream of a plane crash in its impact on their feelings or behavior (Manxiety � 2.00, SD � 1.46; Mavoid flying � 1.28, SD � 1.44) (F � 1). Figure 2. Participants were more likely to report that a dream of a plane crash would affect their travel plans than a conscious thought of a crash or a warning from the federal government. Even an actual plane crash did not exceed a dream of a plane crash in its impact on the reported likelihood that they would engage in air travel (Study 2). Bars represent � 1 standard error. 254 MOREWEDGE AND NORTON  three kinds of information (i.e., an actual plane crash, a dream of a plane crash, and a thought of a plane crash) were attributed different degrees of importance. Those tests revealed significant differences between the importance attributed to the three kinds of information within both groups of participants, �Freudians 2 (2, N � 168) � 55.84, p � .001; and, �non-Freudians 2 (2, N � 173) � 70.18, p � .001. Next, we compared the rank orderings made by the two groups using Schucany and Frawley’s (1973) statistic (as sug- gested by Dekle, Leung, & Zhu, 2008), which revealed that their rank orderings were not concordant (Z � �.88, p � .48). As the two groups’ rank orderings were discordant, they were examined separately. Both participants endorsing the Freudian theory and participants endorsing the alternative theories considered dreams to be more meaningful than similar conscious thoughts (ZWilcoxon Signed Ranks � 7.35, p � .001 and ZWilcoxon Signed Ranks � 6.81, p � .001, respec- tively). As expected, however, the theory of dreaming participants endorsed affected the relative importance they attributed to dreams. Participants who did not endorse the Freudian theory reported that an actual plane crash would be more influential than a dream of a plane crash (ZWilcoxon Signed Ranks � 1.96, p � .05), whereas participants who endorsed the Freudian theory reported that a dream of a plane crash would be marginally more influential than an actual plane crash (ZWilcoxon Signed Ranks � 1.90, p � .06;.see Figure 3). Discussion Irrespective of the theory of dreaming that they endorsed, par- ticipants considered dreams to provide meaningful information. Participants endorsing Freudian and non-Freudian theories of dreams reported that a dream of their plane crashing would make them more likely to avoid a future flight than a similar conscious thought. Remarkably, participants who endorsed the Freudian the- ory considered information distilled from dreams to be even more influential than similar real-world events, whereas participants endorsing non-Freudian theories of dreaming considered informa- tion distilled from dreams to be only slightly less important than similar real-world events. Regardless of the theory of dreams that they endorsed, participants thus considered dreams to be more important than similar thoughts occurring to them while awake and almost as important as or more important than the real-world event that the dream reflected. These findings suggest not only that dreams are considered to be more meaningful sources of informa- tion than similar conscious thoughts but also that the meaning accorded to dreams is influenced by the theory of dreams that the dreamer endorses. Although all participants considered a dream of a plane crash to be meaningful, participants endorsing the Freudian theory of dreams were more affected by that dream than partici- pants endorsing other theories. STUDIES 4, 5, AND 6: THE (MOTIVATED) INTERPRETATION OF DREAMS The first three studies elucidate two important and interrelated aspects of laypersons’ beliefs about dream content: Most people believe that dreams reveal meaningful information about them- selves and their world and that their dreams hold more meaning than similar waking thoughts. Using a plane crash to test the latter point constitutes a strong test of the importance of dreams: If people have any desire to dismiss their dreams as meaningless, then they should be most likely to do so with negative dreams, as such dreams provide undesirable information about one’s self and world. Indeed, it would be much more pleasant to simply consider that dream meaningless and fly without the lingering image of a plane crash in the back of one’s mind. Although dreams of this negative event were considered meaningful, many participants were not willing to accord a dream of a plane crash the same significance as an actual plane crash, suggesting that people do engage in some correction when determining the importance of negative dreams. Subsequent studies explored whether participants would be less motivated to engage in correction when dreams contained information confirming their beliefs and desires than when dreams contained disconfirming or negative information. Having found support in Studies 1, 2, and 3 for our first hypothesis—that people consider dream content to be meaningful and provide important insight into their waking lives—we tested our second hypothesis in Studies 4, 5, and 6: That people do not consider all information in dreams to be equally important, but rather that their dream interpretations are influenced by the extent to which events in dreams accord with their views of reality when awake. Just as more effort is exerted when critiquing undesirable real-world information (Ditto et al., 2003, 1998; Klein & Kunda, 1992), we predicted that people would take a motivated approach when interpreting dream content—ascribing greater meaning to dreams that matched their preexisting beliefs and desires. Initial evidence for a motivated interpretation of dreams is present in Study 3. Although participants were unable to entirely dismiss an unsettling dream of a plane crash (or dismiss that dream as easily as a similar conscious thought), the dream theory participants Figure 3. Participants who preferred the Freudian theory reported that a dream of their plane crashing would be more likely to induce them to cancel a flight than an actual plane crash on their scheduled route, whereas participants who preferred alternative theories of dream content reported that a dream of their plane crashing would be less likely to induce them to cancel a flight than an actual plane crash. Both groups considered the dream to be more influential than a similar conscious thought (Study 3). Bars represent � 1 standard error. 255 THE (MOTIVATED) INTERPRETATION OF DREAMS  endorsed influenced the extent to which they dismissed it. Partic- ipants in Study 3 who endorsed the Freudian theory of dreams were relatively less likely to dismiss the dream than participants endorsing one of the other three theories. These results suggest not only that dreams are considered important sources of information but also that the extent to which a dream is ascribed meaning may vary according to the beliefs of the dreamer. Thus, Studies 4–6 were designed to examine in greater depth whether dreams are selectively interpreted. This possibility was first tested in the domain of social relation- ships, as emotional dreams about friends and loved ones are particularly frequent (McNamara, McLaren, Smith, Brown, & Stickgold, 2005). In Study 4, participants recalled a real dream that involved another person. We assessed the extent to which the correspondence between the positivity of dream content and atti- tudes toward those people influenced the meaning participants attributed to their dreams. We expected that participants who reported dreams involving a liked person would evaluate dreams containing positive content to be more meaningful than dreams containing negative content, whereas participants who reported dreams involving a disliked person would rate dreams containing positive content to be less meaningful than dreams containing negative content. In Study 5, we held the positivity toward the person who appeared in dreams constant by asking all participants to imagine they dreamt about a randomly selected friend but varied the positivity of the dream imagined. In addition to assessing the meaning ascribed to those dreams, we also assessed whether the meaning ascribed to those dreams impacted the strength of their friendship. Finally, we examined the influence of important general beliefs and the desirability of dream content on dream interpretation by exploring how religious believers and skeptics differed in their interpretations of dreams containing divine communications. In Study 6, participants assessed the meaningfulness of a dream in which God commanded them to engage in hedonistic or self- abnegating behavior. We predicted that religious believers would be motivated to consider all such otherworldly dreams as mean- ingful irrespective of the desirability of the commandment, whereas skeptics might be motivated to consider otherworldly dreams meaningful only when the commandment reflected their worldly desires. Study 4 Method Participants A representative national sample of two hundred seventy Amer- icans (144 men; Mage � 44.2, SD � 14.1) completed a short Internet survey. Participants were randomly selected from a data- base of 2.5 million respondents with parameters ensuring that the sample reflected the gender, age, education level, and income distribution reported by the most recent (2000) U.S. Census. The sample included citizens from all 50 states and the District of Columbia. Procedure After completing an unrelated survey, participants were asked to think of a dream about a person they knew that they could clearly recall and to briefly describe its content. Participants then rated the extent to which they considered that dream to be meaningful, to provide insight into their relationship with that person, and to be pleasant on 5-point scales where 1 � not at all, slightly, 2 � slightly, 3 � somewhat, 4 � moderately, and 5 � extremely. Participants then reported the extent to which they liked or disliked the person who appeared in their dream on a 7-point scale ranging from 1 (dislike extremely) to 7 (like extremely). Results We predicted that participants would exhibit a motivated inter- pretation of their dreams, ascribing more importance to pleasant dreams about liked individuals than disliked individuals, and as- cribing more importance to unpleasant dreams about disliked individuals than liked individuals. As the meaningfulness and insight ascribed to dreams was highly correlated, r(268) � .81, p � .001, those reports were averaged into a single measure of the importance attributed to the dream. We then analyzed participants’ reports of importance with re- gression by using centered values for the predictors (liking for the person in the dream and the pleasantness of the dream), which yielded a significant linear fit (R2 � .13), F(3, 266) � 13.14, p � .001. Importance ascribed to dreams varied according to the extent to which participants liked the person who appeared in the dream (� � .24), t(268) � 3.52, p � .001, such that dreams about liked individuals were seen as more meaningful than dreams about disliked individuals, but overall importance did not vary according to the pleasantness of the dream (� � .01), t(268) � .10, p � .92. More important, the model revealed that the importance ascribed to dreams varied according to the predicted interaction of liking and pleasantness (� � .28), t(268) � 2.52, p � .01. To clarify the nature of the interaction, we calculated the simple slopes for dreams with pleasant and unpleasant content (i.e., one standard deviation above and below the mean of pleasantness). As illus- trated by Figure 4, greater meaning was ascribed to pleasant dreams about liked individuals than disliked individuals (� � .33), Figure 4. Participants ascribed more importance to pleasant than to unpleasant dreams that they recalled about liked individuals, and more importance to unpleasant than to pleasant dreams that they recalled about disliked individuals (Study 4). 256 MOREWEDGE AND NORTON  t(267) � 7.09, p � .001, whereas equal meaning was ascribed to unpleasant dreams about liked individuals and disliked individuals (� � .04), t(267) � .29, ns. These results suggest a motivated interpretation of dreams: Participants ascribed more importance to pleasant dreams about liked individuals than to unpleasant dreams about liked individuals but relatively more importance to unpleas- ant dreams about disliked individuals than to pleasant dreams about disliked individuals. Posttest Although participants appeared to interpret the dreams they recalled in accordance with their beliefs about the individuals who appeared in them, it is possible that participants instead recalled dreams in accordance with their beliefs about the individual that first came to mind (Conway & Ross, 1984; Sanitioso, Kunda, & Fong, 1990). To address the possibility that rather than exhibiting a motivated interpretation of dreams, participants exhibited a mo- tivated recollection of dreams, we conducted an ancillary posttest in which participants were asked to recall the most recent dream they could clearly remember, indicate when that dream occurred, and rate the valence of that dream. We tested for the presence of motivated recollection in two ways. First, if people routinely engage in a motivated recollection of dreams, then it could be expected that the overall valence of the dreams they remembered would be positive, rather than neutral or negative. Second, if motivated recollection unfolds over time, then negative dreams should be less likely to be remembered than positive dreams as time passes, suggesting that the valence of dreams and the amount of time between their recollection and occurrence should be pos- itively correlated. We conducted the posttest among a representative national sample of 309 Americans (174 women; Mage � 42.6, SD � 17.6) who were randomly drawn from the database described in Study 4. Participants were first asked whether they could clearly recall a dream from the previous evening. Participants who could recall a dream from the previous evening were then asked to describe that dream. Participants who could not recall a dream from the previous evening were asked to report the date of the most recent dream they could clearly recall and then describe that dream. Each participant then rated the pleasantness of their dream on a 7-point scale ranging from 1 (extremely unpleasant) to 7 (extremely pleasant). We first tested for the overall valence of the dreams people recalled to see whether the valence of the dreams participants recalled was skewed positive. The mean pleasantness rating of the dreams participants reported (M � 4.05, SD � 1.63) was remark- a...