Download - Thorisdottir & Jost (2012). Political Psychology

Motivated Closed-Mindedness Mediates the Effect ofThreat on Political Conservatismpops_840 785..812

Hulda ThórisdóttirUniversity of Iceland

John T. JostNew York University

In this article we synthesize theory and research from several areas of psychology andpolitical science to propose and test a causal model of the effects of threat on politicalattitudes. Based in part on prior research showing that fear, threat, and anxiety decreasecognitive capacity and motivation, we hypothesize that under high (vs. low) threat, peoplewill seek to curtail open-ended information searches and exhibit motivated closed-mindedness (one aspect of the need for cognitive closure). The subjective desire for cer-tainty, control, and closure, in turn, is expected to increase the individual’s affinity forpolitical conservatism, insofar as resistance to change and adherence to authority figuresand conventional forms of morality are assumed to satisfy these epistemic motives moresuccessfully than their ideological opposites. Consistent with this account, we find inStudies 1a and 1b that putting people into a highly threatened mindset leads them to exhibitan increase in motivated closed-mindedness and to perceive the world as more dangerous.Furthermore, in Study 2 we demonstrate that a subtle threat manipulation increases self-reported conservatism (or decreases self-reported liberalism), and this effect is mediated byclosed-mindedness. In Study 3, we manipulated closed-mindedness directly and found thathigh (vs. low) cognitive load results in a greater affinity for the Republican (vs. Democratic)party. Finally, in Study 4 we conducted an experiment involving political elites in Icelandand found that three different types of threat (to the self, group, and system) all ledcenter-right politicians to score higher on closed-mindedness and issue-based politicalconservatism. Implications for society and for the theory of ideology as motivated socialcognition are discussed.

KEY WORDS: cognitive load, closed-mindedness, fear, need for closure, political ideology, politicalconservatism, threat

Political Psychology, Vol. 32, No. 5, 2011doi: 10.1111/j.1467-9221.2011.00840.x

785

0162-895X © 2011 International Society of Political PsychologyPublished by Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148, USA, 9600 Garsington Road, Oxford, OX4 2DQ,

and PO Box 378 Carlton South, 3053 Victoria Australia

“Alas,” said the mouse, “the whole world is growing narrower every day.At the beginning it was so big that I was afraid, I kept running and running,and I was glad when I saw walls far away to the right and left, but theselong walls have narrowed so quickly that I am in the last chamber already,and there in the corner stands the trap that I must run into.”

“You only need to change your direction,” said the cat, and ate it up.

Franz Kafka, “A Little Fable”

Shrewd and opportunistic politicians, like Kafka’s cat, have long recognizedthe power of fear and threat to influence their constituency (e.g., Bennett, 1995;Mack, 2004; Robin, 2004). Machiavelli advised his prince in the sixteenth centurythat it was “better to be feared than loved” (Machiavelli, 1999/1513). Deep into theGreat Depression in 1933, Franklin D. Roosevelt famously declared to the Ameri-can public that “we have nothing to fear but fear itself,” which he described as a“nameless, unreasoning, unjustified terror which paralyzes needed efforts toconvert retreat into advance.” The program of research on which we report hereaims to elucidate the often noted but seldom understood psychological process bywhich threat exerts its influence on political attitudes.

Threat and Political Attitudes

Research in the social and behavioral sciences has consistently found thatthreat is associated with certain manifestations of political conservatism (e.g., Jost,Glaser, Kruglanski, & Sulloway, 2003). In this prior research, the operationaliza-tion of both constructs has varied widely. Threat has generally been operational-ized in terms of threatening societal periods, threats to the legitimacy or stabilityof the social system, the salience of terrorism, rising immigration, and a host ofother threatening stimuli, including thoughts about one’s mortality. Conservatismhas been gauged in terms of self-reported political orientation, voting behavior,support for right-wing authoritarianism, and endorsement of various issue posi-tions that are typically associated with conservatism or right-wing politics (for areview see Jost, Federico, & Napier, 2009).

Authoritarianism

Authoritarianism was perhaps the first ideological construct to be linkedtheoretically and empirically to a personal sense of threat (Adorno, Frenkel-Brunswick, Levinson, & Sanford, 1950).1 Fromm (1941) theorized that people

1 Authoritarianism is a syndrome characterized by overemphasis on submission and identification withstrong leaders, rigid conformity to conventional norms and rejection of those who violate them,cognitive rigidity, ethnocentric prejudice, right-wing orientation, and rejection of the subjective andtender-minded (Adorno et al., 1950; Altemeyer, 1998; Napier & Jost, 2008).

786 Thórisdóttir and Jost

want to “escape from freedom” and embrace authoritarianism when confrontedwith a threatening, uncertain world. Consistent with this theoretical tradition,Lavine and colleagues showed almost 60 years later that people who score high onauthoritarianism tend to be more sensitive to threatening words and messages incomparison with low scorers (Lavine, Lodge, Polichak, & Taber, 2002). Archivalstudies have also supported the link between fear and authoritarianism by dem-onstrating that authoritarianism in the general population increases under high (vs.low) societal threat periods (e.g., periods of unemployment, crime, and war; Doty,Peterson, & Winter, 1991; Sales, 1973; Stenner, 2005).

Several studies conducted in the aftermath of 9/11 indicate that the morethreatened or concerned people are about their security following terrorist acts,the less supportive they are of civil liberties (Cohrs, Kielmann, Maes, &Moschner, 2005; Davis & Silver, 2004; Huddy, Khatid, & Capelos, 2002).Duckitt and Fisher (2003) found that presenting participants with a threateningscenario (in which their country’s crime rate, economy, and general prosperitywere depicted as worsening in the future) increases their affinity for authoritari-anism. Studies on individual differences reveal a strong correlation between thetendency to perceive the world as dangerous and right-wing authoritarianism, asmeasured with the RWA scale (e.g., Altemeyer, 1998). Thus, an empirical con-nection exists between threat—whether measured or manipulated as a state ortrait variable—and authoritarianism.

Support for Politically Conservative Leaders and Opinions

Research on the effects of threat in the political realm has extended far beyondauthoritarianism to a variety of politically relevant attitudes. The 9/11 terroristattacks in New York City and the 3/11 attacks in Madrid launched a new wave ofstudies on the effect of fear on political attitudes. Willer (2004) showed, forinstance, that whenever the department of Homeland Security raised the nationalterror alert level, President Bush’s approval ratings rose. Bonanno and Jost (2006)found that high-exposure survivors of the attack on the World Trade Centerbuildings became more conservative and more authoritarian in the 18 monthsfollowing 9/11, regardless of political party affiliation or previous voting decisions(see also Echebarria & Fernández, 2006; Nail & McGregor, 2009).

Experiments that evoke fear by asking people to contemplate their ownmortality have shown that death anxiety increases support for conservative opin-ions and leaders, including President George W. Bush (e.g., Cohen, Ogilvie,Solomon, Greenberg, & Pyszczynski, 2005; Gailliot, Schmeichel, & Baumeister,2006; Landau et al., 2004b). Jost, Fitzsimons, and Kay (2004) demonstrated thatmortality salience also increases endorsement of politically conservative opinionswith respect to several issues, including taxation, capital punishment, and stem cellresearch (see also Nail & McGregor, 2009). The basic finding that threateningsituations increase one’s affinity for politically conservative opinions has been

787Threat, Closed-Mindedness, and Conservatism

replicated using several different experimental paradigms and is probably notlimited to manipulations of mortality salience (e.g., see Eibach & Libby, 2009;Lambert et al., 2010; Nail, McGregor, Drinkwater, Steele, & Thompson, 2009).

Political Conservatism as Motivated Social Cognition

In 2003, Jost and colleagues surveyed a wide range of studies on the psy-chological predictors of political attitudes published between the years 1958 to2002 (including some studies mentioned above). Based on nearly 90 studies,they concluded that political ideology should be understood as reflecting moti-vated social cognition and that people who embrace conservative, right-wingideologies do so in part because it serves their (chronically or temporarilyheightened) needs to manage uncertainty and threat. Subsequent research hassupported this general model of politically conservative attitudes (e.g., seeBonanno & Jost, 2006; Carney, Jost, Gosling, & Potter, 2008; Jost et al., 2004;Nail & McGregor, 2009; Nail et al., 2009; Thórisdóttir, Jost, Liviatan, & Shrout,2007). For instance, Jost et al. (2007) conducted three studies in which theyfound that psychological needs to reduce uncertainty (e.g., need for order, intol-erance of ambiguity, and lack of openness to experience) and threat (e.g., deathanxiety, system threat, and perceptions of a dangerous world) each contributedindependently to self-reported political conservatism (vs. liberalism), even afteradjusting for the effects of ideological extremity. No support was obtained forthe alternative hypothesis suggested by Greenberg and Jonas (2003) that uncer-tainty and threat management would be associated with extreme liberalism aswell as extreme conservatism.

Recent work suggests that differences between liberals and conservatives interms of orientations toward uncertainty and threat may even be instantiated at thelevel of neurophysiological functioning. Specifically, Amodio, Jost, Master, andYee (2007) found that politically liberal participants performed better than con-servatives on a computer task requiring cognitive flexibility, conflict monitoring,and the ability to override habitual response patterns. During a “go/no-go” task,liberals also showed stronger neural activity in the anterior cingulate cortex (asmeasured in terms of event-related potentials), indicating greater neurocognitivesensitivity to potentially conflicting pieces of information. With regard to threatsensitivity, Oxley and colleagues (2008) observed that participants who stronglyendorsed conservative attitudes on a range of issues (e.g., military spending,capital punishment, gun control, foreign aid, immigration, etc.) exhibited strongerstartle blink responses and galvanic skin response conductivity (both classicphysiological measures of threat sensitivity), as compared with participants whoendorsed liberal attitudes on these issues. Thus, studies indicate that individuals’orientations concerning threat and uncertainty, whether measured behaviorally orphysiologically, are correlated with political orientation. One of the major goalsof the current research program is to more fully explicate the nature of the


relationship between these two distinct but related motives of threat and uncer-tainty management, that is, the connection between existential and epistemicmotives in producing ideological outcomes.

There are at least two important characteristics of political conservatism as anideology that may help to explain its special appeal under circumstances of threatand uncertainty (see also Jost, 2006; Jost et al., 2003, 2007, 2009). One of these,which can be traced back to the writings of Edmund Burke (1790/1999) is overtresistance to the notion that societal arrangements and systems should be delib-erately changed according to contemporary ideals (Kirk, 1982; Lipset & Raab,1978; Minogue, 1967; Muller, 2001). The status quo, no matter how aversive, is aknown condition and is therefore easier to predict and imagine than a potentiallydifferent state of affairs that could be either better or worse. As Kirk (1982) put it,conservatives “prefer the devil they know, to the devil they don’t know” (p. xv).Thus, status quo conservatism offers more familiarity and certainty than do alter-native ideologies that are characterized by preferences for social change andinnovation; for this reason, it is likely to be more accessible and attractive whenpeople are seeking stability and security.

Political conservatism is also characterized by deference to authorities, bothreligious and political (e.g., Burke, 1790/1999; Haidt & Graham, 2007). This mayhelp to explain why conservatives are more enthusiastic about hierarchy and lessenthusiastic about social and economic equality, in comparison with liberals (e.g.,Bobbio, 1996; Jost et al., 2003; Laponce, 1981; Sidanius & Pratto, 1999). Aligningoneself with conventional traditions and deferring to authority figures takes lessmental effort than conceiving of alternative social arrangements and is generallyless cognitively taxing and therefore more beneficial when the motivation forcognition is reduced than when it is not. Thus, to the extent that threat reduces thedesire for open-ended information search and processing, it should increase thepsychological affinity for political conservatism.

Psychological Responses to Threat

Research has yet to specify precisely the psychological processes by whichthreat may affect ideological preferences (but see Lambert et al., 2010 for severalstudies highlighting the role of anger). When process explanations are fully devel-oped, they will need to take into account the physiology of the threat response aswell as downstream effects on cognition, motivation, and political preferences.The literature in political psychology has rarely been very exact in its conceptualor operational definitions of threat, at least compared to cognitive neuroscienceand related fields (Spezio & Adolphs, 2007). This difference can be partiallyattributed to the fact that the two subdisciplines emphasize quite different levels ofexplanation or analysis. In this article we focus on the cognitive, motivational, andbehavioral consequences of threat.


Effects of Threat on Cognitive Capacity and Motivation

Threat and its typical emotional consequences, namely fear and anxiety, havebeen found to result in a number of cognitive, motivational, and behavioral impli-cations, including risk aversion, attentional biases, and impaired performance ona variety of working memory and decision-making tasks (e.g., Jameson, Hinson,& Whitney, 2004; Mogg, Mathews, Bird, & Macgregor-Morris, 1990; Preston,Buchanan, Stansfield, & Bechara, 2007). For one thing, rumination and worryconstrain the availability of mental resources required for memory and cognition(Eysenck & Calvo, 1992; Mathews & Mackintosh, 1998; Miu, Heilman, &Houser, 2008; Preston et al., 2007). Fear also leads to narrow, selective attentionfocused on threatening stimuli (Easterbrook, 1959). This can improve perfor-mance on a task such as the Stroop task in which there is a good match betweenthe narrowed attention and the optimal response (Chajut & Algom, 2003).

Under many circumstances, however, such “cognitive narrowing” will resultin suboptimal performance. For instance, Keinan (1987) found that when facedwith the threat of electric shock, people engage in less systematic scanning ofalternatives on a multiple choice analogy task. This results in hasty, inferiordecisions made under high versus low threat. Threatening research participants bycommunicating (false) negative performance feedback impairs their digit spanmemory (Hodges & Spielberger, 1969) and reduces message elaboration in per-suasion (Sengupta & Johar, 2001). Hillier, Alexander, and Beversdorf (2006)showed that an auditory stressor (90 dB white noise) reduced cognitive flexibilityon a word association task.

Thus, it seems likely that threat—like cognitive load (Ford & Kruglanski,1995; Gilbert & Osborne, 1989)—reduces both the motivation and capacity forinformation processing. In many situations, the two effects may be inseparable.Studies show that fatigue, time pressure, and exposure to loud noises all decreasethe motivation to engage in extensive information searches and to perform elabo-rate cognitive operations (e.g., Kruglanski, 2004; Roets, van Hiel, Cornelis, &Soetens, 2008). Emotions research suggests that threat is experienced as aversiveat least in part because it leads people to perceive that they lack both control andcertainty (e.g., see Dechesne & Kruglanski, 2004; Ellsworth & Smith, 1988;Raghunathan & Pham, 1999; Tiedens & Linton, 2001). As a consequence, whenpeople feel threatened they should be highly motivated to restore a subjectivesense of certainty and control. As a result, they might resort to various cognitive-motivational “strategies” that, according to Kruglanski (2004), are tantamount to“closed-mindedness.”

Motivated Closed-Mindedness

Lay epistemic theory offers a useful way of linking what is known about theeffects of threat on human cognition to social and political outcomes (Kruglanski,


1990, 2004). According to this theory, people vary (for dispositional as well assituational reasons) on a single motivational dimension ranging from a strong needto avoid cognitive closure to a strong need to reach closure. High need forcognitive closure represents a desire for “an answer to a question on a given topic,any answer . . . compared to confusion and ambiguity” (Kruglanski, 1990, p. 337),and it often leads to “black-and-white” thinking. The need for closure increases thetendency to “seize” and subsequently “freeze” on highly salient or accessibleinformation when making decisions (Kruglanski, 1990). It arises when extendedinformation processing is perceived to be costly or when the perceived benefits ofpossessing closure are high.

We propose that because threat reduces cognitive capacity and motivation andis associated with feelings of uncertainty and a lack of control, it should increasethe perceived benefits of reaching closure. Consistent with this proposal, scores onan individual difference measure known as the Need for Cognitive Closure scale(NFCS; Webster & Kruglanski, 1994) correlate fairly highly with both state andtrait measures of anxiety (Colbert, Peters, & Garety, 2006). Kruglanski (2004) alsonoted that many of the effects of inducing mortality salience (i.e., increasing deathanxiety) mimic the effects of inducing cognitive closure through time pressure,mental fatigue, or environmental noise. Furthermore, as we have seen with regardto fear, threat, and anxiety, the need for closure is often associated with a decreasein systematic information processing (e.g., DeDreu, Koole, & Oldersma, 1999;Kruglanski, 2004).

The NFCS possesses five subscales that were designed to capture slightlydifferent aspects of epistemic motivation: need for order, ambiguity intolerance,decisiveness, predictability, and closed-mindedness. Several psychometric studieson the NFCS have suggested that the five subscales do not always relate uniformlyto each other or to other constructs (Kossowska & van Hiel, 2003; Neuberg,Judice, & West, 1997; Webster & Kruglanski, 1994). Specifically, the closed-mindedness subscale correlates weakly with the other four subscales and rarelyloads on a second-order latent factor along with them (Mannetti, Pierro, Kruglan-ski, Taris, & Bezinovic, 2002; Neuberg et al., 1997). It also appears to correlatemore strongly (and negatively) with measures of cognitive complexity and thepersonal “need for cognition” (e.g., Houghton & Grewal, 2000; Webster &Kruglanski, 1994).

Items on the closed-mindedness subscale are quite general, relativelycontext-free, and focused on motivation rather than skill per se. They include thefollowing: “I do not usually consult many different opinions before forming myown view,” “I prefer interacting with people whose opinions are very differentfrom my own,” and “When thinking about a problem, I consider as many differentopinions on the issue as possible.” In terms of face validity, the items seemappropriate for tapping into motivational effects of threat, and—unlike measuresof authoritarianism (e.g., Altemeyer, 1998)—their content is not overtly politicalin any way. Thus, any observed association between closed-mindedness and


political orientation would provide clear support for the theoretical model ofideology as motivated social cognition and could not be dismissed as spurious dueto item content. There are both conceptual and empirical reasons, then, to thinkthat the deleterious effects of threat on epistemic motivation (i.e., a decreaseddesire to engage in extensive information search and processing) could bedetected by the closed-mindedness subscale of the NFCS and that closed-mindedness is a theoretically plausible mediator of the threat-conservatism link.

Prior research has connected the need for closure to social and politicalstrategies that are likely to increase the subjective sense of control and certainty.These include the rejection of opinion deviates, the encouragement of autocratic(vs. democratic) leadership, and preferences for opinion uniformity within thegroup (e.g., Kruglanski, Pierro, Mannetti, & De Grada, 2006). There is evidencethat time pressure and cognitive load produce “conservative shifts” in politicalopinion, even among liberal respondents (Hansson, Keating, & Terry, 1974;Skitka, Mullen, Griffin, Hutchinson, & Chamberlin, 2002). Individuals’ scores onthe NFCS are significantly and positively correlated with right-wing authoritari-anism (Webster & Kruglanski, 1994) and political conservatism (Chirumbolo,Areni, & Sensales, 2004; Jost et al., 2003). The need for cognitive closure is alsoassociated with support for the death penalty and for military action against Iraqfollowing 9/11 (Federico, Golec, & Dial, 2005; Jost, Kruglanski, & Simon, 1999).Thus, we hypothesize that motivated closed-mindedness would mediate the effectof threat on political orientation.

Overview of Research

In four experimental studies involving both university students and politicalelites, we investigated the possibility that situational manipulations of perceivedthreat would increase motivated closed-mindedness. That is, threatened individu-als should be narrowly focused (like Kafka’s mouse) to restore a subjective senseof certainty and control, in comparison with nonthreatened individuals. Motivatedclosed-mindedness, we also hypothesized, would be associated with an increasedaffinity for politically conservative (vs. liberal) labels (i.e., self-identifications)and opinions. These studies, when taken in conjunction, help to illuminate thepsychological processes by which threat facilitates increased conservatism(e.g., see Jost et al., 2003, 2007).

Study 1a: Effects of Threat on Need for Cognitive Closure

In our initial study we sought to demonstrate that a situational manipulationof threat would increase motivated closed-mindedness. To our knowledge, nopublished study to date has sought to capture situationally induced changes inepistemic motivation through the use of questions from the NFCS. This was ourgoal in Study 1a, focusing especially on the closed-mindedness subscale.


Method

Participants. Participants were 48 undergraduates (28 females and 20 males)at New York University who participated in the study for partial course credit.Participants were randomly assigned to one of two conditions (high vs. low threat).

Procedure. To provide a cover story for the manipulation of threat, participantswere informed that the experimenter was helping the NYU Wellbeing Center todevelop a new relaxation technique. They were told that the Center was collectingstudent’s real-life past instances of threatening life experiences, and they wereasked to volunteer theirs. Everyone agreed. Participants were then asked to fill outa list of either 3 (low threat) or 12 (high threat) instances in which they “felt terrifiedand/or severely threatened.” Participants were not required to complete the assignednumber of instances; if participants asked whether they needed to complete theentire list, the experimenter merely said “just fill out as many as you can.”

Participants were also asked: “How threatened do you feel right now?”Responses, which were given on a scale ranging from 1 (“Not at all threatened”) to10 (“Very threatened”), served as manipulation checks. They also completed theentire 42-item need for closure scale (NFCS; Webster & Kruglanski, 1994) so thatwe could analyze the effects of threat on specific subscales. Participants werethoroughly debriefed after the study and probed for suspicion. No one expressedsuspicion.

Results

Participants who were assigned to the low-threat condition were asked towrite about three threatening instances; all of them generated three instances.Participants in the high-threat condition were asked to write down 12 instances,and they wrote an average of 11.08 (SD = 2.37) instances, with 22 out of 26participants listing all 12. The fact that so many participants completed the list(despite not being required to do so) suggests that they did not find the task to bedifficult even when asked to generate 12 instances (cf. Schwarz, Bless, Strack,Klumpp, Rittenauer-Schatka, & Simons, 1991). Examples of typical responseswere: “When I was walking home alone at night and someone was following me”;“When my parents used to argue when I was little”; and “When my best friend wasin a car accident.”

In response to the manipulation check item, participants assigned to thehigh-threat condition were somewhat more likely to report feeling threatened(M = 2.32, SD = 1.82) than were participants assigned to the low-threat condition(M = 1.68, SD = .84). However, self-reported feelings of threat were very low inabsolute terms, and this difference did not attain conventional levels of statisticalsignificance, t (46) = 1.53, p = .13.

The intercorrelations and reliability coefficients for the subscales of theNFCS are listed in Table 1. Reliability was good for the NFCS as a whole


(a = 0.88) and for each of the subscales, with the exception of ambiguity intol-erance. All of the subscales except for decisiveness correlated significantly andpositively with each other and with the overall NFCS. To determine whetherscores on the NFCS as a whole or on any of its subscales were affected by thethreat manipulation, a series of t-tests were computed. As can be seen in Table 2,random assignment to the high-threat condition caused participants to score sig-nificantly higher on the closed-mindedness subscale (M = 5.00, SD = 0.74) incomparison with the low-threat condition (M = 4.55, SD = 0.61), t (46) = 2.24,p < .05, two-tailed. No other differences between the two conditions wereobserved.2

2 We also subjected scores on the closed-mindedness subscale to a principal component analysis toinvestigate its structure. The analysis resulted in a solution of one component according to a combinedestimate from a scree plot and eigenvalues (first-component eiginvalue = 2.88, second-componenteigenvalue = 1.24). One item (“I dislike questions that could be answered in many different ways”)

Table 1. Intercorrelations Among the Subscales of the Need for Closure Scale (NFCS) and TheirReliabilities (Study 1a)

Closed-Mindedness Order Predictability Ambiguity Decisiveness

NFCS (Total) 0.64** 0.83** 0.78** 0.67** 0.49**Closed-Mindedness – 0.38** 0.36* 0.30* 0.21Order – 0.63** 0.51** 0.26Predictability – 0.70** 0.20Ambiguity – -0.09Decisiveness –Reliability (a) 0.71 0.79 0.78 0.56 0.81

*p < .10. **p < .05. ***p < .001. All tests are two-tailed.

Table 2. Effects of Assignment to High vs. Low Threat Conditions on NFCS Subscale Scores; Means(Standard Deviations) and Significance Tests (Study 1a)

Low-Threat Condition High-Threat Condition Difference: t (46)

Closed-Mindedness 4.55 5.00 2.24*(0.61) (0.74)

Order 4.52 4.53 .08(0.81) (0.77)

Predictability 4.39 4.52 .57(0.94) (0.68)

Decisiveness 5.62 5.31 1.12(1.06) (0.86)

Ambiguity 4.06 4.03 .16(0.62) (0.61)

NFCS (Total) 4.86 4.91 .28(0.66) (0.47)

*p < .05. All tests are two-tailed.


Study 1b: Effects of Threat on Motivated Closed-Mindedness andPerceptions of a Dangerous World

In Study 1b we investigated the hypothesis that a situational manipulation ofthreat would lead people to (a) exhibit increased closed-mindedness, as in Study1a, and (b) project their feelings of threat onto society in general. Perceivingsociety as dangerous carries with it a host of social and political consequences. Forexample, previous research has shown that perceptions of a dangerous worldcorrelate strongly (approximately r = .50) with right-wing authoritarianism, aswell as punitiveness, xenophobia, and social intolerance (e.g., Altemeyer, 1998;Duckitt, Wagner, du Plessis, & Birum, 2002).

Method

Participants. Participants were 50 undergraduates (37 females and 13 males)at New York University who participated in the study for partial course credit.Participants were randomly assigned to one of two experimental conditions (highvs. low threat).

Procedure. We employed the same manipulation and cover story for inducingthreat as in Study 1a. Participants were again told that the experimenter washelping the NYU Wellbeing Center to develop a new relaxation technique, andthey were asked to fill out a list of either 3 (low threat) or 12 (high threat) instancesfrom their past.

In this study participants were also asked to complete eight items from theperception of a dangerous world scale (PDW; Duckitt et al., 2002)3 and theeight-item closed-mindedness (CM) subscale of the need for closure scale(Webster & Kruglanski, 1994). Responses to both CM and PDW items wereprovided on scales ranging from 1 (“Strongly disagree”) to 9 (“Strongly agree”).Sample items from the PDW scale (a = .84) are: (1) “Despite what one hears about‘crime in the street,’ there probably isn’t any more now than there ever has been”(reverse-scored), (2) “There are many dangerous people in our society who willattack someone out of pure meanness, for no reason at all,” and (3) “If a persontakes a few sensible precautions, nothing bad is likely to happen to him or her; wedo not live in a dangerous world” (reverse-scored). For this sample, the 8-item

did not reach the conventional .30 loading on the 1st component, but all other items did. When theitem was removed from the scale, Cronbach’s alpha rose to .75 (from .71) but results did not change.Participants still scored higher on closed-mindedness in the high-threat (M = 5.23, SD = 0.80) versusthe low-threat condition (M = 4.75, SD = .62), t (46) = 2.37, p < .05.

3 Two items from the original PDW scale were not administered. One of these was worded so stronglythat we expected little variability in responses to it (“Any day now chaos and anarchy could eruptaround us. All the signs are pointing to it.”). The other fairly apocalyptic item made references to the“end” and God and was therefore excluded to prevent potential confounds due to religiosity orpolitical orientation (i.e., “The ‘end’ is not near. People who think that earthquakes, wars, and faminesmean God might be about to destroy the world are being foolish”).


measure of closed-mindedness exhibited low scale reliability (a = .42).4 Partici-pants were thoroughly debriefed and probed for suspicion. No one expressedsuspicion.

Results

Participants who were asked to remember three previously threateninginstances listed on average 2.89 (SD = .41) instances, with 26 out of 28 partici-pants listing all three. Participants assigned to the high-threat condition, who wereasked to recall 12 instances, listed an average of 9.85 (SD = 4.17) instances, withmore than half generating 12 instances.

As can be seen in Table 3, participants who were assigned to the high-threatcondition scored significantly higher on closed-mindedness than did participantsassigned to the low-threat condition, F(1, 48) = 5.82, p < .05, two-tailed. Thisreplicates the major finding of Study 1a. People assigned to the high-threat con-dition also scored marginally higher on perceptions of a dangerous world, F (1,48) = 3.60, p = .06. Scores on closed-mindedness and perceptions of a dangerousworld were positively intercorrelated, r (50) = .46, p < .01.

Discussion of Studies 1a and 1b

Taken in conjunction, Studies 1a and 1b lay the theoretical and methodologi-cal groundwork for further inquiry into the role of threat in shaping social andpolitical attitudes. Study 1a provided evidence that scores on the closed-mindedness subscale of the NFCS can be influenced by a situational threatmanipulation, and Study 1b replicated this effect. Although the data for themanipulation check in Study 1a was weak, we found in Study 1b that the manipu-

4 Because of the low reliability of the eight-item scale, we subjected the scale to a principal componentanalysis. examination of a scree plot and eigenvalues (first-component eigenvalue = 2.00, second-component eigenvalue = 1.27) indicated that a one-component solution was best. Five items loadedhigher than .30 on the first component (items 3–8). This reduced five-item scale had a substantiallyhigher reliability (a = .61). When all results reported for this study were reanalyzed using thefive-item scale, the pattern of results was identical.

Table 3. Effects of Assignment to High- vs. Low-Threat Conditions on Closed-Mindedness andPerceptions of a Dangerous World Scores; Means (with Standard Deviations) and Significance Tests(Study 1b)

Low-Threat Condition High-Threat Condition Difference: F (1, 48)

Closed-Mindedness 3.73 4.26 5.82**(0.87) (0.62)

Perceptions of aDangerous World

4.12 4.85 3.60*(1.24) (1.47)

*p < .06 **p < .05. All tests are two-tailed.


lation of high (vs. low) threat did affect participants’ perceptions of a dangerousworld. Although individual differences in the perception of a dangerous worldhave consistently been linked to right-wing conservatism (Altemeyer, 1998), theconclusion cannot be drawn from this study alone that threat leads people toendorse politically conservative opinions and identifications more strongly (butsee Jost et al., 2004; Nail & McGregor, 2009; Nail et al., 2009). Our remainingexperiments address ideological outcomes more explicitly, in addition to replicat-ing the effect of threat on closed-mindedness and investigating the possibility thatclosed-mindedness mediates the effect of threat on political conservatism.

Study 2: Motivated Closed-Mindedness as a Mediator of theThreat-Conservatism Link

In Study 2, we induced threat with the use of a subtle situational primingtechnique in which participants were less consciously aware of threat-relatedthoughts than in Studies 1a and 1b. The manipulation was based loosely on astudy by Rothman, Haddock, and Schwarz (2001) in which they manipulatedparticipant’s perceptions of how risky their own sexual behavior was. Partici-pants in that study perceived their own sexual behavior as riskier when theyreported their number of previous sexual partners on a scale that was anchoredwith relatively low (vs. high) numbers.5 We used an analogous procedure tomanipulate a sense of threat. Specifically, we expected that people would feel atleast somewhat more threatened when they answered questions about terrorismon a rating scale that was anchored with scale labels assuming a high degree ofthreat, as compared to scale labels suggesting low or moderate degrees of threat.We hypothesized that random assignment to the high (vs. low) threat conditionwould lead people to exhibit increased closed-mindedness and, therefore, astronger affinity for political conservatism.

Method

Participants. Seventy-one undergraduate students from New York University(52 females and 19 males) participated in exchange for partial course credit.

5 Rothman et al. (2001) asked participants to indicate their number of sexual partners (an absolutenumber with objective meaning) on a scale that led them to infer that they were either above averageor below average with respect to promiscuity. We manipulated the scale labels in our experiment fora different purpose and made different theoretical and methodological assumptions. For one thing, wewere interested in the subjective experience of threat (with no objectively correct answer). Foranother, we assumed that (only a few years after 9/11) nearly all New York residents would expresssome degree of concern about terrorism. Thus, we expected that participants would feel morethreatened when they were asked to provide their ratings on a (subjective) scale with high (vs. low)threat labels, insofar as the high-threat labels would invoke a mindset about terrorism that assumedgreater minimum and maximum levels of threat.


Procedure and materials. A sense of threat was induced by manipulating thelabels of the endpoints on a rating scale for seven questions concerning thoughtsabout terrorism. Sample questions include: (1) “I worry that terrorists might strikeanytime anywhere in the United States,” (2) “I worry that a terrorist attack couldhappen in the subway system,” and (3) “I worry that terrorists might attack NewYork City with chemical weapons.” In the low-threat condition, the labels for thescale endpoints ranged from “Not at all” to “Somewhat”; in the high-threatcondition, they ranged from “Somewhat” to “To a great extent.” The endpointlabels constituted the only difference between the two conditions; the items them-selves were the same.

Following the threat manipulation, participants completed a filler task inwhich they were asked to write a “to-do” list as if they needed to apply for a jobon the following day. This task was assigned in order to divorce the manipula-tion from the questionnaire that contained the main dependent measures of thestudy. After the filler task, participants completed the closed-mindedness sub-scale from Webster and Kruglanski’s (1994) need for closure scale (a = .61)6

and also placed themselves on a left-right ideological continuum ranging from-5 (extremely liberal) to +5 (extremely conservative). In addition to being highlyparsimonious, measuring political orientation with a single left-right ideologicalself-placement item is both valid and reliable (e.g., see Jost, 2006; Knight,1999). All participants were thoroughly debriefed and probed for suspicion.None of the participants reported seeing any connection between the terrorismquestions and the dependent measures.

Results and Discussion

The mean self-reported fear of terrorism in the high-threat condition, in whichthe scale labels ranged from “Somewhat” to “To a great extent,” was 3.49(SD = 1.96). This was not significantly different from the mean observed in thelow-threat condition in which the scale labels ranged from “Not at all” to “Some-what” (M = 4.14, SD = 1.64), t (69) = 1.52, ns. Thus, any differences between thetwo conditions in terms of threat and its ideological consequences can be assumedto be nonconscious, at least for some of the participants (cf. Arndt, Greenberg, &Cook, 2002).

As hypothesized, participants assigned to the high-threat condition scoredsignificantly higher on closed-mindedness (M = 4.29, SD = 0.89) than did thoseassigned to the low-threat condition (M = 3.81, SD = 0.94), t (69) = 2.18, p < .05.This finding replicates the results of Studies 1a and 1b. Furthermore, participantswho were assigned to the high-threat condition also identified themselves as more

6 A principal component analysis was conducted again on the eight items. This time, all items loadedhigher than .30 on the first component and the highest reliability coefficient (a = .61) was obtainedusing all eight items.


conservative (or less liberal) on the ideological self-placement item (M = -.75,SD = 2.38), compared to those participants who were assigned to the low-threatcondition (M = -1.84, SD = 2.01), t (68) = 2.09, p < .05. This finding supplementsthe results of prior research in which threat was found to increase politicallyconservative opinions (Jost et al., 2004) and support for a conservative leader(Landau et al., 2004b; Willer, 2004).

Finally, we conducted a bootstrap bias-corrected mediational analysis(Shrout & Bolger, 2002) using AMOS software to examine a model in which theeffect of high (vs. low) threat on liberal-conservative self-placement was medi-ated by closed-mindedness (see Figure 1). The analysis indicated mediation forthe effect of threat on ideological self-placement, insofar as the direct effect wasno longer significant after adjusting for closed-mindedness (b = .82, SE = .52,ns), whereas the indirect effect was significant (b = .27, SE = .19, p < .05).Threat significantly predicted closed-mindedness in the mediational model(b = .48, SE = .22, p < .05), and it, in turn, predicted ideological self-placement(b = .56, SE = .28, p < .05).7

Thus, Study 2 demonstrated that participants who are subtly primed with high(vs. low) threat subsequently identify themselves as more conservative (or lessliberal) on the ideological spectrum. Importantly, the relationship between threatand conservative self-placement was statistically mediated by closed-mindedness,as hypothesized. This study, therefore, corroborates the notion that when peopleare threatened, they are less motivated to elaborate on their cognitions, and this, in

7 We also analyzed a reverse-mediation model in which the effect of threat on closed-mindedness wasmediated by political orientation. Results from this reversed mediation model revealed a marginallysignificant indirect effect (b = .11, SE = .07, p = .05), but the effect was considerably weaker than theeffect obtained for the hypothesized causal order.

*Significant at the p < .05 level.

Closed-mindedness

1.09*(.52)

[.82 ns (.52)]

)82.(*65.)22.(*84.

High (vs. low)threat

Conservative (vs.liberal) self-placement

Figure 1. Mediational analysis illustrating the direct and indirect effects of high vs. low threat onmotivated closed-mindedness and conservative self-placement (Study 2).The figure displays unstandardized regression coefficients (and standard errors). The numbers inbrackets represent the coefficients after adjusting for closed-mindedness.


turn, draws them toward conservative ideology (or away from liberal ideology). Itis worth noting that the findings from Study 2 contradict expectations derived fromterror management theory (Anson, Pyszczynski, Solomon, & Greenberg, 2009;Greenberg et al., 1992; Greenberg & Jonas, 2003), which suggests that whenpeople experience existential anxiety, they should seek to reaffirm their existingvalues, whether liberal or conservative. Because participants in this study werequite liberal on average, terror management theory would predict that the threatmanipulation should have made most of them more liberal rather than moreconservative. Instead, we found that participants on average exhibited a “conser-vative shift” (Bonanno & Jost, 2006; Nail & McGregor, 2009), consistent with thenotion that certain characteristics of conservative ideology make it more attractivethan liberal ideology under conditions of threat (see also Jost et al., 2003, 2004,2007).

Study 3: Manipulating Motivated Closed-Mindedness Directly

Study 2 provides solid statistical evidence consistent with the causal chain wehave proposed, linking threat to closed-mindedness and, in turn, political conser-vatism. Spencer, Zanna, and Fong (2005) have pointed out, however, that when itcomes to causal claims, experimental studies in which the hypothesized mediatoris directly manipulated and the dependent variable is measured are methodologi-cally superior to statistical analyses of mediation alone. This is because media-tional analyses (such as the analysis reported in Study 2) rarely fulfill twoimportant requirements of causality, namely temporal order and isolation. In Study3, we directly manipulated the proposed mediator, closed-mindedness, in order toinvestigate its impact on political self-identification.

Method

Thirty-one undergraduate students (24 females, 7 males) from PrincetonUniversity were randomly assigned to either a high or low cognitive-load manipu-lation adapted from procedures used by Gilbert and Osborne (1989) and Ford andKruglanski (1995). In the high-load condition, participants were given 10 secondsto memorize a fairly difficult eight-digit number (29364095) and instructed to keepit in memory while they filled out several questionnaires (about 10 minutes). In thelow-load condition the number was an easy two-digit number (22). After complet-ing several questionnaires pertaining to another study, participants answered twoquestions pertaining to political orientation. Specifically, they located themselveson a left-right ideological self-placement item ranging from -4 (Liberal) to +4(Conservative) and also indicated the strength of their party affiliation, rangingfrom -4 (Strong Democrat) to +4 (Strong Republican). Responses to these twoitems were positively intercorrelated, r (31) = .79, p < .01.


Results and Discussion

Participants who were randomly assigned to the high cognitive-load conditionrated themselves as slightly more conservative (or less liberal; M = -.94,SD = 1.81), in comparison with those who were randomly assigned to the lowcognitive-load condition (M = -1.53, SD = 1.46), but this difference did not reachstatistical significance, t (29) = 1.01, ns. With regard to party affiliation, however,the difference between the two conditions was statistically significant. Participantsassigned to the high cognitive-load condition were, on average, less enthusiasticabout their identification with the Democratic party (M = -.94, SD = 1.34) thanwere participants assigned to the low cognitive-load condition (M = -2.00,SD = 1.51), t (29) = 2.07, p < .05. Thus, a direct experimental manipulation of theproposed mediator was generally successful at increasing the extent of conserva-tive self-identification (or, in this case, undermining liberal self-identification).

Study 4: A Study of Political Elites

Whereas Studies 1–3 were conducted with samples of predominantly liberalcollege students in the United States, Study 4 investigated the effects of threat onthe social and political attitudes of conservative (center-right) political elites (i.e.,party delegates) in Iceland. This provided an excellent opportunity to evaluate thegeneralizability of our findings. Study 4 also afforded a theoretical opportunity toinvestigate three different types of threats (i.e., threats to the self, group, andsystem) to determine whether their effects on closed-mindedness and politicalconservatism would be similar or dissimilar. Among other things, we were able toexamine whether threats to the economic system or to the party’s political successwould exert ideological effects that were comparable to those elicited by mortalitysalience threats in other experiments.

Method

Participants. Participants were official delegates of the Icelandic Indepen-dence Party’s national convention in 2005. The Independence Party is a center-right party that espouses probusiness, procapitalist policies as well as liberal“Scandinavian style” positions on cultural, gender, and welfare issues. The studywas conducted over the Internet using createsurvey.com. We received 481 validresponses (a 45.4% response rate). The gender distribution of respondents closelyreflected that of the delegates as a whole (72% male and 28% female).

Materials. Delegates were randomly assigned to one of four conditions,including three different high-threat conditions in which they were asked to readand reflect on a 90-word passage that described a threat to (a) the economic systemof Iceland, (b) their group (the Independence Party), or (c) their personal well-being. For example, the system threat passage read as follows:


Although most people agree that Icelanders have done very well forthemselves since gaining independence in 1944, we always live in theshadow of a great variety of threats to our independence and economy.We still rely first and foremost on a single natural resource to create themajority of our national income which makes our economy unstable. TheIcelandic currency, although remarkably strong right now, is after all verysmall and vulnerable in the international context. Finally, foreign inves-tors could start large scale investing in Icelandic companies, threateningour autonomy.8

Delegates assigned to the control condition were not exposed to any passage priorto answering the remaining questions. All participants completed the closed-mindedness subscale from the NFCS, a measure of issue-based liberalism-conservatism, and an ideological self-placement scale that ranged from 1(extremely left-wing) to 11 (extremely right-wing). For this study, we calculatedmean scores on closed-mindedness based on responses to the best five items fromthat subscale (a = 0.59).9,10

To make the questions more suitable to the contemporary Icelandic context,the items used to measure issue-based liberalism-conservatism were selected andmodified from Sidanius and Ekehammar’s (1980) S-Scale, which was originallydesigned to gauge social and political attitudes in the Swedish context. On a7-point scale ranging from “very negative” to “very positive,” respondents indi-cated how positively or negatively they felt about 16 specific ideas or statements,namely: (1) “Affirmative action in job hiring where one gender is underrepre-sented” (reverse-scored); (2) “Increased taxation of the richest” (reverse-scored);(3) “Privatization of the Icelandic National Broadcasting Service”; (4) “Statesupported agriculture” (reverse-scored); (5) “Greater equality in salaries” (reverse-scored); (6) “Privately run elementary schools”; (7) “Increased aid to the lowestwage earners” (reverse-scored); (8) “Completely government funded nationalhealth care” (reverse-scored); (9) “Legalization of marijuana” (reverse-scored);(10) “Capitalism”; (11) “Libertarianism”; (12) “Privatization of state owned busi-nesses”; (13) “Lower taxes on companies”; (14) “Privately run health care industry

8 This study was conducted three years before the collapse of the Icelandic economy in late 2008, wellbefore the first signs of trouble started to emerge in the banking industry.

9 An examination of the seven closed-mindedness items revealed that most of them had a skew statisticabove 1.0, indicating a positive skew high enough to potentially violate assumptions of normality.The skewness was adjusted for by square root transformation. All means and regression coefficientsreported in the results section are expressed in the variable’s original metric for ease of interpreta-tion, but all statistical tests and the path model were conducted using the transformed variable.

10 A principal component analysis of the closed-mindedness subscale resulted in one component,according to a joint examination of a scree plot and eigenvalues (1st component = 1.98, 2ndcomponent = 1.14). Five of the seven items loaded higher than .30 on the first component. The twoitems that were dropped were: “I dislike questions which could be answered in many different ways,”and “I feel irritated when one person disagrees with what everyone else in a group believes.”


(but publicly funded)”; (15) “Adoption of tuition at the college level”; and (16)“Iceland being on the ‘coalition of the willing’ list.” The resulting scale wasinternally reliable (a = 0.81) and composite scores correlated significantly withideological self-placement on the left-right scale (r = .39, p < .01).

Results

The descriptive statistics for closed-mindedness, issue-based liberalism-conservatism, and left-right self-placement for each of the four conditions arelisted in Table 4. Delegates assigned to the three experimental (high threat) con-ditions were significantly more likely than delegates assigned to the control con-dition to endorse politically conservative opinions on the composite measure ofissue-based conservatism, F (3, 478) = 4.29, p < .01. None of the means in thethree high-threat conditions differed from each other, but all were significantlydifferent from the control condition.

The effects of threat on ideological self-placement did not attain conventionallevels of significance in this study, F (3, 478) = .67, ns, presumably because alldelegates were already strongly and self-consciously identified with the right wing(approximately 8.90 on a scale from 1 to 11). When closed-mindedness scoreswere pooled for those delegates assigned to the three high-threat conditions, theresulting average (M = 3.08, SD = 1.26) was significantly higher than the averagescore for those delegates assigned to the control condition (M = 2.82, SD = .05),F (1, 480) = 5.23, p < .05.

To investigate the possibility that closed-mindedness mediated the relation-ship between threat and political attitudes, we conducted a bias-corrected media-

Table 4. Effects of Assignment to Self, Group, or System Threat Conditions on Closed-Mindedness,Issue-Based Conservatism, and Right-Wing Self-Placement; Means (with Standard Deviations;Study 4)

No Threat/Control Self Threat Group Threat System Threat

Closed-mindedness 2.82a 3.10 3.10 3.04(1.10) (1.25) (1.25) (1.27)

Issue-based conservatism 4.56b 5.00c 4.88c 4.89c

(0.95) (0.97) (0.89) (1.00)Right-wing self-placement 8.74d 8.93 8.90 8.93

(1.33) (1.18) (1.05) (1.43)n 129 110 123 116aAlthough pairwise comparisons were nonsignificant, closed-mindedness scores were significantlylower (on average) in the no threat/control condition than in the three threat conditions (combined;M = 3.08, SD = 1.26), F (1, 480) = 5.23, p < .05.b, cIssue-based conservatism was significantly lower (on average) in the no threat/control conditionthan in any of the three threat conditions (p < .05), which did not differ from one another.dNo significant differences in right-wing self-placement were observed between the no threat/controlcondition and the other three condition, possibly because of restricted range (see text).


tional analysis using AMOS (Shrout & Bolger, 2002). The three threat conditions(self, group, and system) were combined and compared to the control conditionto create a binary independent variable suitable for a path analysis (0 = no threat,1= threat). As illustrated in Figure 2, threat (vs. no threat) predicted closed-mindedness (b = .26, SE = .13, p < .05), which in turn predicted the endorsementof more conservative political attitudes (b = .07, SE = .03, p < .01). The direct pathfrom threat to political attitudes was reduced slightly from b = .33, SE = .10, tob = .31 SE = .10. Accordingly, there was a significant, albeit small indirect effectof threat on political attitudes (b = .02, SE = .01, p < .05), providing evidence forpartial mediation.

General Discussion

In several experiments, we have shown that laboratory manipulations of threatserve to increase motivated closed-mindedness and, in so doing, to increase thepsychological affinity for politically conservative opinions and identifications.This overall pattern of results was observed with respect to American and Euro-pean research participants, college students and political elites, and in predomi-nantly liberal as well as predominantly conservative samples. Taken as a whole,these findings provide reasonably strong support for Jost et al.’s (2003) theory ofpolitical conservatism as motivated social cognition and contradict the notion thatthreat increases individuals’ adherence to their own political ideology, even if thatideology happens to be a liberal one (e.g., see Anson et al., 2009; Greenberg &Jonas, 2003). Because the experiments here employed situational manipulations ofthreat to document “conservative shift” (see also Bonanno & Jost, 2006, for alongitudinal field study), our results complement those gleaned from personality

*Significant at the p < .05 level.

Threat (vs. no threat) condition

Conservative (vs. liberal) attitudes

Closed -mindedness

.26*(.13) .07*(.03)

.33*(.10)

[.31*(.10)]

Figure 2. Mediational analysis illustrating the direct and indirect effects of threat on motivatedclosed-mindedness and issue-based conservatism in a sample of political elites (study 4).The figure displays unstandardized regression coefficients (and standard errors). The numbers inbrackets represent the coefficients after adjusting for closed-mindedness.


or dispositional studies in which individual differences pertaining to threat sensi-tivity and uncertainty reduction have been shown to be associated with politicallyconservative (vs. liberal) orientation (e.g., Amodio et al., 2007; Jost et al., 2007;Oxley et al., 2008). Thus, dispositional and situational variables seem to exertparallel effects on ideological outcomes.

One of the most important empirical contributions of this set of experimentsis the demonstration that motivated closed-mindedness statistically mediates therelationship between threat and political conservatism. This does not mean,however, that closed-mindedness is the only mediator of the threat-ideology link,especially in light of the relatively weak indirect effect observed for political elitesin Study 4. It seems likely that threat exerts other psychological effects in additionto reducing epistemic motivation and that these can also affect political ideology.Along these lines, Lambert et al. (2010) provided evidence suggesting that theemotion of anger may also play an important role. We think that future researchwould do well to consider the possibility that threat elicits anxiety, which, in turn,reduces cognitive capacity and increases the need for cognitive closure, ultimatelyleading to an increased affinity for conservatism. Unfortunately, we did not includeany direct measures of anxiety, so we were unable to assess this possibility withthe present set of studies. By demonstrating that closed-mindedness mediates theeffect of threat on conservatism, however, we have shed further empirical lighton the complex relationship between existential motives to manage threat andepistemic motives to reduce uncertainty (e.g., Dechesne & Kruglanski, 2004;Hogg, 2007, pp. 108–110; Landau et al., 2004a; van den Bos et al., 2005). Previ-ous research has shown that each motive contributes significantly and indepen-dently to political conservatism (Jost et al., 2007). In the research programsummarized here, we have found that a causal relationship also holds between theexperience of threat and the need to reduce uncertainty, again with the end resultof increased conservatism.

To be clear, we do not assume that threat will always make people moreconservative (cf. Weise et al., 2008). Although we observed no evidence of liberalshift in these experiments, we did not observe conservative shifts for everymeasure of political orientation that we administered (see also Nail & McGregor,2009). We think that threat is especially likely to yield conservative results whenpeople are relatively unaware of the connection between their own emotional andmotivational states (e.g., fear, anxiety, anger, and closed-mindedness) and theirself-proclaimed political attitudes. However, when people are made explicitlyaware of such connections—as when Senator Joseph Biden (D-DE) criticized thespeeches of fellow presidential hopeful Rudy Giuliani (R-NY) as containing only“a noun, a verb, and 9/11” during the 2008 electoral campaign—they may besignificantly less prone to exhibit increased conservatism. Under such circum-stances, processes of psychological reactance are probably more likely to produceideological polarization rather than a general conservative shift (e.g., see Brehm &Mann, 1975).


In the present set of experiments, epistemic motivation was measured usingscores on the closed-mindedness subscale of the NFCS (Webster & Kruglanski,1994). Although attempts to capture situationally induced changes in epistemicmotivation with the NFCS have not (to our knowledge) been reported previouslyin the literature, it turned out to be a successful method, insofar as the effect ofthreat on closed-mindedness was consistently observed in four studies. In light ofthe fact that scores on the closed-mindedness subscale were relatively low inabsolute terms, it may be more accurate to suggest that the threat manipulationsresulted in decreased open-mindedness rather than increased closed-mindedness.In any case, what is clear is that threat affects epistemic motivation and, in sodoing, appears to (temporarily) alter one’s social and political attitudes. Researchsuch as this suggests that ideological belief systems can be “genuine,” “meaning-ful,” and “consequential,” and yet also subject to some degree of situationalmalleability or—in the hands of experimental social psychologists or Machiavel-lian political operatives—manipulation (see also Jost, 2006).

The mouse in Kafka’s parable cited in the epigram of this article was over-come with fear. Initially it was afraid of the vast, uncertain surroundings, and so itwelcomed the confining walls to the left and right. By the time the reader happensupon the situation, the mouse’s thinking has narrowed along with the walls. Themouse fails to realize that a change of direction is possible and so continues on anultimately self-destructive path. The cat understands better but uses the knowledgeto its own advantage. The parallels between the lessons of “The Little Fable” andthe research described in this article may be indirect, but they are probably not loston the reader. Human beings, like many other species, are sensitive to threat andexperience the lack of certainty and control as highly aversive. Consequently, thereis a risk of becoming so preoccupied with relieving the threat that cognitivealternatives simply fall away, and the easiest or most accessible course of thoughtand action prevails—whether it leads us to a rewarding morsel of cheese or rightinto the cat’s maw.

ACKNOWLEDGMENTS

This article is based in part on a dissertation submitted by the first author tothe Department of Psychology at NewYork University under the supervision of thesecond author. We would like to thank the other members of the dissertationcommittee, namely Elizabeth Phelps, Patrick Shrout, Yaacov Trope, and TomTyler. The thesis itself received honorable mentions from the International Societyfor Political Psychology (ISPP) and Society for the Psychological Study of SocialIssues (SPSSI) dissertation award committees. The research described here wasfunded by the NYU Center for Catastrophe Preparedness and Response. Thearticle was written during a period in which the second author also benefited froma National Science Foundation Award (BCS-0617558). This work has been pre-sented at several colleges and universities, including Princeton, Ohio State, Bard


College, Kyoto University, the IDC, and the University of Illinois, and at confer-ences sponsored by ISPP and the Society for Personality and Social Psychology(SPSP). We thank Ira Carmen, Alan Lambert, Howard Lavine, and two anonymousreviewers for helpful comments on earlier versions of the article. Correspondenceconcerning this article should be sent to Dr. Hulda Thorisdottir, Oddiv/Saemundargotu, Department of Political Science, University of Iceland, 101Reykjavík, Iceland. E-mail: [email protected].

REFERENCES

Adorno, T. W., Frenkel-Brunswick, E., Levinson, D. J., & Sanford, R. N. (1950). The authoritarianpersonality. New York: Harper & Brothers.

Altemeyer, R. A. (1998). The other “authoritarian personality.” In M. P. Zanna (Ed.), Advances inexperimental social psychology (Vol. 30, pp. 47–91). New York: Academic Press.

Amodio, D. M., Jost, J. T., Master, S. L., & Yee, C. M. (2007). Neurocognitive correlates of liberalismand conservatism. Nature Neuroscience, 10, 1246–1247.

Anson, J., Pyszczynski, T., Solomon, S., & Greenberg, J. (2009). Political ideology in the 21st century:A terror management perspective on maintenance and change of the status quo. In J. T. Jost,A. C. Kay, & H. Thórisdóttir (Eds.), Social and psychological bases of ideology and systemjustification (pp. 210–240). New York: Oxford University Press.

Arndt, J., Greenberg, J., & Cook, A. (2002). Mortality salience and the spreading activation ofworldview-relevant constructs: Exploring the cognitive architecture of terror management.Journal of Experimental Psychology: General, 131, 307–324.

Bennett, D. H. (1995). The party of fear (2nd ed.). New York: Random House.

Bobbio, N. (1996). Left and right: The significance of a political distinction. Cambridge: Polity Press.

Bonanno, G. A., & Jost, J. T. (2006). Conservative shift among high-exposure survivors of theSeptember 11th terrorist attacks. Basic and Applied Social Psychology, 28, 311–323.

Brehm, J. W., & Mann, M. (1975). Effect of importance of freedom and attraction to group memberson influence produced by group pressure. Journal of Personality and Social Psychology, 31,816–824.

Burke, E. (1790/1999). Reflections on the revolution in France. New York: Oxford University Press.

Carney, D. R., Jost, J. T., Gosling, S. D., & Potter, J. (2008). The secret lives of liberals andconservatives: Personality profiles, interaction styles, and the things they leave behind. PoliticalPsychology, 29, 807–840.

Chajut, E., & Algom, D. (2003). Selective attention improves under stress: Implications for theories ofsocial cognition. Journal of Personality and Social Psychology, 85, 231–248.

Chirumbolo, A., Areni, A., & Sensales, G. (2004). Need for cognitive closure and politics: Voting,political attitudes, and attributional style. International Journal of Psychology, 39, 245–253.

Cohen, F., Ogilvie, D. M., Solomon, S., Greenberg, J., & Pyszczynski, T. (2005). American roulette:The effect of reminders of death on support for George W. Bush in the 2004 presidential election.Analyses of Social Issues and Public Policy, 5, 177–187.

Cohrs, J. C., Kielmann, S., Maes, J., & Moschner, B. (2005). Effects of right-wing authoritarianism andthreat from terrorism on restriction of civil liberties. Analyses of Social Issues and Public Policy,5, 263–276.

Colbert, S. M., Peters, E. R., & Garety, P. A. (2006). Need for closure and anxiety in delusions: Alongitudinal investigation in early psychosis. Behaviour Research and Therapy, 44, 1385–1396.

Davis, D. W., & Silver, B. D. (2004). Civil liberties vs. security: Public opinion in the context of theterrorist attacks on America. American Journal of Political Science, 48, 28–46.


Dechesne, M., & Kruglanski, A. W. (2004). Terror’s epistemic consequences: Existential threats andthe quest for certainty and closure. In J. Greenberg, S. Koole, & T. Pyszczynski (Eds.), Handbookof experimental existential psychology (pp. 247–262). New York: Guilford.

DeDreu, C. K., Koole, S. L., & Oldersma, F. L. (1999). On the seizing and freezing of negotiatorinferences: Need for cognitive closure moderates the use of heuristics in negotiation. Personalityand Social Psychology Bulletin, 25, 348–362.

Doty, R. M., Peterson, B. E., & Winter, D. G. (1991). Threat and authoritarianism in the United States,1978–1987. Journal of Personality and Social Psychology, 61, 629–640.

Duckitt, J., & Fisher, K. (2003). The impact of social threat on world view and ideological attitudes.Political Psychology, 24, 199–222.

Duckitt, J., Wagner, C., du Plessis, I., & Birum, I. (2002). The psychological bases of ideology andprejudice: Testing a dual process model. Journal of Personality and Social Psychology, 83,75–93.

Easterbrook, J. A. (1959). The effect of emotion on cue utilization and the organization of behavior.Psychological Review, 74, 16–28.

Echebarria, A., & Fernández, E. (2006). Effects of terrorism on attitudes and ideological orientation.European Journal of Social Psychology. 36, 259–265.

Eibach, R. P., & Libby, L. K. (2009). Ideology of the good old days: Exaggerated perceptions of declineand conservative politics. In J. T. Jost, A. C. Kay, & H. Thórisdóttir (Eds.), Social and psychologi-cal basis of ideology and system justification (pp. 402–426). New York: Oxford University Press.

Ellsworth, P. C., & Smith, C. A. (1988). From appraisal to emotion: Differences among unpleasantfeelings. Motivation & Emotion, 12, 271–302.

Eysenck, M. W., & Calvo, M. G. (1992). Anxiety and performance: The processing efficiency theory.Cognition and Emotion, 6, 409–434.

Federico, C. M., Golec, A., & Dial, J. L. (2005). The relationship between the need for closure andsupport for military action against Iraq: Moderating effects of national attachment. Personalityand Social Psychology Bulletin, 31, 621–632.

Ford, T. E., & Kruglanski, A. W. (1995). Effects of epistemic motivations on the use of accessibleconstructs in social judgment. Personality and Social Psychology Bulletin, 21, 950–962.

Fromm, E. (1941). Escape from freedom. New York: Holt, Rinehart, & Winston.

Gailliot, M. T., Schmeichel, B. J., & Baumeister, R. (2006). Self-regulatory processes defend againstthe threat of death: Effects of self-control depletion and trait self-control on thoughts and fears ofdying. Journal of Personality and Social Psychology, 91, 49–62.

Gilbert, D. T. & Osborne, R. E. (1989). Thinking backward: Some curable and incurable consequencesof cognitive busyness. Journal of Personality and Social Psychology, 57, 940–949.

Greenberg, J., & Jonas, E. (2003). Psychological and political orientation—The left, the right, and therigid: Comment on Jost et al. (2003). Psychological Bulletin, 129, 376–382.

Greenberg, J., Simon, L., Pyszczynski, T., Solomon, S., & Chatel, D. (1992).Terror management andtolerance: Does mortality salience always intensify negative reactions to others who threatenone’s world-view? Journal of Personality and Social Psychology, 63, 212–220.

Haidt, J., & Graham, J. (2007). When morality opposes justice: Conservatives have moral intuitionsthat liberals may not recognize. Social Justice Research, 20, 98–116.

Hansson, R. O., Keating, J. P., & Terry, C. (1974). The effects of mandatory time limits in the votingbooth on liberal-conservative voting patterns. Journal of Applied Social Psychology, 4, 336–342.

Hillier, A., Alexander, J. K., & Beversdorf, D. Q. (2006). The effect of auditory stressors on cognitiveflexibility. Neurocase, 12, 228–231.

Hodges, W. F., & Spielberger, C. D. (1969). Digit span: An indicant of trait or state anxiety? Journalof Consulting and Clinical Psychology, 33, 430–434.

Hogg, M. A. (2007). Uncertainty-identity theory. In M. P. Zanna (Ed.), Advances in experimental socialpsychology (Vol. 39, pp. 69–126). San Diego, CA: Academic Press.


Houghton, D. C., & Grewal, R. (2000). Please, let’s get an answer—any answer: Need for consumercognitive closure. Psychology and Marketing, 17, 911–934.

Huddy, L., Khatid, N., & Capelos, T. (2002). Trends: Reactions to the terrorism attacks of September11, 2001. Public Opinion Quarterly, 66, 418–450.

Jameson, T. L., Hinson, J. M., & Whitney, P. (2004). Components of working memory and somaticmarkers in decision making. Psychonomic Bulletin and Review, 11, 515–520.

Jost, J. T. (2006). The end of the end of ideology. American Psychologist, 61, 651–670.

Jost, J. T., Federico, C. M., & Napier, J. (2009). Political ideology: Its structure, functions, and electiveaffinities. Annual Review of Psychology, 60, 307–337.

Jost, J. T., Fitzsimons, G. M., & Kay, A. C. (2004). The ideological animal: A system justification view.In J. Greenberg, S. L. Koole, & T. Pyszczynski (Eds.), Handbook of experimental existentialpsychology (pp. 263–283). New York: Guilford Press.

Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J. (2003). Political conservatism as motivatedsocial cognition. Psychological Bulletin, 129, 339–375.

Jost, J. T., Kruglanski, A. W., & Simon, L. (1999). Effects of epistemic motivation on conservatism,intolerance, and other system-justifying attitudes. In L. L. Thompson, J. M. Levine, & D. M.Messick (Eds.), Shared cognition in organizations: The management of knowledge (pp. 91–116).Mahwah, NJ: Erlbaum.

Jost, J. T., Napier, J. L., Thórisdóttir, H., Gosling, S. D., Palfai, T. B., & Ostafin, B. (2007). Are needsto manage uncertainty and threat associated with political conservatism or ideological extremity?Personality and Social Psychological Bulletin, 33, 989–1007.

Keinan, G. (1987). Decision making under stress: Scanning of alternatives under controllable anduncontrollable threats. Journal of Personality and Social Psychology, 52, 639–644.

Kirk, R. (Ed.) (1982). The conservative reader. Harmondsworth, UK: Penguin.

Knight, K. (1999). Liberalism and conservatism. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman(Eds.), Measures of political attitudes (pp. 59–158). San Diego, CA: Academic Press.

Kossowska, M., & van Hiel, A. (2003). The relationship between need for closure and conservativebeliefs in Western and Eastern Europe. Political Psychology, 24, 501–518.

Kruglanski, A. W. (1990). Lay epistemic theory in social-cognitive psychology. Psychological Inquiry,1, 181–197.

Kruglanski, A. W. (2004). The psychology of closed-mindedness. New York: Psychology Press.

Kruglanski, A. W., Pierro, A., Mannetti, L., & De Grada, E. (2006). Groups as epistemic providers:Need for closure and the unfolding of group-centrism. Psychological Review, 113, 84–100.

Lambert, A., Scherer, L., Schott, J. P., Olson, K., Andrews, R., O’Brien, T., & Zisser, A. (2010). Rallyeffects, threat, and attitude change: An integrative approach to understanding the role of emotion.Journal of Personality and Social Psychology, 98, 886–903.

Landau, M. J., Johns, M., Greenberg, J., Pyszczynski, T., Martens, A., Goldenberg, J. L., & Solomon,S. (2004a). A function of form: Terror management and structuring the social world. Journal ofPersonality and Social Psychology, 87, 190–210.

Landau, M. J., Solomon, S., Greenberg, J., Cohen, F., Pyszczynski, T., Arndt, J., Miller, C. H., Ogilvie,D. M., & Cook, A. (2004b). Deliver us from evil: The effects of mortality salience and remindersof 9/11 on support for President George W. Bush. Personality and Social Psychology Bulletin, 30,1136–1150.

Laponce, J. A. (1981). Left and right: The topography of political perceptions. Toronto: University ofToronto Press.

Lavine, H., Lodge, M., Polichak, J., & Taber, C. (2002). Explicating the black box through experi-mentation: Studies of authoritarianism and threat. Political Analysis, 10, 343–361.

Lipset, S. M., & Raab. (1978). The politics of unreason. Chicago: University of Chicago Press.

Machiavelli, N. (1999/1513). The prince. London and New York: Penguin Classics.


Mack, A. (Ed.) (2004). Fear: Its political uses and abuses. Social Research. Special Issue.

Mannetti, L., Pierro, A., Kruglanski, A. W., Taris, T., & Bezinovic, P. (2002). A cross-cultural study ofthe Need for Cognitive Closure Scale: Comparing its structure in Croatia, Italy, the USA and theNetherlands. British Journal of Social Psychology, 41, 139–156.

Mathews A., & Mackintosh B. (1998). A cognitive model of selective processing in anxiety. CognitiveTherapy and Research, 22, 539–560.

Minogue, K. (1967). Conservatism. In P. Edwards (Ed.), The encyclopedia of philosophy (Vol. 2,pp. 195–198). New York: Macmillan.

Miu, A. C., Heilman, R. M., & Houser, D. (2008). Anxiety impairs decision-making: Psychophysi-ological evidence from an Iowa Gambling Task. Biological Psychology, 77, 353–358.

Mogg, K., Mathews, A., Bird, C., & Macgregor-Morris, R. (1990). Effects of stress and anxietyon the processing of threat stimuli. Journal of Personality and Social Psychology, 59, 1230–1237.

Muller, J. Z. (2001). Conservatism: Historical aspects. In N. J. Smelser & P. B. Baltes (Eds.),International encyclopedia of the social and behavioral sciences (pp. 2624–2628). Amsterdam:Elsevier.

Nail, P. R., & McGregor, I. (2009). Conservative shift among liberals and conservatives following9/11/01. Social Justice Research, 22, 231–240.

Nail, P. R., McGregor, I., Drinkwater, A. E., Steele, G. M., & Thompson, A. W. (2009). Threatcauses liberals to think like conservatives. Journal of Experimental Social Psychology, 45,901–907.

Napier, J. L., & Jost, J. T. (2008). The “anti-democratic personality” revisited: A cross-nationalinvestigation of working class authoritarianism. Journal of Social Issues, 64, 595–617.

Neuberg, S. L., Judice, T. N., & West, S. G. (1997). What the Need for Closure Scale measures andwhat it does not: Toward differentiating among related epistemic motives. Journal of Personalityand Social Psychology, 72, 1396–1412.

Oxley, D. R., Smith, K. B., Alford, J. R., Hibbing, M. V., Miller, J. L., Scalora, M., Hatemi, P. K., &Hibbing J. R. (2008). Political attitudes vary with physiological traits. Science, 32, 1667–70.

Preston, S. D., Buchanan, T. W., Stansfield, R. B., & Bechara, A. (2007). Effects of anticipatory stresson decision making in a gambling task. Behavioral Neuroscience, 121, 257–263.

Raghunathan, R., & Pham, M. T. (1999). All negative moods are not equal: Motivational influences ofanxiety and sadness on decision making. Organizational Behavior and Human Decision Pro-cesses, 79, 56–77.

Robin, C. (2004). Fear: The history of a political idea. New York: Oxford University Press.

Roets, A., van Hiel, A., Cornelis, I., & Soetens, B. (2008). Determinants of task performance andinvested effort: A need for closure by relative cognitive capacity interaction analysis. Personalityand Social Psychology Bulletin, 34, 779–792.

Rothman, A. J., Haddock, G., & Schwarz, N. (2001). “How many partners is too many?” Shapingperceptions of vulnerability. Journal of Applied Social Psychology, 31, 2195–2214.

Sales, S. M. (1973). Threat as a factor in authoritarianism: An analysis of archival data. Journal ofPersonality and Social Psychology, 28, 44–57.

Schwarz, N., Bless, H., Strack, F., Klumpp, G., Rittenauer-Schatka, H., & Simons, A. (1991). Ease ofretrieval as information: Another look at the availability heuristic. Journal of Personality andSocial Psychology, 61, 195–202.

Sengupta, J., & Johar, G. V. (2001). Contingent effects of anxiety on message elaboration andpersuasion. Personality and Social Psychology Bulletin, 27, 139–150.

Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: Newprocedures and recommendations. Psychological Methods, 7, 422–445.

Sidanius, J., & Ekehammar, B. (1980). Sex-related differences in socio-political ideology. Journal ofPsychology, 21, 17–26.


Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy andoppression. New York: Cambridge University Press.

Skitka, L. J., Mullen, E., Griffin, T., Hutchinson, S., & Chamberlin, B. (2002). Dispositions, scripts, ormotivated correction? Understanding ideological differences in explanations for social problems.Journal of Personality and Social Psychology, 83, 470–487.

Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). Establishing a causal chain: Why experiments areoften more effective than mediational analyses in examining psychological processes. Journal ofPersonality and Social Psychology, 89, 845–851.

Spezio, M. L., & Adolphs, R. (2007). Politics and the evolving neuroscience literature. In W. R.Neuman, G. E. Marcus, A. Crigler, & M. MacKuen (Eds.), The affect effect: Dynamics of emotionin political thinking and behavior (pp. 71–95). Chicago: University of Chicago Press.

Stenner, K. (2005). The authoritarian dynamic. New York: Cambridge University Press.

Thórisdóttir, H., Jost, J. T., Liviatan, I., & Shrout, P. (2007). Psychological needs and values underlyingleft-right political orientation: Cross-national evidence from Eastern and Western Europe. PublicOpinion Quarterly, 71, 175–203.

Tiedens, L. Z., & Linton, S. (2001). Judgment under emotional certainty and uncertainty: The effectsof specific emotions on information processing. Journal of Personality and Social Psychology,81, 973–988.

van den Bos, K., Poortvliet, P. M., Maas, M., Miedema, J., & van den Ham, E. J. (2005). An enquiryconcerning the principles of cultural norms and values: The impact of uncertainty and mortalitysalience on reactions to violations and bolstering of cultural worldviews. Journal of ExperimentalSocial Psychology, 41, 91–113.

Webster, D. M., & Kruglanski, A. W. (1994). Individual differences in need for cognitive closure.Journal of Personality and Social Psychology, 67, 1049–1062.

Weise D. R, Pyszczynski, T., Cox, C. R, Arnd, J., Greenberg, J., Solomon, S., & Kosloff, S. (2008).Interpersonal politics: The role of terror management and attachment processes in shapingpolitical preferences. Psychological Science, 19, 448–55.

Willer, R. (2004). The effects of government-issued terror warnings on presidential approval ratings.Current Research in Social Psychology, 10, 1–12.
