Why Men (and Women) Do and Don't Rebel: Effects of System Justification on Willingness to Protest

Three studies examined the hypothesis that system justification is negatively associated with collective protest against ingroup disadvantage. Effects of uncertainty salience, ingroup identification, and disruptive vs. nondisruptive protest were also investigated. In Study 1, college students who were exposed to an uncertainty salience manipulation and who scored higher on system justification were less likely to protest against the governmental bailout of Wall Street. In Study 2, May Day protesters in Greece who were primed with a system-­‐‑justifying stereotype exhibited less group-­‐‑based anger and willingness to protest. In Study 3, members of a British teachers’ union who were primed with a “system-­‐‑rejecting” mindset exhibited decreased system justification and increased willingness to protest. The effect of system justification on nondisruptive protest was mediated by group-­‐‑based anger. Across very different contexts, measures, and methods, the results reveal that, even among political activists, system justification plays a significant role in undermining willingness to protest.

anger, and disruptive vs. nondisruptive forms of collective action.

An Integrated Perspective on Collective Protest
According to system justification theory, most individuals possess a (largely nonconscious) motivation to defend, bolster, and justify the status quo (e.g., Jost, Liviatan, van der Toorn, Ledgerwood, Mandisodza, & Nosek, 2010;Kay, Gaucher, Peach, Laurin, Friesen, Zanna, & Spencer, 2009). This motivation, which varies according to both situational and dispositional factors (Jost & Hunyady, 2005), is clearly at odds with support for protest and collective action aimed at changing the extant social system, especially for members of disadvantaged groups who might otherwise push to improve the status of their own group (Jost, Burgess, & Mosso, 2001). Whereas group interest (or group justification motivation) might encourage members of disadvantaged groups to engage in collective protest aimed at helping the ingroup, the endorsement of system---justifying beliefs is predicted to be negatively associated with collective protest on behalf of the ingroup (H1). We test this hypothesis in all three studies using both experimental manipulations and measures of system justification.

Uncertainty Salience
But why would members of disadvantaged groups possess system justification motivation in the first place, given that it conflicts with personal and collective self--interest? One reason is that system---justifying beliefs serve epistemic needs to reduce uncertainty and ambiguity about the social and political world; making peace with the System Justification and Protest 5 status quo allows one to feel certain and secure, whereas rebelling (even ideologically) brings risk and unpredictability (Jost & Hunyady, 2005). Accordingly, many studies show that heightened personal needs for certainty, order, structure, and closure are associated with a clear preference for conservative, system---justifying stereotypes and ideologies (Calogero & Jost, 2011;Jost, Glaser, Kruglanski, & Sulloway, 2003;Jost, Napier, Thorisdottir, Gosling, Palfai, & Ostafin, 2007;Stapel & Noordewier, in press).
Thus, even if social change is seen as desirable by some members of disadvantaged groups, it is inherently unsettling. It is simply not possible for members of any group engaging in collective protest to know beforehand the outcome of their struggle, given that their group's status could improve, worsen, or remain the same once the conflict subsides.
A system justification perspective differs from other theoretical approaches, including uncertainty---identity theory (Hogg, 2005), which treats virtually all belief systems as (equivalently) tied to the satisfaction of epistemic needs (see also Greenberg & Jonas, 2003;McGregor & Marigold, 2003). Indeed, Hogg (2005) has argued that "uncertainty can lead to social stasis or social change" and, more provocatively, that "uncertainty is just as likely to produce system---justifying or hierarchy---enhancing ideologies as system---challenging or hierarchy---attenuating ideologies" (p. 222). By contrast, we hypothesize that increasing the salience of uncertainty will lead to decreased rather than increased support for system---challenging protest behavior (H2). We investigate this System Justification and Protest 6 hypothesis using an experimental manipulation of uncertainty salience in Study 1 and also consider the more speculative possibility that uncertainty salience would interact with system justification levels. The idea is that members of disadvantaged groups who are high in system justification should be unlikely to protest in general, whereas those who are low in system justification might be more willing to protest when uncertainty salience is low rather than high; in other words, support for protest might occur only when uncertainty salience and system justification are both low.

Group Identification and Group---Based Anger
Research on social identity theory indicates that strongly identified group members are especially likely to fight against their (perceived) state of disadvantage by means of collective protest (e.g., Abrams, 1992;Simon & Klandermans, 2001;Tajfel & Turner, 1979;see also O'Brien & Major, 2005). Thus, we would expect on the basis of prior research that group identification would be positively associated with protest behavior among members of disadvantaged groups.
A number of theoretical perspectives, including Gurr's (1970) version of relative deprivation theory, assume that the experience of anger is an important antecedent of rebelliousness. Research confirms that anger in response to felt injustice (i.e., moral outrage) is indeed one of the strongest predictors of participation in collective protest (e.g., Martin, Scully, & Levitt, 1990;Montada & Schneider, 1989;Van Zomeren, Postmes, & Spears, 2008). This emotion is considered to be action---oriented insofar as it elicits System Justification and Protest 7 tendencies to move against the source of disadvantage (Mackie, Devos, & Smith, 2000;Van Zomeren, Spears, Fischer, & Leach, 2004). Therefore, we would expect that group--based anger will be positively associated with collective protest on behalf of the ingroup, as in past research.
It has been suggested that system---justifying beliefs serve the palliative function of decreasing negative affect (including anger), thereby increasing adherents' satisfaction with the status quo (Jost & Hunyady, 2002;Kluegel & Smith, 1986;Napier & Jost, 2008). An experiment by Wakslak, Jost, Tyler, and Chen (2007) used a mindset priming technique to induce a high (vs. low) system---justifying mindset and demonstrated that this mindset decreased moral outrage (i.e., feelings of anger and distress concerning inequality) and willingness to help the disadvantaged. However, this experiment focused on reactions to circumstances of outgroup disadvantage, whereas the present research focuses on reactions to ingroup disadvantage. We would hypothesize that system justification will be negatively associated with group---based anger (H3), which is an important antecedent of willingness to participate in collective action, as noted above. We investigate hypothesis 3 in Studies 2 and 3, using experimental inductions of system justification motivation.

Disruptive vs. Nondisruptive Forms of Protest
Researchers of collective action often distinguish between normative and non--normative forms of protest. However, as highlighted by Simon and Klandermans System Justification and Protest 8 (2001), power struggles that call into question an overall system of authority focus on whether the existing norms are legitimate. For this reason, the normative/non---normative distinction can be ambiguous insofar as ingroup and outgroup norms are likely to be in conflict, and so it is unclear whose norms are being breached (either, neither, or both). It is therefore useful to consider instead whether the actions themselves are disruptive or nondisruptive to the overarching social order (as defined by the system). Disruptive actions, such as strikes and riots, upset the social order and daily routines of ordinary citizens, whereas nondisruptive forms of protest, such as petition---signing and letter--writing, do not.
The purpose of disruptive action is likely to be complex and effortful; it creates highly visible events in order to draw a wider audience's attention to the putative legitimacy of the group'ʹs concerns. Disruptive action, then, represents an extreme end of an action continuum that, according to social identity theory, should arise from a "social change" belief structure among high group identifiers (Abrams & Grant, 2011), that is, when group members strongly believe that a structural or system change is required in order for their group to attain a just outcome. Disruptive action is likely to be taken by people who value and identify strongly with their own group and can articulate an ideology that explains why the system is unjustified (Jost et al., 2001). For them, disruptive action, such as participating in a violent demonstration, is a rational response to a broken system (cf. Reicher & Hopkins, 2001), even though it might well produce deleterious personal consequences such as being arrested or injured. The goals of disruptive action may be quite distal, such as challenging the rights of authorities to define the rules or establishing the existence of the group as a coherent entity, creating a shared experience, and reinforcing social identification (cf. Smith, Seger, & Mackie, 2007), thereby enhancing the basis for further collective action (see also Reicher, Spears & Postmes, 1995).
In contrast, nondisruptive action, such as petition---signing and letter---writing, probably serves a more proximal goal, such as venting anger and frustration or trying to persuade those in power to concede a specific point of argument or action. Thus, we believe that nondisruptive action, being less costly and easier to enact, may be a more readily elicited by emotion---based response to injustice. Insofar as system justification reduces feelings of frustration and anger (see also Jost, Wakslak, & Tyler, 2008), the temporary activation of system justification motivation should alleviate the kind of emotional distress that is associated with nondisruptive protest. That is to say, system justification should be negatively associated with nondisruptive protest behavior to the extent that it eliminates short---term feelings of frustration and anger that might otherwise lead someone to sign a petition or engage in other nondisruptive reactions.
Differences in the antecedents of disruptive and nondisruptive protest behavior can be observed in previous research (e.g., Chaikalis---Petritsis, 2010;Rojas, 2007;Simon & Klandermans, 2001). For example, a study of anti---globalization protestors by System Justification and Protest 10 Cameron and Nickerson (2006) found that social dominance orientation (the ideological endorsement of group---based hierarchical systems) was negatively associated with participation in nondisruptive forms of protest, but it was unrelated to or disruptive protest. Furthermore, Tausch, Becker, Spears, Christ, Saab, Singh, and Siddiqui (in press) discovered that anger predicted willingness to participate in nondisruptive protest but was unrelated to participation in disruptive protest. Thus, to the extent that (a) social dominance orientation is a system---justifying ideology and (b) system justification undermines support for collective protest by decreasing group---based anger (or moral outrage), we might expect that system justification will exert a negative (indirect) effect on nondisruptive (but not necessarily disruptive) protest through group---based anger (H4).
We assessed this mediational hypothesis in Study 3.

Summary of Hypotheses and Research Overview
To sum up, we investigated the following novel hypotheses in three studies: (H1) System justification will be negatively associated with collective protest.
(H2) Uncertainty salience will be negatively associated with collective protest.
(H3) System justification will be negatively associated with group---based anger.
(H4) The negative effect of system justification on (nondisruptive) collective protest will be mediated by group---based anger.
In Study 1, we explored college students' reactions to the controversial governmental bailout of Wall Street financial firms in 2008. Arguably, such bailouts System Justification and Protest 11 constituted "upward redistribution" from U.S. taxpayers to large multinational corporations that subsequently rewarded their executives with enormous bonuses.
According to Pfeffer (2010), a system justification perspective can help to explain the lack of moral outrage that followed the global financial crisis of 2008: [System justification theory] helps us understand why there is less anger and pressure to change regulatory oversight than might be expected by the economic disaster visited on millions of Americans by events not of their own doing. To feel good about America and its system, people make sense of what happened in ways that do not undermine the legitimacy of the existing status order, including the huge salaries earned by people working in finance.
Thus, we hypothesized that system justification and uncertainty salience would be negatively associated with protest against the bailouts. Studies 2 and 3 were field experiments involving real---world political activists; we manipulated system justification motivation and measured willingness to participate in various forms of collective action. In Study 2, we exposed protestors at a May Day demonstration in Greece to complementary vs. noncomplementary stereotypes (Kay & Jost, 2003) and measured their levels of system justification, anger, and support for disruptive and nondisruptive protest. In Study 3, we employed a new "system---rejecting" mindset priming technique in the context of a survey of schoolteachers on strike in the United Kingdom, measuring their subsequent levels of system justification, ingroup identification, anger, and support for disruptive and nondisruptive protest. To the extent that the various operationalizations of theoretical variables in different studies yield similar patterns of results, this research program as a whole will provide convergent evidence that general processes of system justification are at work and are not attributable to specific features of the groups or contexts under investigation.

Study 1: Reactions to the Wall Street Bailout
In our first study we sought to investigate the effects of system justification and uncertainty salience on support for collective protest in the context of the U.S. government's decision to bail out Wall Street. We measured system justification and manipulated uncertainty salience using a mindset priming technique. This study also enabled us to compare effects on disruptive and nondisruptive forms of protest.

Method
Participants. One---hundred and eight New York University students (58% female, mean age = 20 years, SD = 1.51) completed a paper---and---pencil questionnaire for partial course credit. the uncertainty manipulation had no effect on either subscale, both t's < 1. Affect was also unrelated to support for protest (see Table  1), so it cannot explain the effects of uncertainty salience on protest tendencies.
All participants read an article adapted from the New York Times entitled "U.S.
Expands Plan to Buy Banks' Troubled Assets," including the following passage: WASHINGTON - The US administration's new plan to liberate the nation's banks from a toxic stew of bad home loans and mortgage--related securities is bigger and more generous to private investors than expected, but it also puts taxpayers at great risk. Taken  Geithner acknowledged at a briefing for reporters. "The question is how best to do it." Afterward, participants indicated their willingness to engage in disruptive and nondisruptive forms of protest against the government bailout and completed manipulation checks, as described below. Willingness to protest. On 7---point scales (ranging from 1 = not at all, 7 = very much so) we asked participants how willing they were to engage in nondisruptive ("I am willing to send with fellow NYU students a letter/email message of protest to the government") and disruptive ("I am willing to take part in occupying an NYU building as a sign of protest") forms of protest. Manipulation checks. Following Van den Bos et al. (2006), we administered four manipulation check items, two that were specific to uncertainty salience (α = .97) and two that were specific to television salience (α = .99). Participants assigned to the System Justification and Protest 15 uncertainty salience condition did indeed report higher levels of uncertainty (M = 6.25, SD = .79) than did those in the control condition (M = 3.08, SD = 1.67), t (71.29) = 12.46, p < .001. In addition, participants assigned to the television salience condition reported thinking about television more (M = 5.66, SD = 1.40) than did those in the uncertainty salience condition (M = 1.22, SD = .73), t (75.67) = ---20.43, p < .001.

Results and Discussion
Descriptive statistics are summarized in Table  1. To investigate the effects of system justification and uncertainty salience on willingness to engage in disruptive and nondisruptive protest, we conducted a path model using MPLUS 6.1 (Muthén & Muthén, 1998---2010. The experimental variable was effect---coded as 1 for the uncertainty salience condition and ---1 for the control condition. Economic system justification was centered at its mean. We allowed the residuals for nondisruptive and disruptive protest to correlate (b = .48, SE = .18, ß = .26, z = 2.60, p < .01), because they were assumed to share predictors that were not included in the model. were both negatively associated with support for disruptive protest (i.e., willingness to engage in building occupation). These results corroborate hypotheses 1 and 2, respectively. The interaction between the two predictor variables was not reliable, b = .04, SE = .13, ß = .03, z = .31, p = .76.
This finding contradicts the notion that uncertainty salience would encourage system--challenging behavior for those who were prone to such behavior, namely those who are chronically low in system justification (e.g., Hogg, 2005).
It is also possible to contrast the simple effects of system justification in the two experimental conditions. The effect of economic system justification on nondisruptive protest was negative and marginally significant in the control condition, b = ---.61, SE = .34, ß = ---.27, z = ---1.81, p = .07, but it was not significant in the uncertainty salience condition, b = .35, SE = .28, ß = .15, z = 1.22, p = .22. Thus, it appears that the uncertainty manipulation was strong enough to overcome individual differences in system justification-but only with respect to nondisruptive protest. Even low system justifiers were unlikely to register a complaint under circumstances of high uncertainty salience.

Study 2: May Day in Greece
System Justification and Protest 17 In the first study, we found that system justification scores were negatively associated with support for protest, as hypothesized. In Study 2, we sought to manipulate system justification motivation directly by exposing participants to complementary (e.g., "poor but happy") stereotype exemplars, adapting methods used by Kay and Jost (2003). To increase external as well as internal validity, we conducted our experiment in the field, in the context of a May Day demonstration in Greece. In addition to manipulating system justification motivation, we measured ingroup identification and group---based anger.

Materials and procedure. One male and one female experimenter approached
individual protestors on a random basis. Following Kay and Jost (2003) participants first read about a poor individual who was described as either happy or unhappy: Nick is from a large Greek city. He is married and has two children, has brown hair, and is 1.80m tall. Nick liked to play football as a child and still closely

System Justification and Protest 18
After reading this passage, participants rated how likely or unlikely they thought it was that Nick was arrogant, funny, generous, content, socially competent, fulfilled, likeable, and modest (using 9---point scales). The contentment ratings were used to check on the manipulation of perceived happiness; the others were administered to strengthen the cover story. Participants also indicated their level of agreement on a 9---point scale with a single item from the Kay and Jost (2003) scale that we considered most relevant: "Most governmental policies serve the greater good." Next participants answered questions concerning a pension bill that the Greek government was seeking to pass; the bill sought to increase the age at which Greek citizens would have the right to draw a state pension. We focused questions on that bill because its announcement immediately preceded May Day, so the majority of demonstrating organizations made the governmental bill their focal issue. We asked protestors how angry they were with the government, how much they identified with the other protestors gathered at the rally, and how willing they were to protest against the bill (in both disruptive and nondisruptive ways).
Because of time constraints and the unique circumstances of the study, we relied on single items to measure moderating and dependent variables. For instance, we asked participants: "How willing are you to do the following in order to protest against the pension bill?" Response options were "Sign a petition" (nondisruptive) and "Take part in occupying a public building" (disruptive). To measure ingroup identification we System Justification and Protest 19 asked participants the extent to which they agreed that: "Being part of the people gathered here today is important to me." To tap anger we asked: "When I think about the pension bill the government is trying to pass, I feel angry." Responses were provided on 9---point scales (1 = not at all, 9 = very much).
Finally, participants reported their gender, age, political orientation, and organizational membership (e.g., whether they were members of any of the trade unions or political parties attending the march; we coded 1 for "yes" and 0 for "no" and asked them to specify which organization(s) if they replied "yes"). Fifty---two percent of the respondents belonged to some trade union or political party. We measured political orientation on a continuum ranging from 1 (left---wing) to 9 (right---wing). A strong majority (87.5%) circled a number that was left of the scale midpoint (M = 2.79, SD = 1.61). Given that May Day is a traditionally leftist, working---class holiday, this skew was expected. We decided to expose participants to the "poor but happy" stereotype exemplar rather than the "poor but honest" exemplar because Kay, Czaplinski, and Jost (2009) found that the former affected leftists' levels of system justification, whereas the latter affected rightists' system justification scores. In all analyses, we adjusted for main effects of age, gender, political orientation, and organization membership.

Results and Discussion
Manipulation checks. To determine whether participants attended to the vignette, To determine whether the manipulation affected participants' system justification scores, we conducted an additional ANCOVA. As anticipated, participants assigned to the "poor but happy" condition scored higher on system justification (M = 1.53, SE = .13) than did participants assigned to the "poor and unhappy" condition (M = 1.06, SE = .13), F (1, 17) = 6.18, p < .05, η² = .27, although both groups scored low in absolute terms. participants assigned to the "poor but happy" condition were slightly less likely to sign a petition (M = 7.86, SE = .47) than those assigned to the "poor and unhappy" condition (M = 8.61, SE = .50).

Effects on anger, group identification, and willingness to protest. We conducted a one---
Correlations among study variables. In terms of correlational results (see Table  2), we observed that anger was positively and significantly associated with nondisruptive protest, r(23) = .56, p < .01, and marginally significantly with disruptive protest, r(23) = .36, p = .08. Similarly, group identification was positively and significantly associated with nondisruptive protest, r(23) = .69, p < .001, and marginally significantly with disruptive protest, r(23) = .36, p = .08. Because of the small sample size and the use of System Justification and Protest 22 single---item measures, we were unable to conduct a proper test of Hypothesis 4, which posits that anger mediates the negative effect of system justification on willingness to protest (cf. Wakslak et al., 2007). Thus, we conducted an additional experiment with multiple---item measures and a larger "real---world" sample, namely members of a British school teachers' union that was on strike.

Study 3: A Teachers' Strike in Britain
To investigate the effects of system justification, ingroup identification, and anger on support for disruptive and nondisruptive protest, we conducted a web---based experiment involving a national teachers' union in the U.K. Specifically, we manipulated system justification motivation using a new experimental technique and measured ingroup identification, group---based anger, and willingness to protest. This study provided our best opportunity to integrate system justification theory with other perspectives on collective action on behalf of one's own group, including social identity theory.

Method
Participants. Fifty---nine members (80% female; mean age = 33.7, SD = 8.21) of the National Union of Teachers (NUT) volunteered to take part in a web---based experiment.
Participants were notified of the study through an email sent by a union representative a few days before the strike. Materials and procedure. Participants were randomly assigned to one of two experimental conditions: a system rejection condition and a control condition. The experimental manipulation was disguised as a "warm---up" question that would allow participants to think about some issues before they answered a number of questions about the NUT strike. Participants assigned to the system rejection condition read the following instructions: Please think about ways that things are organised or arranged politically, legally, socially or economically in the UK. Which of these things would you strongly recommend other countries should NOT follow because they work particularly badly and are bad ways to run things for the country as a whole? You might think of laws, policies or institutions such as parliament, employment, education, family, social norms and roles, cultural traditions or religion. Please write a few lines about one of these things that you would recommend other countries should NOT follow and why.
Participants assigned to the control condition instead read the following: Please think about different teaching methods you have used with primary or secondary school children. Which of these methods would you strongly recommend other teachers should NOT follow because they work particularly badly for children'ʹs learning? Please write a few lines about one of these methods that you would recommend other teachers should NOT follow and why.

System Justification and Protest 24
Inspection of the open---ended responses led to the conclusion that in all cases participants generated criticisms of the appropriate target. That is, participants assigned to the system rejection condition wrote about circumstances in the U.K., whereas participants assigned to the control condition wrote about educational methods.
Next participants completed Kay and Jost's (2003) 8---item system justification scale (α = .80). Afterward, they rated their identification with the ingroup, their level of anger, and their intentions to engage in disruptive and nondisruptive forms of protest.
Willingness to protest. We administered four items to measure support for nondisruptive protest. Each began with the prompt "I intend to . . ." and ended with: (1) "sign a NUT petition to be sent to the government," (2) "send a letter or email message with my NUT colleagues to my local MP," (3), "attend a local NUT meeting to discuss teachers' pay and the strike action," and (4) "hand out leaflets to the public with NUT colleagues" (α = .82). These items were based on a list of protest behaviors suggested by the union (NUT, 2008) as well as items used by Leach, Iyer, and Pedersen (2007).
A pretest involving a small sample of NUT members (N = 18) confirmed that the System Justification and Protest 25 nondisruptive behaviors were regarded as nondisruptive and that the disruptive behaviors were seen as disruptive to the functioning of schools.
Ingroup identification. We administered a 5---item measure of identification with the union derived from Van Zomeren et al. (2004) and Kessler and Hollbach (2005): (1) "I view myself as a member of the NUT," (2) "I am glad to be a member of the NUT," (3) "I feel connected to other NUT members," (4) "I am proud to be a member of the NUT," and (5) "Being a NUT member is important to me" (α= .93).

Anger. We tapped into anger against the government with two items adapted
from Van Zomeren et al. (2004): "Because of the government's position on teachers' pay, I feel … angry" and "frustrated" (α= .78).

Results and Discussion
Means, standard deviations, and intercorrelations are presented in Table  3.
Inspection of this table reveals that random assignment to the self---generated system rejection condition was indeed associated with increased willingness to engage in both disruptive and nondisruptive forms of protest. It was also associated with decreased system justification and increased anger, as expected. System justification was negatively correlated with anger, as in previous studies (e.g., Wakslak et al., 2007).
We hypothesized that assignment to the self---generated system rejection condition would lead to decreased system justification, which should be associated with greater anger. Anger, in turn, was expected to foster the motivation to participate in nondisruptive protest but not necessarily disruptive protest (see Tausch et al., in press).
To test these hypotheses, we performed path analyses in MPLUS 6.1 (Muthén & Muthén, 1998---2010. The system rejection condition was coded as 1 and the control condition as ---1. The model included a test of the relationship between system rejection and anger, as mediated by system justification and group identification. In addition, the model investigated the extent to which system justification and group identification predicted both forms of protest, and whether these relationships were mediated by anger. We employed the bootstrapping technique outlined in Preacher and Hayes (2008); this method is believed to be better than standard ways of estimating mediation, as with the Sobel Test (Shrout & Bolger, 2002). We requested 95% bias---corrected confidence intervals using 5,000 resamples. The residuals of nondisruptive and disruptive protest tendencies were correlated, as in the previous studies, as were the residuals for system justification and group identification.

System Justification and Protest 28
Finally, we tested for mediational effects using a bootstrapping analysis. We found that system justification mediated the effect of assignment to the system rejection condition on anger. Because the interval of {.03, .46} did not contain 0, the indirect effect was considered to be significant. For ingroup identification, a confidence interval of {.02, .43} was obtained, also indicating significant mediation. These findings are consistent with the interpretation that system rejection decreased system justification and increased group identification, and these two variables displayed opposite relations with anger. Anger did not significantly mediate the

General Discussion
This research program constitutes the first direct empirical investigation of the causal effects of system justification (and, to our knowledge, uncertainty salience) on willingness to protest on behalf of the ingroup. Using a diverse range of settings, methods, and samples from the U.S., Greece, and Britain, the present set of studies has produced generalizeable evidence for all four hypotheses. Specifically, we found that system justification is negatively associated with anger and willingness to protest-and that anger mediates the effect of system justification on willingness to protest. As hypothesized, uncertainty salience was also associated with decreased motivation to engage in disruptive protest and (when system justification was low) nondisruptive protest. No support was obtained for the alternative hypothesis that uncertainty salience would increase protest levels among low system---justifiers (Greenberg & Jonas, 2003;Hogg, 2005).
Taken as a whole, these findings speak to the utility of incorporating system justification processes when it comes to explaining collective forms of protest (cf. Simon & Klandermans, 2001). Our studies build on previous research indicating that belief in a just world (Hafer & Olson, 1993;Rubin & Peplau, 1973) and social dominance orientation scores (Cameron & Nickerson, 2006) are negatively correlated with willingness to protest. In Study 1 we observed that individual differences in system justification were also negatively correlated with disruptive forms of protest in the context of the Wall Street bailout (see also Pfeffer, 2010). In Studies 2 and 3 we demonstrated that situational inductions of system justification motivation are capable of undermining support for protest-even among people who are otherwise highly motivated to engage in collective action, namely protesters and union members.
The results we have obtained substantiate previous calls to distinguish between disruptive and nondisruptive forms of protest (e.g., Cameron & Nickerson, 2006;Chaikalis---Petritsis, 2010;Rojas, 2007). Although the two dependent variables were positively intercorrelated, and the general patterns of effects were similar for disruptive and nondisruptive forms of protest, we did find that system justification was more strongly associated with a reluctance to engage in disruptive protest in Studies 1 and 2.
At the same time, we replicated Tausch et al.'s (in press) findings that anger was significantly related to nondisruptive protest but unrelated to disruptive protest. Thus, in Study 3 we observed that the indirect effect of system justification through anger (the so---called "palliative effect") was present only with respect to nondisruptive protest. In the context of the teachers' strike, willingness to engage in disruptive protest was driven by identification with the union rather than anger (or system justification).
In addressing Gurr's famous question of why men (and presumably women as well) rebel against governmental and other authorities, it is probably not enough to say that, "men are quick to aspire beyond their social means and quick to anger when those means prove inadequate" (Gurr, 1970, p. 58). Or that: "Feelings of illegitimate inequality or injustice typically result when social comparisons reveal that one's in---group is worse off than relevant outgroups" (Simon & Klandermans, 2001, p. 324). These generalizations are far truer with respect to those whom we would describe as low (as opposed to high) "system---justifiers." We have found that when system justification motivation is heightened either chronically or temporarily-that is, following situational inductions of uncertainty salience or complementary stereotype exposure, or in the absence of a "system---rejecting" mindset-individuals are unlikely to take System Justification and Protest 31 meaningful action against sources of disadvantage. Thus, a deeper understanding of the motivational antecedents of willingness to protest, which incorporates what we know about system justification processes as well as factors such as group identification and group---based anger, is needed. This research, we think, takes one step in the right direction, but we are still some ways from being able to predict the occurrence of revolutions and regime change on the basis of social psychological theories alone. In the long run, it will be necessary to connect psychological, sociological, political, economic, historical, and cultural levels of analysis. Only then will we really understand what inspires the "rabble to riot."