This is an open access article distributed under the terms of the Creative Commons
Attribution License (Organizations and societies are in constant transition, evolving and changing from one form to another. At the same time, people usually prefer keeping things the way they are, and resist change (Eidelman, Crandall, & Pattershall, 2009; Jost, Banaji, & Nosek, 2004). As the default ideological position is conservative and status quo maintaining (Skitka, Mullen, Griffin, Hutchinson, & Chamberlin, 2002), proponents of change are often met with resistance (Jost et al., 2004), and are negatively perceived (Bashir, Lockwood, Chasteen, Nadolny, & Noyes, 2013). According to system justification theory, the motivation to prefer the status quo is similar to the motivation to enhance oneself and one’s ingroup, and thus is social in nature (Jost et al., 2004). Moreover, humans are cognitively biased to prefer what is known to something new and less known (Eidelman et al., 2009). This implies that it may take more of an effort and a risk to challenge the status quo than to support its existence.
Differing attitudes towards the status quo could be seen as boundaries between groups of different opinions. Previous research indicates that similarity in opinions or attitudes may function as group boundaries (Kenworthy & Miller, 2002), eliciting typical intergroup biases such as ingroup favoritism and outgroup derogation (Pettigrew, 1979; Robbins & Krueger, 2005; Tajfel & Turner, 1986). Specifically, in cases where attitudes constitute group boundaries people typically attribute negative reasons for opponents’ attitudes and positive reasons for allies’ attitudes (Bäck, 2013; Kennedy & Pronin, 2008; Kenworthy & Miller, 2002; Reeder, Pryor, Wohl, & Griswell, 2005). In previous research on perceptions of attitude opponents, perceivers’ position toward the status quo has not been taken into account. However, given that the social risks associated with challenging versus defending the status quo may differ, it seems reasonable that attitude position may be a significant moderator of this type of biases.
Since biased perceptions of opponents in a situation where opinions differ may escalate into conflict (Kennedy & Pronin, 2008), it is important to understand how individuals construct mental representations of allies and opponents, and also which mechanisms might be driving such representations. Thus, the overall aim of the present research was to investigate how attitude position on an issue affects people’s mental representations of attitude allies and opponents.
Position to the Status Quo and Intergroup Biases
Despite the fact that the default position is status quo maintaining and proponents of change are usually negatively stereotyped (Bashir et al., 2013), some individuals and groups, such as the Black Power revolution, or Womens’ Right activists, still choose to challenge the status quo and fight for their cause. However, this choice does not come without social costs. Studies show that advocates of change may be accused of being driven primarily by self-interest (O’Brien & Crandall, 2005). Moreover, people evaluate challengers of the status quo as less flexible, less reasonable, more contributing to the conflict (Kray & Robinson, 2001), and more extreme than defenders (Keltner & Robinson, 1997). Since challengers of the status quo are generally viewed negatively (Bashir et al., 2013; Crandall, Eidelman, Skitka, & Scott Morgan, 2009), it is possible that being the object of such negative perceptions may impact the individuals’ own view of themselves and others. For example, Kray and Robinson (2001) have suggested that challengers are likely to be aware of the negative perception of them, which may lead to an increased need to defend and justify their own position, by placing them in a vulnerable position.
The Need to Justify Own Position
Several authors argue that challengers are in general in a more vulnerable position than defenders (De Dreu, Kluwer, & Nauta, 2008; O’Brien & Crandall, 2005). Such a vulnerable and exposed position should also imply an increased need to justify or defend one’s position, for example by undermining the position of one’s opponents. Although there are no studies investigating challengers’ perceptions of their attitude allies and opponents, there are indications that challengers do display increased intergroup biases. For example, in their comparative study on conflict resolution in symmetrical and asymmetrical dyad conflict structures, De Dreu and colleagues (2008, Experiment 2) found that in asymmetrical structures (where one party wants to change and one party wants to keep the status quo), challengers of the status quo perceived defenders as less friendly and more dominant than defenders perceived challengers. Thus, we expected challengers to display stronger intergroup biases by ascribing more negative reasons for opponents’ than allies’ attitude position than defenders, and that the need to justify one’s position is one potential explanation to this asymmetry.
Social Identity and Affect
Being a challenger of the status quo usually implies that one advocates one specific issue (e.g., women’s rights). This could lead challengers to an increased sense of ‘groupness’ or entitativity (Campbell, 1958), which, according to self-categorization theory, leads people to an increased differentiation of in- and out-groups on important dimensions (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987).
When identification with the own group becomes salient, group based appraisals elicit specific emotions (Mackie, Devos, & Smith, 2000). According to Intergroup Emotions Theory, negative emotions to an outgroup could be elicited by ingroup saliency, and predict negative behavior tendencies to the group (Mackie et al., 2000). Appraisals are cognitive evaluations or interpretations of an emotional event or object (Mackie et al., 2000). As such, appraisals are, in a sense, evaluations of the ingroup and outgroup, and can be captured in perceptions of these groups. Thus, when group membership becomes salient and a (threatening) outgroup becomes salient, emotions are automatically experienced which feed into the explicit evaluations of the two groups. In addition to a greater need to justify one’s position, another psychological mechanism explaining why challengers may be more negative to their opponents could then be differences in experienced emotions elicited by an increased sense of group identification among challengers as opposed to defenders. Hence, we propose a model for explaining why position to the status quo may increase biases of allies and opponents. In this model, we integrate two possible explanatory variables, the need to justify one’s position, and the possible increase in negative affect that should be experienced by challengers.
Position to the Status Quo and Numerical Status
In previous research it has been shown that perceived threat to social identity leads to increased biases (Branscombe & Wann, 1994; Kenworthy & Miller, 2002; Stephan, Boniecki, Ybarra, Bettencourt, Ervin, Jackson, & McNatt, 2002). The role of threats in the racial attitudes of blacks and One factor associated with feelings of threat is the perceived numerical status of one’s group. Kenworthy and Miller (2002) showed that when participants were told they were in the minority, they displayed stronger intergroup biases of the perceived attitude origins, or reasons for why an individual holds a certain attitude. Based on the notion of false consensus (Kelley, 1967), the belief that the preference held by the majority is the objectively correct one, Kenworthy and Miller argued that knowing that one belongs to the minority may increase the need to defend and justify one’s position hence increasing intergroup biases (Kenworthy & Miller, 2002). Holding an “incorrect” preference thus threatens one’s positive self-image (Festinger, 1954) and one’s positive ingroup image (Tajfel & Turner, 1986); in the present situation that means one’s opinion allies. Manipulations of numerical status have been successfully used to influence group-based biases in several empirical studies (Bettencourt, Miller, & Hume, 1999; Corneille, Yzerbyt, Rogier, & Buidin, 2001; Kenworthy & Miller, 2001; Kenworthy & Miller, 2002; O’Brien & Crandall, 2005; Otten, Mummendey, & Blanz, 1996).
Drawing on the literature that presents the status quo as the default, one could argue that being a challenger could also imply being in the minority. In line with this, Keltner and Robinson (1997) showed that both challengers and defenders of the status quo underestimated the actual prevalence of challengers and overestimated the actual prevalence of defenders. Thus, when studying the possible effects of status quo on perceptions of allies and opponents, it is important to take numerical status into account.
Overview of Studies
The main objective of the current research was to investigate the role of attitude position on biased views of attitude allies and opponents. Specifically, we tested the hypothesis that challengers of the status quo should attribute more negative reasons for their opponents’ attitudes, and more positive reasons for their allies’ attitudes, than the other way around. We tested this in three studies, using a range of social issues.
We have not limited challengers and defenders of the status quo to a specific attitude issue, but rather adopted a higher-level definition in that any individual choosing a standpoint that opposes the status quo is defined a challenger and any individual choosing a standpoint in agreement with the status quo is defined a defender. This means that neither challenging nor defending the status quo should be overtly associated with positive or negative connotations in a general sense.
In Study 1, we investigated the main hypothesis that challengers are more biased than defenders. In Study 2, we extended Study 1 by including several social issues, controlling for the possible confound of positive or negative connotations attached to the challenging/defending position. In addition, we added an experimental manipulation of numerical status to control for implied numerical status as another possible confound. In Study 3, we aimed to replicate Study 1 and 2 by using another social issue, and another numerical status manipulation. In addition, we added possible mediating variables and tested our overall model.
Study 1
To examine how status quo challengers and defenders attribute the attitude origins of their allies and opponents, participants in Study 1 were presented with a hypothetical scenario regarding a decision on a potentially important issue. The issue should be involving to the participants, while at the same time preference for or against the status quo should vary considerably. Hence, in the present study, high school students were asked if teachers should be allowed to confiscate students’ cell phones during class hours in cases where they disturb the class. This issue has been debated somewhat in Sweden (the country in which the studies were conducted) during the past years, as young people increasingly use their cell phones during class. The status quo maintaining (i.e., defending) position was to not allow this.
In line with previous research (Bäck, 2013; Kenworthy & Miller, 2002) we expected that both challengers and defenders would show intergroup biases in attitude attributions, and ascribe more favorable origins of preferences to those who agree with them as opposed to those who disagree with them. However, we expected this bias to be magnified among challengers of the status quo.
Method
Participants and Design
Participants were 311 high school students (136 males, 162 females, and 13 who did not state a gender) from the larger Stockholm area. Mean age was 17 years (SD = 0.81), ranging from 16 to 20 years. The design was a 2 (target group: agree/disagree with own preference) × 2 (position: challenging/defending the status quo position) mixed design with target group being the repeated factor and position varied between participants. Position was self-selected and not manipulated. The dependent variables were ratings of the origins of preferences of those who agreed and disagreed with own preference for both challengers and defenders. The distribution of preferences was 208 defenders (67.0%), and 100 challengers (32.0%). Three (1.0%) participants did not state a preference. These were excluded from further analyses.
Procedure and Variables
The data collection took place in participants’ schools and was integrated into their lessons. The study was part of a larger project and was introduced as a study on decision-making in school. The experimenter handed out the questionnaires, which began with a brief description of the issue of whether teachers should be allowed to confiscate students’ disturbing cell phones during class hours, or not. To make participants elaborate on the issue, they were instructed to list three values from a list of values provided to them (or make up their own) that they believed should be considered in relation to keeping or changing the status quo (e.g., to create a good learning environment, that each individual should be respected in school, that the school should be characterized by tolerance, etc.). The values that were given as examples were balanced so that there were both positive and negative aspects attached to both positions in order to make sure that one position should not per se be associated with positive or negative connotations. After some unrelated items, participants stated their own position in the issue (challenging or defending the status quo). Following this were further unrelated items, and then participants were asked to argue their position. Participants were told that this was ostensibly to be used in an upcoming study on persuasion. However, its actual purpose was rather to increase participant involvement. This was followed by some more unrelated items. Then participants were instructed to answer questions about the origins of the preferences of (a) students who shared their preference on the issue of confiscating disturbing cell phones (their allies), and of (b) students who did not share their preference (their opponents). Following the procedures first outlined by Kenworthy and Miller (2002), and later used by Bäck and colleagues (2010), the questions included four of the items measuring the rationality and externality dimensions of origins of attitudes derived by Kenworthy and Miller (2002). These items constituted the main dependent variables in the study. Two externality ratings asked participants to indicate to what extent they perceived the target as being influenced by (a) friends and family (i.e. the opinion of those who have [a different preference than I do/the same preference as I do] is influenced by what their family and friends prefer), and by (b) the potential disapproval of others if the target had the opposite preference. Correlations for these items were .53 and .66 (ps < .001) for the ingroup and outgroup ratings respectively. The two rationality items assessed to what extent they perceived that the target based their preference on (a) careful consideration of decision alternatives and (b) on their own, independent thinking about the issue. The items were assessed on a 7-point scale (1 = strongly disagree to 7 = strongly agree). The correlation between these items were .56 and .53 (ps < .001) for the ingroup and outgroup ratings respectively. The emotionality dimension was excluded in the current research since it yielded only weak effects in the Kenworthy and Miller (2002) study, and its implications are unclear as stated by Kenworthy and Miller themselves.
Results
Target Group and Position Effects
We conducted two repeated measures Analyses of Variance (ANOVAs), one for each dimension (rationality and externality), using position to the status quo as an independent, between participants factor (challenger/defender), and target group (allies/opponents) as a repeated factor. These ANOVAs showed, first, that there was a significant main effect of target group, such that allies were rated as more rational and less external than were opponents, regardless of participants’ own position, F(1, 294) = 46.56, p < .001, ηp2 = .14, and F(1, 298) = 37.38, p < .001, ηp2 = .11, for rationality and externality respectively. Means and standard deviations are shown in Table 1. These analyses thus confirm earlier demonstrations (Bäck, 2013; Bäck et al., 2010; Kenworthy & Miller, 2002) that people in general explain the opinions of those who agree with them in a more positive way than the opinions of those who disagree. Moreover, the ANOVAs revealed a significant interaction effect between target group and own position, such that challengers showed stronger intergroup biases than did defenders, F(1, 294) = 20.25, p <.001, ηp2 = .06, and F(1, 298) = 22.57, p < .001, ηp2 = .07, for rationality and externality, respectively. As can be seen in Table 1, this pattern held true for both dimensions. To test simple effects we performed two separate repeated measures ANOVAs for each dimension testing the difference in ratings of allies and opponents within each position (challenger/defender). F-values were calculated using the error term from the full factor ANOVAs, and p-values were based on the degrees of freedom from the full ANOVAs. Results showed that for the rationality dimension, challengers significantly differed in their ratings of the in- and outgroup, F(1, 294) = 46.86, p < .001, as did defenders, the latter however to a lesser extent, F(1, 294) = 4.10, p = .04. For the externality dimension, challengers again differed in their ratings of the in- and outgroup, F(1, 298) = 42.18, p < .001, while defenders did not significantly differ so, F(1, 298) = 1.79, p = .18. Hence, being a challenger seemed to elicit stronger intergroup biases than being a defender of the status quo.
Table 1
Means and Standard Deviations of Rationality and Externality Ratings for Challengers and Defenders of the Status Quo
| Study 1 (N = 311)
|
||||||
|---|---|---|---|---|---|---|
| Challengers
|
Defenders
|
|||||
| Allies | Opponents | Diff.a | Allies | Opponents | Diff.a | |
| Rationality | 5.15 (1.11) | 4.19 (1.40) | 0.96*** | 4.92 (1.24) | 4.72 (1.30) | 0.20* |
| Externality | 2.66 (1.11) | 3.63 (1.55) | -0.97*** | 2.82 (1.36) | 2.96 (1.40) | -0.14 |
| Study 2 (N = 135)
|
||||||
| Challengers
|
Defenders
|
|||||
| Allies | Opponents | Diff.a | Allies | Opponents | Diff.a | |
| Rationality | 4.73 (1.07) | 3.89 (1.15) | 0.85*** | 4.72 (1.24) | 4.32 (1.27) | 0.41* |
| Externality | 3.26 (1.29) | 3.84 (1.45) | -0.58*** | 3.75 (1.23) | 3.75 (1.42) | -0.01 |
Note. Rationality and externality rated on 7 point Likert-scales, where higher values indicate more ascribed rationality and externality. Higher values on the rationality and lower values on the externality difference indices imply more bias.
aSimple effects.
*p < .05. **p < .01. ***p < .001, two-tailed tests.
Discussion
Previous research has shown that the default ideological position is conservative, and that challengers tend to be perceived more negatively than defenders by observers (Crandall et al., 2009; O’Brien & Crandall, 2005). In the current study, we demonstrated that asking those identifying themselves as challengers and defenders to make judgments of their allies and opponents moderated this pattern. In line with studies on attribution of attitude origins (Bäck, 2013; Bäck et al., 2010; Kenworthy & Miller, 2002), we found that both challengers and defenders of the status quo attributed more rational reasons for the attitudes of those with whom they agreed, and more external reasons for the attitudes of those with a different opinion.
In the current study, we further found that challengers undermined, by ascribing more externality and less rationality, the position of defenders to a larger extent than defenders did of challengers’ position. The results also extend findings by De Dreu and colleagues (2008) in showing that challengers perceive defenders more negatively than the other way around not only in dyad interactions but at the group-level as well.
However, since the default position is status quo maintaining, it is possible that position to the status quo also implies numerical position. Specifically, challengers may perceive themselves to be in the minority, which could be perceived to be threatening and also elicit increased intergroup differentiation (Kenworthy & Miller, 2002). Hence, our finding of increased biases among challengers may not have been an effect of position per se, but rather an effect of the perceived numerical status. It should also be noted that on the particular attitude issue examined in the current study, advocating a change (allowing teachers to confiscate students’ properties) may be associated with being in favor of increased social control, which is generally associated with conservatism. Although it has previously been shown that general political orientation does not explain effects of position to the status quo on intergroup biases (Bäck, 2013), this is a further potential confound that needs be ruled out before drawing strong conclusions.
Study 2
Study 2 was designed to disentangle the effects of numerical status and position, and by using an array of different social issues we sought to eliminate political orientation as another possible confound. In addition to the cell phone issue, we added an issue that has been widely discussed in many European countries lately; whether students in school should be prohibited from wearing religious symbols, such as a veil or a necklace with a crucifix, or not. Here, a prohibition could be viewed as an act of racism, but could also be seen as a way to create more equality in school. Thus, we took care that neither position should have solely positive or negative connotations. The third issue concerned gay rights, specifically whether gay couples should be allowed to adopt children from foreign countries. Being a proponent of this position implies high levels of tolerance for non-normative ways of living, while opposing this implies respect and care for traditional family values – i.e. conservatism. Again, there should not be any obvious positive or negative connotations to either alternative. At the time of the study, the status quo was that teachers were not allowed to confiscate cell phones, religious symbols were not prohibited, and gay couples were not allowed to adopt from foreign countries.
Method
Participants and Design
One hundred and thirty-five undergraduate university students (39 males, 95 females, 1 one who did not respond) at Stockholm University participated in the study in exchange for course credits or lottery tickets. The mean age was 28 (SD = 11.41) years. Participants were randomly assigned one of the three issues, the cell phone issue (n = 44), the religious symbols issue (n = 45), or the gay couples issue (n = 46). About one third of participants in each issue were informed that there was a large majority in favor of the proposition, one third was informed that there was a large majority against the proposition, and one third were given no numerical status information. The design was thus 2 (target group: allies/opponents) × 3 (issue: cell phones, religious symbols, gay couples) × 3 (numerical status: majority, minority, no information) × 2 (position: challenger/defender), with repeated measures on the first factor. All other variables were measured between participants.
Procedure and Variables
Participants were recruited through experimenter visits at lectures. The experimenter waited until after the lectures for the students to fill out the questionnaires, and for this they were given either course credits (if they were psychology students) or lottery tickets (students of other disciplines at the university).
First, participants were provided with information about the issue, and asked to think about the issue for a while before stating their own preferred alternative, which was followed by some unrelated items. Then they were given numerical status information. Participants in the majority in favor of the proposition read the following text:
Before you continue it may be interesting to know that we have conducted an earlier study regarding this issue. It turned out that there was a clear majority in favor of the proposition that teachers should be allowed to confiscate disturbing cell phones.
In the majority against condition the words “majority in favor” was replaced with “majority against”. In the other two issues, the italicized text was replaced with religious symbols should be prohibited and, homosexual couples should be allowed to adopt children from other countries. In the no information condition, the text was completely omitted. After this, participants rated the origins of preferences for both their allies and their opponents on both the rationality and the externality dimension, as in Study 1. Correlations for the rationality dimension were .54 and .52, and for the externality dimension .56 and .62 (all ps < .001), for the ingroup- and outgroup ratings respectively.
This was followed by some unrelated items, and then a manipulation check, adapted from Kenworthy and Miller (2002), asking participants how they felt about there being a clear majority in favor (against) the proposition that teachers should be allowed to confiscate disturbing cell phones (religious symbols should be prohibited in school, or homosexual couples should be allowed to adopt). Participants answered on a 7-point Likert-scale ranging from 1 = very bad to 7 = very good. If participants understood the manipulation they should report feeling bad when the majority was against their own preference and good when the majority was in favor of their own preference.
In total, 68 participants (50.5%) were defined as challengers i.e. they were pro the given proposition, and 65 (48.0%) were defenders. Two persons (1.5%) did not state their position and were hence excluded from further analyses.
Results
Manipulation Check
An independent samples t-test showed a significant difference between those who were informed that the majority shared their preference as opposed to those informed that the majority had the opposite preference as themselves with respect to how they felt, t(100) = 8.25, p < .001. Those whose preference was shared by the majority felt significantly better (M = 5.83, SD = 1.09) than those whose preference was not shared by the majority (M = 2.93, SD = 1.99).
Attitude Issue, Target Group, Position and Numerical Status Effects
Two separate ANOVAs, one for each dimension, with position (challenger/defender), numerical status (majority/minority/no information), and attitude issue (cell phones/religious symbols/gay adoption) as between participants factors, and target group as a repeated factor were conducted. Importantly, analyses showed that there were no effects of attitude issue per se, Fs < 1.70, ps >.18. The ANOVAs replicated the effect of target group from Study 1, such that allies were rated more positively than were opponents, F(1, 124) = 35.97, p < .001, ηp2 = .23, and F(1, 121) = 7.18, p = .008, ηp2 =.06, for rationality and externality respectively. Means and standard deviations are shown in Table 1. Moreover, analyses revealed interaction effects of position with target group, F(1, 124) = 4.83, p = .04, ηp2 = .03, and F(1, 121) = 8.42, p = .004, ηp2 = .07, for rationality and externality respectively. Replicating results of Study 1, challengers again showed stronger intergroup bias than did defenders. Means and standard deviations are shown in Table 1. Again, we performed simple effects analyses using the error term and degrees of freedom from the first ANOVAs to test the difference in ratings of the in- and outgroup within the challenging and defending group separately for each dimension. Results showed that for the rationality dimension, both challengers and defenders significantly differed in their ratings of the in- and outgroup, F(1, 124) = 26.00, p<.001, and F(1, 124) = 4.40, p = .04, for challengers and defenders respectively, but defenders did so to a lesser extent. This mirrored the results of Study 1. For the externality dimension, challengers again differed in their ratings of the in- and outgroup, F(1, 121) = 11.91, p < .001, while defenders did not, F(1, 121) = 0.13, p = .72. Numerical status did not affect biases, Fs < 0.74, ps > .48, nor did the interaction between numerical status and position, Fs < 0.97, ps > .38.
Discussion
Study 2 confirmed findings from the first study showing that challengers of the status quo tended to be more biased in their perceptions of their allies and opponents than were defenders. In Study 2 we used the numerical status manipulation previously used by Kenworthy and Miller (2002). In contrast to their results, we did not find any effects of numerical status, which could indicate that position is more important than numerical status. Even though the manipulation check indicated that the participants seemed to have understood whether their opinion was shared by the majority or not, it might not have been strong enough to affect participants. One could speculate that the implications of the statements ‘a clear majority in favor of the proposal’ or ‘a clear majority against the proposal’ was not clear to participants. As Martin and colleagues (2002) have shown, effects of numerical status may differ according to what specific information is provided. These authors (Experiment 2) did not find effects when only labels were used (‘majority’ vs ‘minority, but effects were strong when both label and percentages were given, that is, when participants were informed that ‘a majority of 82%...’ vs ‘a minority of 18%...’.
In addition, if position really does affect intergroup biases regardless of numerical status, it is not known why. The third study was designed to address these issues.
Study 3
Study 3 was designed to replicate the effects of position when using a stronger numerical status manipulation (Martin et al., 2002). We hypothesized that position to the status quo would significantly predict biased intergroup perceptions as was found in Studies 1 and 2, and that numerical status would not influence such biases, neither main or interaction effects.
Moreover, we wanted to look at possible mediating factors of the effect of position on biases. As stated by several authors previously, challengers are in a vulnerable and exposed position (De Dreu et al., 2008; O’Brien & Crandall, 2005), which could lead them to feel that they need to defend and justify their own position more strongly than defenders. This could be achieved through increased biases to undermine the opponents’ position. Thus, we expect the need to justify own position to mediate the effect of position to the status quo on biases.
Challengers may also experience higher levels of negative affect against their opponents, which could be a foundation for the creation of more conscious cognitive representations of the opponents as having less desirable reasons for their attitudes. As suggested by intergroup emotions theory, when social identity is salient, group based appraisals elicit emotions which may lead to negativity to the outgroup (Mackie et al., 2000). We expect that being the proponents of a specific issue may lead to an increased identification with like-minded others, which should be a pre-condition for negative affect to arise. Thus, we expected that challengers should more strongly identify with like-minded others than defenders, and that they would experience spontaneous negative affect toward their opponents. We hypothesized that both the need to justify one’s own position and negative affect towards the outgroup would lead to increased intergroup biases of attitude reasons.
In addition, we extended on Study 1 and 2 by adding more dependent measures of motives for others’ opinions (Reeder et al., 2005). We expected that these motives would follow the same pattern as the rationality and externality dimension of attitude attribution.
Finally, we chose to look at yet another social issue, the possible implementation of affirmative action to get more women to company boards. The suggestion to make this a law, forcing companies to have at least 40% women in their boards, has been debated in Sweden and Europe lately, and most people seem to be negative towards this. As an example, the former Swedish minister of equality stated that Sweden opposed this suggestion at the EU-level (Haag, 2012). Reasons that have been discussed are that women who are chosen based on such affirmative action will not be respected, or that their enrollment could be regarded in some manner as cheating. Thus, we could be fairly certain that the challenging option would not be associated with obvious positive connotations. However, to be sure, we conducted a short pilot study (n = 110 university students), where we asked participants on a continuous scale from 1 = clearly for to 7 = clearly against, what their take on affirmative action to increase the number of women in company boards was. The mean was 4.43 (SD = 2.07), and it was significantly different from the midpoint of the scale, t(109) = 4.70, p < .001. This confirmed that our student sample was mostly negatively oriented towards affirmative action, and hence we could be sure that the challenging position was negatively connoted.
Method
Participants and Design
One hundred and nine undergraduate university students (35 males, 74 females) at Lund University participated in exchange for lottery tickets. Mean age was 23 (SD = 3.50). Participants were randomly assigned one of the two numerical information conditions, majority (n = 55) or minority (n = 54). The design was thus 2 (target group: allies/opponents) × 2 (numerical status: majority/minority) × 2 (position: challenger/defender), with repeated measures on the first factor.
Procedure and Variables
Participants were recruited through experimenter visits at lectures and in common areas at the university. Participation was reimbursed with a lottery ticket.
First, participants were provided with information about the issue and asked to think about the issue for a while before stating their own preferred alternative. Then they were given the numerical status information. Participants in the majority condition read the following text:
Before you continue it may be interesting to know that we have conducted an earlier study regarding this issue. It turned out that there was a clear majority, 82%, in favor of the proposition to implement a law regarding affirmative action in company boards.
In the minority condition, the words ‘majority, 82%’, was replaced with ‘minority, 18%’. Following the manipulation, we had a manipulation check as in Study 2 asking participants how they felt about there being a clear majority/minority in favor of the proposition. Answers ranged from 1 = very bad, to 7 = very good. Then we assessed to what extent they identified with likeminded others, using three items: ‘I see myself as a person who advocates my opinion about affirmative action’, ‘I identify with other people who have the same opinion as I do about affirmative action’, and ‘I feel belongingness to other people who share my opinion about affirmative action’. Answers ranged from 1 = do not agree at all, to 7 = completely agree. Cronbach’s alpha was high, .84, so we created a mean index.
After this, participants rated the origins of preferences on the rationality and externality dimensions, as well as the motives for their opponents and allies. The dimensions were assessed as in the previous studies. Correlations for the rationality dimension were .53 and .50, and for the externality dimension .36 and .18 (all ps < .001, except for outgroup externality, which only approached significance, p = .07), for the ingroup- and outgroup ratings respectively.
The perceptions of attitude motives were adapted from Reeder et al. (2005) and read: ‘Those with a different [the same] attitude as me about the proposal to implement affirmative action are ethical/driven by self-interest/have a hidden agenda’. Answers ranged from 1 = do not agree at all to 7 = completely agree. After this we assessed emotions to the outgroup with items adapted from Mackie et al. (2000). The question read: ‘When you think about other people who have a different opinion than you about the proposal to implement affirmative action, to what extent do you feel…’. And this was followed by ‘angry/worried/anxious/contemptuous’. Answers ranged from 1 = do not feel angry/worried/anxious/ contemptuous at all, to 7 = feel very angry/worried/anxious/ contemptuous. In order to derive an index of negative affect we first assessed internal reliability for our emotion items (experienced anger, worry, anxiety and contempt for the outgroup). Since Cronbachs alpha was high, .86, we created an index of the mean of the emotion items.
Finally, we assessed the extent to which participants felt a need to justify their position. The question read: ‘I feel a need to justify or defend my opinion’. Answers ranged from 1 = do not agree at all to 7 = completely agree. This was followed by some unrelated items before we assessed demographics.
In total, 48 participants (44%) were defined as challengers, that is, they were pro the proposal to implement a law on affirmative action, and 61 (56%) were defenders.
Results
Manipulation Check
An independent samples t-test showed a significant difference between experimental groups regarding participants’ feelings of being in the numerical minority/majority, t(107) = -9.63, p < .001. Those who were informed that the majority shared their opinion (M = 5.52, SD = 1.33) felt significantly better than those who were informed that the minority shared their opinion (M = 3.09, SD = 1.33). Thus, the participants seemed to have understood whether they were in the majority or minority.
Target Group, Position and Numerical Status Effects
Five separate ANCOVAs, one for each dimension and one for each motive, with position (challenger/defender) and numerical status (majority/minority) as between participants variables and target group (allies/opponents) as a repeated measure were conducted. Participant gender was added as a covariate given the gender-based nature of the attitude issue. The F-values for all effects and dependent variables are shown in Table 2. The results showed main effects of target group for all dependent variables, indicating intergroup differentiation. Thus, the target group effects from Study 1 and 2 were replicated such that allies were rated more rational and less external than opponents. In addition, allies were rated as more ethical, less driven by self-interest, and less likely to have a hidden agenda than opponents. These results are in line with prior research (Bäck, 2013; Bäck et al., 2010; Kenworthy & Miller, 2002; Reeder et al., 2005).
Table 2
F-Values for All Effects and Dependent Variables (Study 3, N = 109)
| Predictors
|
|||||
|---|---|---|---|---|---|
| Outcome | Target Group | Group x Gender | Group x Numerical Status | Group x Position | Group x Numerical Status x Position |
| Rationaliy | 10.77*** | 2.05 | 1.64 | 22.09*** | 0.39 |
| Externality | 7.24** | 2.98 | 0.23 | 1.22 | 0.02 |
| Ethical | 5.14* | 1.92 | 0.49 | 24.67*** | 0.001 |
| Self-interest | 14.80*** | 10.15** | 1.41 | 8.57** | 1.36 |
| Hidden agenda | 5.64* | 1.81 | 1.81 | 13.08*** | 0.51 |
*p < .05. **p < .01. ***p < .001.
Gender significantly predicted biases only for self-interest such that men made a larger difference in the ratings of the ingroup and outgroup (M = 3.17, SD = 1.25, and M = 4.17, SD = 1.49, respectively), than did women (M = 3.99, SD = 1.47, and M = 4.09, SD = 1.59, for in- and outgroup ratings respectively). Thus, men ascribed more self-interest to their opponents than their allies, while women did not make such a large distinction between their allies and opponents.
Moreover, analyses revealed interaction effects of position with target group for all dependent variables except the externality dimension. Again, replicating the results of Studies 1 and 2, challengers showed stronger intergroup biases than defenders. Results are shown in Table 3. Again, we tested the simple effects of the difference between the ratings for the allies and opponents within each position for each dependent variable. Simple effects were again calculated using the error term and degrees of freedom from the first ANCOVAs. Results showed that for the rationality dimension, both challengers and defenders significantly differed in their ratings of the in- and outgroup, F (1, 103) = 58.70, p < .001, and F (1, 103) = 3.91, p = .05, for challengers and defenders respectively, but defenders did so to a lesser extent. The same was true for the externality dimension, F (1, 104) = 10.15, p = .002, and F (1, 104) = 5.54, p = .02 for challengers and defenders respectively. Furthermore, challengers significantly differed in their ratings of the in- and outgroup on all motives, F(1, 103) = 30.29, p <.001; F(1, 104) = 11.94, p < .001; F(1, 104) = 25.67, p < .001, for ethical, self-interest, and hidden agenda respectively, whereas defenders did not, F(1, 103) = 1.03, p = .31; F(1, 104) = 0.16, p = .70; F(1, 104) = 0.32, p = .57 for ethical, self-interest, and hidden agenda respectively.
The general pattern that challengers displayed stronger biases in favor of their ingroup, as compared to defenders, was again confirmed. On all dependent variables where challengers were more biased than defenders and for the motives, the differences in ratings between allies and opponents were not significant for defenders.
Table 3
Means and Standard Deviations of Ratings for Challengers and Defenders of the Status Quo (Study 3, N = 109)
| Challengers
|
Defenders
|
|||||
|---|---|---|---|---|---|---|
| Allies | Opponents | Diff.a | Allies | Opponents | Diff.a | |
| Rationality | 5.07 (0.84) | 3.65 (1.26) | 1.45*** | 4.40 (1.18) | 4.07 (1.17) | 0.33* |
| Externality | 4.04 (1.05) | 4.52 (0.90) | -0.48** | 3.99 (1.12) | 4.30 (1.04) | -0.31* |
| Ethical | 4.91 (1.27) | 3.65 (1.48) | 1.32*** | 3.80 (1.36) | 4.02 (1.30) | -0.21 |
| Self-interest | 3.65 (1.67) | 4.44 (1.83) | -0.79*** | 3.79 (1.25) | 3.87 (1.25) | -0.08 |
| Hidden agenda | 1.98 (1.21) | 2.98 (1.58) | -1.00*** | 2.46 (1.36) | 2.56 (1.18) | -0.10 |
Note. Rationality and externality rated on 7 point Likert-scales, where higher values indicate more ascribed rationality and externality. Higher values on the rationality and lower values on the externality difference indices imply more bias.
aSimple effects.
*p < .05. **p < .01. ***p < .001, two-tailed tests.
Moreover, as in Study 2, there were no effects of numerical status, all Fs < 1.99, ps > .16. Thus, even when using a very strong numerical status manipulation, position to the status quo yielded significant effects while numerical status did not. This lends support to our notion that being a challenger takes precedence over numerical information.
Mediation Analyses
Finally, we sought to understand why challengers might display stronger intergroup biases than defenders. Our argument was that challengers on the one hand should feel a need to justify their position which should lead to intergroup differentiation. In addition, being the proponents of a specific proposal may lead challengers to experience increased negative affect toward the outgroup, in turn influencing biases. This is because they should feel more strongly identified with like-minded others than defenders. Our proposed model can be seen in Figure 1.
Figure 1
Proposed model, with needs to justify own opinion and negative affect toward the outgroup as mediators predicting biased perceptions. The paths refer to the paths being tested in the mediation analyses.
First, we wanted to control that challengers did indeed feel more strongly identified to like-minded others than defenders. For this purpose, we performed an independent group t-test with position (challenger/defender) as grouping factor and the social identification index as dependent variable. Indeed challengers experienced a higher degree of identification with like-minded others, t(106) = -4.29, p < .001. The mean for the challengers was 4.45 (SD = 1.54), as compared to 3.24 (SD = 1.39) for the defenders. Thus, it could be expected that challengers also would experience increased negative affect.
To test our model, we followed the procedure for multiple mediation outlined by Preacher and Hayes (2008), using their macro for SPSS. This analysis provides effects of the independent variable on both mediators simultaneously, direct effects of mediators on the outcome, the total effect of the independent variable on the outcome (regressing only the IV on the outcome), and the direct effect of the independent variable on the outcome (controlling for the mediators). This corresponds to paths a, paths b, path c and path c’ in Kenny, Kashy, and Bolger’s (1998) model. Position to the status quo was dummy-coded with defender as the reference category (0). In order to run the regressions, we created a difference index for each of the dependent variables by subtracting the outgroup ratings from the ingroup ratings. Thus, positive values on the rationality dimension and ethical motive indices imply ingroup favoritism, while negative values on the externality dimension, the self-interest motive and the hidden agenda motive imply ingroup favoritism.
Results showed a significant total effect (path c) of position on biases for all dependent variables except externality, as can be seen in Table 4. These results parallel the findings of the ANCOVAs. There were also significant effects of position on both negative affect and justification needs for all outcome variables. The results indicated that challengers experienced more negative affect and felt a higher need to justify own position, compared to defenders.
As can be seen in Table 4, only negative affect predicted biases (path b1). Thus, even though challengers felt a higher need to justify their own position this need did not predict increased biases, while negative affect did when both variables were entered into the model. Finally, when looking at the effect of position on the outcomes, controlling for the mediators (path c’), it can be seen that there was still a significant, although less strong, effect. Bootstrapped estimates along with Sobel statistics are shown in Table 5.
Table 4
Unstandardized Regression Coefficients and Standard Errors (in Parentheses) for all Effects in the Mediation Analyses Using Position, Need for Justification and Affect as Predictors of Biases (Study 3, N = 109)
| Predictors
|
||||||
|---|---|---|---|---|---|---|
| Outcomes | IV to affect (path a1) | IV to need (path a2) | Affect on DV (path b1) | Need on DV (path b2) | Total effect (IV on DV, path c) | Direct effect (IV on DV, path c’) |
| Rationality | 1.21*** (0.25) | 0.79* (0.35) | 0.44*** (0.09) | -0.06 (0.07) | 1.12*** (0.25) | 0.64* (0.25) |
| Externality | 1.15*** (0.26) | 0.84* (0.35) | -0.09 (0.07) | -0.02 (0.06) | -0.17 (0.20) | -0.05 (0.22) |
| Ethical motives | 1.21*** (0.26) | 0.79* (0.36) | 0.29* (0.12) | -0.04 (0.08) | 1.53*** (0.32) | 1.22*** (0.35) |
| Self-interesta | 1.72*** (0.26) | 0.74* (0.35) | -0.25* (0.11) | 0.09 (0.09) | -0.88*** (0.31) | -0.66† (0.34) |
| Hidden agenda | 1.15*** (0.26) | 0.84* (0.35) | -0.23* (0.10) | 0.07 (0.32) | -0.90*** (0.26) | -0.70* (0.29) |
aGender included as covariate.
†p < .10. *p < .05. **p < .01. ***p < .001, two-tailed tests.
Table 5
Bootstrapped Estimates of (Unstandardized) Indirect Effects (and Standard Errors) of Position to the Status Quo on the Dependent Variables Through the Mediators (ab Paths), and Sobel Statistics (Study 3, N = 109)
| Predictors
|
|||||
|---|---|---|---|---|---|
| Outcome variable | Negative affect | Sobel’s S | Justification needs | Total | Sobel’s S |
| Rationality | 0.51 (0.15) | 3.40*** | -0.05 (0.06) | 0.47 (0.14) | -0.90 |
| Externality | -0.10 (0.08) | -1.37 | -0.02 (0.05) | -0.12 (0.08) | -0.36 |
| Ethical motives | 0.35 (0.17) | 2.12* | -0.03 (0.07) | 0.35 (0.18) | -0.49 |
| Self-interesta | -0.30 (0.14) | -1.95* | 0.07 (0.08) | -0.22 (0.14) | 1.02 |
| Hidden agenda | -0.27 (0.12) | -2.06* | 0.07 (0.07) | -0.20 (0.13) | 0.93 |
aGender included as covariate.
*p < .05. **p < .01. ***p < .001, two-tailed tests.
Discussion
Overall our results indicate that our model may be valid in that at least affect should be seen as a mediator between position and biases. Even though challengers displayed a higher need to justify their position, this need did not explain biases. Hence, justification needs did not seem to be a mediator when affect was also included in the model. Even though mediation was only partial since there was still a direct effect of position on biases, negative affect significantly explained variation in the outcome variables. Thus, negative affect against the outgroup can predict biased perceptions of the sources of others’ attitudes. The results conform to intergroup emotions theory (Mackie et al., 2000) in that when group boundaries become salient, affect is spontaneously experienced. Affect in turn is a foundation for later, more cognitive representations of objects.
There were no significant effects for the externality dimension. The externality dimension usually yields weaker results than the rationality dimension (Bäck et al., 2010; Kenworthy & Miller, 2002). One reason could be that the implications of assigning external reasons for others’ attitudes may not necessarily imply negativity. For instance, by attributing external reasons one can still keep a relatively positive image of a person who disagrees with oneself. Thus, it may not be because my friend lacks the ability to think rationally that he/she disagrees with me, but rather because he or she is too influenced by his/her peers or the media. This reasoning is line with naive realism, which states that others who disagree with oneself are seen, for instance, as lazy, and if they thought about the issue they would agree with oneself (Ross & Ward, 1996).
Gender significantly predicted biases for the self-interest dimension, such that men made a larger difference between their allies and opponents that did women. A possible reason for this could be that men who challenge and propose more gender equality are particularly likely to meet criticism from other men as men would not benefit from a change. Thus, challenging men may feel even more negatively towards defending men and see them as only protecting of their own interests and as being unwilling to share the power.
General Discussion
In line with previous studies (Bäck, 2013; Bäck et al., 2010; Kenworthy & Miller, 2002; Reeder et al., 2005), the present research has shown that people tend to be biased in their perceptions of others whom they agree with as opposed to disagree with, such that people tend to attribute more positive motives to those holding the same as opposed to different opinion as oneself. Specifically, those sharing one’s own opinion are seen as more rational and less externally influenced than those who do not share one’s opinion, who are instead seen as more externally influenced and less rational. In addition, opponents are seen as less ethical, more driven by self-interest, and as more likely to have a hidden agenda. These results have implications for people’s tolerance, respect, and understanding of conflicting perspectives, with potentially important consequences for how people behave towards disagreeing others (Kennedy & Pronin, 2008).
In the current research, we further demonstrated what we label a challenger bias; people who challenge the status quo are more biased in their perceptions of allies’ and opponents’ attitude origins than are people who want to keep the status quo. Our aim was also to test the robustness of these effects, and control for possible confounds. Our results support the challenger bias in that it seems robust to perceived numerical status and positive/negative connotations attached to different positions in relation to different social issues. Even though challengers were informed that the majority shared their position, biases did not decrease. Hence, being in the challenging position per se may override possible effects of numerical status. For the future, this finding is particularly interesting as it suggests that position and numerical status may carry qualitatively different information. Future research on minority and majority influence should take position to the status quo into account. For example, minorities are often seen as rigid, as opposed to flexible (Mugny, 1982) and consistent (Moscovici & Lage, 1976). However, similar perceptions seem to apply to challengers of the status quo. The way these mechanisms operate need future attention.
We argued that one potential explanation to this finding is that challengers are in a vulnerable and exposed position (De Dreu et al., 2008; O’Brien & Crandall, 2005), increasing the need to justify and defend themselves. One way to accomplish this is by increasing biases regarding others’ attitude origins, and hence undermining the validity of the opponents’ preferences (O’Brien & Crandall, 2005). However, we did not receive support for this interpretation. Another explanation may be that such a position is associated with perceptions of belonging to a group that advocates the position, thus making social identification with like-minded others salient. When identity is salient, negative emotions and following negative perceptions (Bodenhausen, Sheppard, & Kramer, 1994) and even behavior may result (Mackie et al., 2000). We received partial support for our model in that negative affect to the outgroup partially mediated the negative biases. However, this is only preliminary evidence and needs to be replicated.
Given that challengers are less tolerant of their opponents’ point of view, they may be perceived as strenuous and stubborn, rather than innovative. This could possibly decrease the likelihood that defenders will take them seriously, which in turn may lead challengers to employ more extreme convincing methods. For instance, negative perceptions of the opposing party may easily lead to an escalation of conflict (Kennedy & Pronin, 2008). In line with this reasoning it has been shown that revisionists are often stereotyped as extremists (Keltner & Robinson, 1997). On the political arena, challengers are sometimes also seen as extremists. For instance, political parties or leaders sometimes resort to downgrading their opponents. Negative campaigning, such as smearing, for example in the form of attribution of mental defects to opposing candidates, has been practiced in the U. S. since 1796 (Felknor, 1992). It has even been proposed that without such negative campaigning challengers of the incumbents have little opportunity to gain power (Felknor, 1992).
The fundamental driving force of negative attitudes towards opponents seems to lie in the challenging position rather than the numerical status. Perhaps the essence of challenging could be considered a more conscious act of choosing a point of view, rather than finding oneself to belong to a minority. To challenge what is generally considered to be good and true demands more from the individual in terms of gearing up for fight and protection against attacks, and to connect to other like-minded in order to put forth a strong offense.
Outgroup Derogation or Ingroup Favoritism?
The general pattern that emerges is that there is a difference in ratings between allies and opponents between challengers and defenders, such that challengers perceive that there is a larger difference between allies and opponents than defenders do. We have thus far not discussed where the difference occurs. That is, is this due to challengers being more ingroup favorising or is it that they are outgroup derogating? In order to understand this, we closely examined the means for in- and outgroup ratings for challengers and defenders on both dimensions in all three studies and on the three motives in Study 3. In Study 1, the differences seemed to be due to challengers being outgroup derogating. That is, challengers attributed more externality and less rationality to the outgroup, while both challengers and defenders rated the ingroup about the same. However, in Studies 2 and 3, challengers also displayed ingroup favoritism, by attributing more rationality and ethicality to the ingroup than did defenders, as well as less externality, less self-interest and less ‘having a hidden agenda’. Thus, in Studies 2 and 3, challengers bolster the difference between the groups by attributing more positive aspects to the ingroup at the same time as they derogate the outgroup by attributing more negative aspects to the outgroup. This indicates that challengers both undermined their opponents’ attitudes, but also that they enhanced their own motives. Perhaps this is not surprising as it would seem plausible that in order to challenge the status quo, one would need to be a firm believer of the proposal that one advocates.
Implications
The present results provide important insights to the emergence of intergroup conflict in particular when it comes to factions within a group, such as an organization, or even a society. Our findings indicate that those who want a change are less tolerant of their opponents’ point of view, which may lead to a conflict escalation. This can also explain why some groups that challenge the status quo are seen as extremists, and become strongly polarized, such as the Tea Party movement in the United States. Relatedly, in the U.S. the Democratic and the Republican parties have become increasingly polarized from each other (Bonica, McCarty, Poole, & Rosenthal, 2013), possibly adding to the effect of the challenger bias. Knowledge about the challenger bias could be beneficial in avoiding or reducing intergroup conflicts and may help establish a better understanding of the opposing group and its goals.
Limitations
Even though we sought to use a relatively well established paradigm for investigating others’ origins of attitudes, there seemed to be some problems with how to interpret some of the dimensions. In both previous studies that have used this paradigm, results for the externality dimension have typically been lower and less stable than for the rationality dimension (Bäck, 2013; Kenworthy & Miller, 2002). One interpretation is that people ascribe external sources in order to keep a positive image of someone else. For example, if a friend strongly opposes something that oneself advocates, one can search for external reasons for why he or she holds this position and thus remain friends with this person, while making excuses for his/her position.
In addition, Kenworthy and Miller (2002) investigated a third dimension; emotionality. This dimension refers to someone holding a certain attitude because it feels good (Maio & Haddock, 2007). Although it was expected that basing one’s attitudes on emotions would be considered negative, the results varied between the experiments in Kenworthy and Miller’s study. Thus we chose to exclude that dimension in the present research because it is not clear how it should be interpreted. With that said, we do not consider emotions unimportant as predictors of attitude positions. On the contrary, in our third study we have added an emotional component explaining intergroup biases. However, it may be that different emotions have different evaluative connotations. To have an attitude because of fear may be less desirable than to have an attitude because of anger, or happiness, or sadness. These emotions are highly distinct in that, for example, they trigger different information processing systems (Bless, Schwarz, & Wieland, 1996; Kahneman, 2011). Hence, it is plausible that people ascribe different positive or negative connotations to different emotions as predictors of others’ attitudes. Thus, we agree with Kenworthy and Miller’s conclusion that such connotations of emotionality as an explanation for others’ attitudes should be examined.
Another limitation concerns the manipulation of majority-minority status information. Although we conducted a third study confirming the more weak numerical status manipulation used in Study 2, we cannot rule out that there are other ways that numerical status could influence the results. For instance, to say that a majority is in favor implies that the minority is against, but this framing may have an impact on how people perceive the information (Falomir-Pichastor, Mugny, Quiamzade, & Gabarrot, 2008; Moscovici, 1980). In our third study, we chose to use the framing that either a majority of 82% or a minority of 18% was in favor of the proposal. However, we might have had other results if we had used the opposite framing. That is, a majority of 82% or a minority of 18% opposes the proposal. It logically implies the same thing, but it may have different impact on people (Falomir-Pichastor et al., 2008).
Finally, in the present study we used the term “defender” to refer to people who are aligned with the status quo. However, as this position is the default it is possible that people do not feel the need to actively defend their position since it is not actually being questioned or threatened. It is plausible that if the “defenders” were threatened, for instance by a policy proposal that the status quo is about to be changed, their biases might increase. It is possible that under such circumstances, the asymmetry between challengers and defenders might disappear. Thus, it would be desirable for future research to test the boundary conditions for the found effect, for instance by manipulating a threat to the status quo. However, with that in mind, the defenders did not show increased biases when informed that they were in minority, a commonly used threat manipulation (Branscombe & Wann, 1994; Kenworthy & Miller, 2002; Stephan et al., 2002) that should be perceived as the status quo being threatened.
Conclusions
To conclude, it seems that challengers of the status quo display less tolerance for others’ deviating opinions. This may lead to a perception of challengers as more extreme, which in turn may decrease the likelihood that the defenders want to engage in cooperation. For instance, recent research show that activists, who are proponents of social change, are negatively stereotyped, and people are unwilling to agree with a statement made by an activist as opposed to the same statement made by a non-activist (Bashir et al., 2013). Instead, it may actually lead to an escalation of the disagreement into a conflict. Thus, ironically, challengers’ negative perceptions of their outgroup and their exaggerated view of their own group’s superiority may actually bolster the status quo.