People or Ideology? Social Conservatism and Intergroup Contact Moderate Heterosexuals’ Responses to a State-sponsored Anti-LGBT Campaign

Recent years have witnessed the global rise of illiberal regimes, which built widespread support by initiating hate campaigns against minorities. This research examined a government-led anti-LGBT campaign in Poland and its impact upon attitudes of the wider public. Our study (N = 2061 heterosexual participants) assessed sexual prejudice (blatant dehumanization and modern homonegativity) and pro- and anti-minority action intentions in two representative surveys, conducted before (fall 2018; N = 821) and after (summer 2019; N = 1240) the beginning of the anti-LGBT campaign. Prejudice and anti-minority action tendencies increased during the analyzed period. However, there was also an increase in pro-minority action tendencies, demonstrating public resistance. These effects were moderated by respondents’ levels of social conservatism and their prior contact with LGBT people. Low (compared to high) levels of social conservatism reduced (or even reversed) the impact of the state-sponsored anti-LGBT campaign on sexual prejudice, which further translated into pro- and anti-minority collective action intentions. High (compared to low) levels of contact predicted a reduced effect of the campaign on blatant dehumanization, which translated into lower anti-minority action tendencies. Blatant dehumanization and modern homonegativity differentially predicted participants’ willingness to engage in pro- and anti-minority collective action, highlighting the importance of taking a more granular approach when predicting action intentions. We discuss implications for interventions aimed at protecting vulnerable groups in the face of anti-minority rhetoric.


Introduction
"They try to convince us that LGBT is people -it is simply an ideology," said Andrzej Duda, the incumbent President of Poland, during his election rally in Brzeg, June 13, 2020 (Dellanna & Holroyd, 2020). The homo/transphobic sentiment in Duda's speech was nothing new. His words were a continuation of the tried and tested anti-LGBT rhetoric employed by Law and Justice-Poland's right-wing ruling party and Duda's former political grouping-to win the 2019 European and national parliamentary elections.
By attacking the LGBT (i.e., lesbian, gay bisexual, and transgender) community, the Polish government followed a common strategy employed by powerholders to consolidate their position. Presenting a minority group as a threat to society, the authorities may then adopt the role of its defenders and thereby win the support of the numerical majority (see Subašić et al., 2008). Although the choice of the targeted minority depends on context (e.g., Glick, 2005), similar patterns of state-sponsored persecution observed recently in Russia, Hungary, or the USA suggest that sexual and gender minorities may be especially prone to attacks under right-wing populist rule (Applebaum, 2020;Snyder, 2018). This is because people's stance on LGBT rights defines their place on a number of social and political dimensions (e.g., East vs. West, tradition vs. progress, hierarchy vs. equality, religion vs. secularity, patriarchy vs. gender equality). By rendering LGBT rights a topic of political debate, the authorities thus reach wider audiences than they would by addressing issues specific to particular lines of ideological division.
The current research pursues three goals. First, in the context of Poland's recent anti-LGBT campaign, we aim to document changes in public opinion with respect to both blatant and subtle forms of prejudice against sexual minorities. We do this by comparing representative samples of the Polish public in 2018 (prior to the government's campaign) and 2019 (after the anti-LGBT campaign had started). Second, to shed light on the potential implications of such increases in prejudice, we examine the extent to which different forms of prejudice predict different forms of collective action (i.e., action undertaken to attain the political goals of a group; van Zomeren, 2016). Specifically, we compare blatant dehumanization of sexual minorities-i.e., their explicit and deliberate characterization as less than human (see Kteily et al., 2015)-and modern homonegativity, a more subtle form of sexual prejudice that is primarily characterized by the denial of existing inequalities (see Morrison & Morrison, 2003) as predictors of both anti-and pro-minority collective action intentions (i.e., collective action which harms and action that benefits the outgroup; see Radke et al., 2020). Finally, we examine factors that are likely to shape the response among the cis-heterosexual (i.e., heterosexual individuals whose gender identity matches the sex they were assigned at birth) majority to anti-LGBT rhetoric.
We focus on two key variables that have been shown to powerfully impact political conflict and intergroup relations in a variety of domains and that capture a broad spectrum of psychological and experiential individual differences: political ideology, in particular social conservatism (see Jost, 2017), and intergroup contact (Pettigrew & Tropp, 2006). Social conservatism entails high sensitivity to threat (e.g., Jost et al., 2003) and is associated with moral principles that are incompatible with the idea of LGBT rights (Kivikangas et al., 2021). Intergroup contact, on the other hand, has been shown to immunize majority members against the established antecedents of prejudice, such as parental authoritarianism (Dhont & van Hiel, 2012) or preference for hierarchical intergroup relations (Visintin et al., 2019). Therefore, we propose that social conservatism and contact with gay men and lesbian women moderate the majority's responses to authorities' threat-inducing narrative, such that individuals with low levels of contact and high levels of social conservatism are most likely to respond with increased negative attitudes toward sexual minorities and higher willingness to engage in collective action to harm the minority group. By contrast, we expect that majority members who manifest high intergroup contact and low social conservatism respond with reduced sexual prejudice and increased pro-outgroup collective action intentions.
In the sections below, we first give an overview of the dynamic relations between the authorities, minority groups, and the majority by introducing the tripolar perspective on intergroup relations. We then outline our hypotheses about the impact of anti-LGBT rhetoric, link pro-and anti-outgroup collective action to different types of prejudice, and we provide arguments for the moderating effects of intergroup contact and social conservatism. The conceptual model tested in our research is presented in Fig. 1.

Tripolar View of Intergroup Relations
Classic approaches in social psychology conceptualize social reality in binary terms. For instance, social identity theory (SIT, Tajfel & Turner, 1986) introduces the ingroup-outgroup dyad and specifies conditions under which members of a disadvantaged group are most likely to rebel against an advantaged group. Nevertheless, intergroup relations usually involve more than two groups (e.g., Dixon et al., 2020;Simon & Klandermans, 2001;Subašić et al., 2008;Zagefka, 2019). For example, the political solidarity model of social change (Subašić et al., 2008) distinguishes three main actors: the authority (i.e., those who occupy positions of power and strive to defend the status quo), the minority (i.e., those who are deprived of power and seek to bring about social change), and the majority (i.e., the most numerous group that observes the conflict between the authority and the minority). As proposed by the authors, social change occurs only when majority members share norms and values with the minority group and distance themselves from the authority.
One of the practices the authority may adopt to maintain its position is the "divide and rule" strategy (Dixon et al., 2020). In order to prevent alliances between majority and minority members, powerholders reinforce existing differences between these two groups or create new areas of division. This is achieved, for example, by limiting opportunities for intergroup contact, encouraging competition, or discrediting the minority group as "traitors" who endanger the values of a given society. We believe that the current state-sponsored attacks on LGBT people in Russia, Poland, and Hungary reflect the "divide and rule" strategy and serve to consolidate majority support for the ruling elite (e.g., Żuk & Żuk, 2020). In the narratives developed in these contexts, the cis-heterosexual majority (the so-called healthy part of society) is contrasted with LGBT minority, which is presented as a threat to biological and cultural reproduction of a nation (Mole et al., 2021;Korolczuk & Graff, 2018). The powerholders, in turn, present themselves as the only political actor able to defend the majority against this destructive influence (e.g., Żuk & Żuk, 2020). Such a defense is not limited to the anti-LGBT discourse, but also entails the implementation of policies designed to infringe minority rights (e.g., the homopropaganda ban introduced in Russia). Moreover, these hostile actions discourage LGBT individuals from disclosing their identities and, as a result, limit opportunities for positive intergroup contact. As this makes minority-majority coalitions less likely, it further bolsters the status quo.

Anti-LGBT Rhetoric, Prejudice, and Collective Action?
We propose that the hate-laden rhetoric used by the authority changes the attitudes and behaviors of cis-heterosexual majority members (Hypothesis 1). Specifically, anti-LGBT campaigns are likely to increase the societal level of sexual prejudice and anti-outgroup collective action intentions. At the same time, change in the global level of pro-outgroup collective action intentions may take different directions. On one hand, the increase in negative attitudes toward LGBT community may entail the overall decrease of majority members' willingness to engage on behalf of this group. On the other hand, state-sponsored attacks on LGBT people may mobilize at least some cisheterosexual individuals to defend the persecuted minority, translating into the global increase of pro-outgroup action tendencies.
While social psychological research identified numerous factors prompting members of the advantaged groups to engage in support of disadvantaged groups (see Radke et al., 2020), the antecedents of collective action against disadvantaged groups have received scant research attention (for exceptions, see Górska et al., 2022a, b;Górska et al., 2022a, b;Osborne et al., 2019;Stefaniak et al., 2020). The limited work available suggests, however, that both pro-and anti-outgroup collective action (or system-challenging and system-supporting collective action, see Jost et al., 2017) are associated (albeit differently) with prejudice. Specifically, while prejudice exerts a negative effect on collective action in support of the disadvantaged groups (e.g., Swank et al., 2013), its association with willingness to engage in collective action against the disadvantaged groups is positive (e.g., Górska et al., 2022a, b). In terms of causality flow, previous longitudinal research suggests that the relationship between prejudice and outgroup-related engagement is unidirectional rather than bidirectional. Specifically, while negative attitudes toward the immigrants were longitudinally associated with higher host society members' willingness to engage in collective action against this group, the over-time effect of collective action on outgroup-directed attitudes did not reach significance (Górska et al., 2022a, b).
However, negative outgroup attitudes can take many forms; direct, extremely hostile forms coexist with more indirect, "modern," and subtle forms (for a review, see Brown, 2011). We suggest here that willingness to take anti-outgroup collective action would bear a strong positive association with overtly hostile outgroup attitudes, while pro-outgroup collective action intentions should be inhibited to a greater extent by subtle forms of prejudice (Hypotheses 2a and b,respectively).
Numerous studies have demonstrated that extreme outgroup-directed actions, such as violence, discrimination, or anti-outgroup collective action develop on top of extremely rather than moderately negative attitudes (e.g., Cowan et al., 2005;Tausch et al., 2011). One example of such severe hostility is blatant dehumanization. Prior research (e.g., Kteily et al., 2015;Moore-Berg et al., 2020) has demonstrated that blatant dehumanization facilitates aggressive action intentions beyond subtle dehumanization (i.e., denying outgroup members uniquely human traits; see Haslam et al., 2005) and negative affect. For instance, majority Americans who overtly associated Mexican immigrants and Muslims with animals were more likely to sign anti-outgroup petitions, even when their political conservatism and negative affect were accounted for (Kteily & Bruneau, 2017). Thus, we expect that blatant dehumanization of homosexual people would serve as a strong antecedent of one's willingness to engage in collective action against this group.
On the other hand, collective action in support of an outgroup originates (among other factors) from injustice perceptions (Leach et al., 2006). Modern prejudicea type of outgroup attitude that emerged in response to political claims made by historically disadvantaged groups (Brown, 2011)-reduces injustice perceptions as it denies the existence of unjust inequalities. Specifically, it rests on three ostensibly neutral beliefs: (1) that all groups nowadays enjoy equal rights (meaning that discrimination is no longer an issue), (2) that political demands put forward by the outgroup are unjustified, and (3) that a given outgroup is responsible for its own marginalization (McConahay, 1986). Unlike blatant dehumanization, modern prejudice does not involve the belief in allegedly inferior nature of the disadvantaged outgroup. Instead, it focuses on outgroup members' behaviors that may threaten the existing hierarchy. As such, while individuals high in blatant dehumanization would display outgroup-directed hostility in all situations possible, advantaged group members exhibiting high level of modern homonegativity would express their antipathy only when the disadvantaged outgroup demands social change toward equality.
As evidenced by numerous studies (for a review, see Brown, 2011), modern prejudice is closely associated with the disapproval of policies designed to reduce intergroup hierarchies. This is also the case in the context of sexual minorities; past research found a negative effect of modern homonegativity (i.e., modern prejudice toward gay men and lesbians; Morrison & Morrison, 2003) on support for LGB rights (e.g., Górska et al., 2017). A similar result may occur for collective action in support of sexual minorities.
Taken together, the arguments presented above allow to expect that anti-LGBT campaigns would affect cis-heterosexual individuals' outgroup-related collective action by changing attitudes toward sexual and gender minorities (Hypothesis 3). In particular, we propose that, by increasing blatant dehumanization, state-sponsored attacks on the LGBT community would predict higher intentions to engage in collective action against this group (Hypothesis 3a). On the other hand, the threatinduced increase in modern homonegativity should entail lower readiness to engage in pro-outgroup collective action (Hypothesis 3b).
Nonetheless, not all members of the majority group are determined to embrace prejudice toward the persecuted minority and engage in discriminatory behaviors. There may be some who remain indifferent and others who actively defend the minority group. This kind of heterogeneity has been documented in numerous contexts, most dramatically in the diversity of attitudes toward the persecution of Jewish people (the minority) among non-Jewish residents (the majority) of countries occupied by the Nazi regime (the authority; e.g., Čehajić-Clancy & Bilewicz, 2020). Why, when confronted with the authority's attack on a minority, do some members of the majority group engage in antiminority behavior while the others defend this minority? In other words, which factors moderate majority members' responses to the persecution of the minority group by the established authority? We propose that, in case of sexual and gender minorities, individual differences in intergroup contact (Hypothesis 4) and social conservatism (Hypothesis 5) are likely to play an important role.

Intergroup Contact
Positive intergroup contact is a powerful means of improving intergroup relations; it not only betters attitudes toward other groups (Pettigrew & Tropp, 2006) but also elicits collective action on their behalf (e.g., Hässler et al., 2020;Reimer et al., 2017). In the context of sexual and gender minorities, cis-heterosexuals' contact with LGBT people is associated with lower levels of sexual prejudice (e.g., Smith et al., 2009) and stronger intentions to engage in actions aimed to benefit that group (Hässler et al., 2020;Reimer et al., 2017). Importantly, prior intergroup contact can also buffer the impact of factors known to stimulate prejudice. For instance, Visintin et al. (2020) showed that the impact of intolerant social norms on outgroup attitudes was significantly weaker among individuals with high (vs. low) intergroup contact. In a similar vein, Jasinskaja-Lahti et al. (2011) found that intolerant parental and peer norms did not translate into minority youth's negative attitudes toward the majority when their personal experiences with majority members were positive. Moreover, research demonstrated that adolescents' contact with immigrant peers diminished the positive association between parental right-wing authoritarianism and adolescents' anti-immigrant prejudice (Dhont & van Hiel, 2012) and that the negative association between social dominance orientation on support for multiculturalism was weaker among participants with high levels of contact with immigrants (Visintin et al., 2019).
One reason why contact with minority representatives may immunize majority members against authorities' "divide and rule" strategy is the shift in coalitionary boundaries it likely entails (Cikara, 2021). The common ingroup identity model (Gaertner & Dovidio, 2000) postulates that positive intergroup contact blurs intergroup boundaries and facilitates the development of superordinate identity, the latter including both ingroup and outgroup members. Thus, individuals involved in harmonious intergroup encounters may no longer view themselves as the members of their primary ingroup but as the members of a broader common ingroup. At the same time, individuals perceived previously as belonging to the outgroup become viewed as the members of the superordinate ingroup. When these ingroup members are assaulted by an external force, taking action to defend them, or at least not joining the attacker, is a much more likely response. In the context of anti-LGBT campaigns, seeing minority friends as members of the common ingroup (e.g., "regular people") may inoculate cis-heterosexual majority representatives against authorities' attempts to impose the division for "healthy" and "diseased" part of the society.
Second, the buffering effect of intergroup contact may be attributed to the particular features of attitudes based on direct personal involvement (see Visintin et al., 2020). Specifically, attitudes developed through direct, behavioral experience with an attitude object are stronger and more resistant to change than attitudes based on indirect, nonbehavioral experience (see Fazio & Zanna, 1981;Howe & Krosnick, 2017). 1 Therefore, when negative, secondhand information about an outgroup is confronted with positive, direct experience gained in personal interactions, then the effect of secondhand information on attitudes is reduced. By contrast, when an individual lacks positive experience with the outgroup, their attitudes are formed on the basis of secondary information (see Sechrist & Stangor, 2007).
Based on this evidence, we expect intergroup contact to moderate the impact of threatening rhetoric on majority members' sexual prejudice. Specifically, the effect of anti-LGBT campaigns on blatant dehumanization and modern homonegativity should be the strongest among individuals with low levels of intergroup contact and become weaker among individuals reporting high levels of contact (Hypotheses 4a and 4b, respectively). Consequently, it is legitimate to expect that intergroup contact would moderate the indirect effects of anti-LGBT narrative on pro-and anti-outgroup collective action. In particular, the heightening effect of anti- LGBT campaigns on anti-outgroup collective action intentions via increased blatant dehumanization should be most pronounced when contact with minority members is low and should become weaker at higher levels of intergroup contact (Hypothesis 4c). Likewise, the negative effect of anti-LGBT rhetoric on willingness to engage in pro-outgroup collective action via increased modern homonegativity should be the strongest at low levels of intergroup contact and should become weaker when intergroup contact is high (Hypothesis 4d).
There are also reasons to believe that when a critical threshold of intergroup contact (see MacInnis & Hodson, 2019) is exceeded, majority members would be motivated to act in defense of LGBT people. By witnessing the impact of such campaigns on their LGBT friends and acquaintances, cis-heterosexual individuals may become aware of the injustice faced by LGBT community as a whole. As injustice perception is a key predictor of collective action in solidarity with other groups (e.g., Leach et al., 2006), anti- LGBT campaigns designed to elicit the sense of threat may, paradoxically, prompt some majority members to engage in pro-minority collective action.

Social Conservatism
Another factor likely to shape cis-heterosexual majority members' response to the state-sponsored persecution of LGBT community is political ideology. Individual differences in endorsed ideology impact a wide range of beliefs, attitudes, and behaviors (see Jost, 2017). While political ideology is often conceptualized along a single dimension (i.e., left to right, liberal to conservative), differentiating between social (i.e., preference for traditional social arrangements and collective security) and economic (i.e., preference for limited government interventionism) conservatism may be more accurate in many contexts (see Bilewicz et al., 2015). As attitudes toward sexual and gender minorities are more strongly related to social than to economic conservatism (e.g., Barnett et al., 2018;Rosik et al., 2013), we focus on this dimension and expect that, in comparison to social liberals, social conservatives would be more likely to adopt intolerant narratives and act accordingly. The rationale behind this claim is threefold.

3
First, divergent responses to the victimization of a minority group may originate from identity concerns and affective polarization-the tendency to perceive opposing partisans negatively and copartisans positively (Iyengar & Westwood, 2015). As LGBT rights are an important element of the liberal agenda, social conservatives may view prejudice against sexual and gender minorities as a way to express their collective identity and harm their political adversaries. Moreover, believing that the right-wing authority shares their core values, and thus their identity, socially conservative majority members may have complete confidence in the messages conveyed by the powerholders, and they may modify their beliefs and behavior accordingly (see Haslam & Reicher, 2012). Conversely, for majority members who perceive their values to be at odds with the authority's, anti-minority narratives have little credibility. It is also conceivable that majority members who feel extreme antipathy toward the authority would endorse pro-minority views as a means of weakening the authority and maintaining positive distinctiveness in relation to them. Moreover, socially liberal members of the cis-heterosexual majority may believe that LGBT people share common fate with their ingroup (i.e., experience unfavorable treatment from the right-wing authority). This, in turn, may serve as a cue to shared group membership (i.e., those who are targeted by the ruling party) and motivate liberals to solidarize with the persecuted members (i.e., LGBT individuals) of their newly established common ingroup (see Cikara, 2021).
Second, sexual and gender minorities challenge values and moral principles associated with the conservative worldview (e.g., Golec de Zavala et al., 2021;Mole et al., 2021). For instance, homosexuality remains in conflict with binding moral foundations (Barnett et al., 2018)-a set of moral principles (i.e., purity/sanctity, authority/respect, and ingroup/loyalty) that prioritize broader collective interests over individual interests (see Graham et al., 2011) and show a positive association with conservatism (e.g., Graham et al., 2009;Kivikangas et al., 2021). By accentuating the conflict between binding moral foundations and homosexuality, the authorities may prompt conservative members of cis-heterosexual majority to adopt stronger sexual prejudice and act against sexual minorities. The opposite response may be expected of liberals, who rely on individualizing (i.e., putting individual interests above collective wellbeing) rather than binding moral foundations (e.g., Graham et al., 2009). This is because state-sponsored victimization of sexual minorities violates harm/care and fairness/reciprocity principles. As this challenges their moral beliefs about what is right and wrong, cis-heterosexual liberals are likely to develop moral outrage and actively support sexual minorities (see Radke et al., 2020).
Finally, even beyond gender identity and sexuality contexts, people's reactions to threat and uncertainty are ideologically asymmetric (e.g., Jost, 2017). Compared with liberals, conservatives show stronger existential and epistemic needs (Jost et al., 2003), and they are more responsive to signals of threat and ambiguity (Feldman, 2003). Therefore, when a minority is presented as threatening and ambiguous, the increase in prejudice and hostile behavior toward this group should be most strongly pronounced among conservative members of the majority. This may be particularly so in the case of LGBT people, who challenge the simplistic, binary views of gender and sexuality (e.g., Makwana et al., 2018).
Based on these rationales, we propose that the anti-LGBT rhetoric would be most strongly associated with higher blatant dehumanization of minority members and higher modern homonegativity at high levels of social conservatism and become weaker at lower levels (Hypotheses 5a and 5b, respectively). Likewise, the positive indirect effect of anti- LGBT campaigns on anti-outgroup collective action intentions via increased blatant dehumanization, as well as the negative indirect effect of anti-LGBT campaigns on prooutgroup collective action intentions via increased modern homonegativity, should be strongest when social conservatism is high and the weakest when it is low (Hypotheses 5c and 5d, respectively).

The Current Research Context
We verified our theorizing in the context of recent sociopolitical developments in Poland. Although the communist Poland  did not criminalize homosexual acts, the situation of the local LGBT community was not particularly good at the time. The Security Services surveilled gay men, and homosexuality was viewed as a depravity imported from the West (Mole et al., 2021). Following the country's transition to liberal democracy in 1989, sexual and gender minorities became more visible. However, the increase in visibility was not accompanied by the extension of LGBT rights. Due to the strong position of the Roman Catholic Church, which condemns homosexuality and equates biological sex with gender identity and gender expression (see Marchlewska et al., 2019), the consecutive parliaments did not pass the bills that would recognize same-sex couples or that criminalize hate crime against LGBT people. In addition, Poland's accession to the European Union in 2004 evoked conservative backlash, which construed sexual prejudice as a way to protect the national identity from the threat posed by external political forces (e.g., Mole et al., 2021).
Today, the idea of homosexuality as incompatible with the national-Catholic notion of "Polishness" is an important component of the ideological program offered by Law and Justice-a right-wing populist party that took power in 2015. To mobilize its voters in advance of the 2019 European and parliamentary elections, Law and Justice made the supposed LGBT threat a flashpoint of their election campaign. This focus emerged after Rafał Trzaskowski (a newly elected Mayor of Warsaw) signed the 12-point LGBT + Declaration in February 2019, which pledged to introduce policies intended to support the local LGBT community. Backed by the Polish Catholic Church and the subordinated media, Law and Justice responded by presenting LGBT people as a threat to the integrity of Polish families (Mole et al., 2021;Żuk & Żuk, 2020). Protests against the LGBT + Declaration were organized in Warsaw, and municipalities from the southeastern part of the country (a stronghold of Law and Justice) adopted resolutions against "LGBT ideology" (Żuk & Żuk, 2020). The effects of this campaign on public sentiment in Poland were profound. The number of Polish language tweets containing the words "LGBT" and "danger*" increased sharply after the LGBT + Declaration had been signed and remained high until the 2019 parliamentary elections (see Fig. 2). Relatedly, public opinion research revealed the growth of negative attitudes toward LGBT community. For instance, in the 2019 survey conducted by the Public Opinion Research Center, 43% of respondents (compared to 36% in 2017) strongly opposed the institutionalization of same-sex civil unions (Bożewicz, 2019).
Nonetheless, the societal response to this anti-LGBT campaign was not uniform. Emboldened by the approval from the authorities, traditionally homophobic far-right activists felt free to physically assault LGBT individuals. For instance, participants of the first ever Equality March in Białystok-a city in north-eastern Poland-were violently attacked by the local nationalist groups (e.g., Mole et al., 2021). At the same time, many socially liberal Poles took to the streets to support LGBT rights. For example, the 2019 Warsaw Equality Parade gathered the unprecedented crowd, including thousands of cis-heterosexual participants. The aim of the present research was to explain the diversity of responses to the attack on the LGBT community. We did so by examining prior intergroup contact and social conservatism as the moderators of cis-heterosexual Poles' responses to the victimization of LGBT people. Furthermore, to assess the roles of different forms of prejudice in shaping different types of action, we tested whether blatant dehumanization and modern homonegativity differentially predict intentions to engage in pro-and anti-minority collective action (see Fig. 1).
To test our hypotheses, we utilized data collected in two representative surveys of Poles carried out in autumn 2018 and summer 2019 (i.e., just before and after the Law and Justice party identified the LGBT community as an enemy to the nation; see Mole et al, 2021). We operationalized the presence of the state-sponsored campaign against the LGBT community as measurement timing (i.e., before versus after the beginning of the anti-LGBT campaign). There are at least two reasons to believe that such an operationalization is appropriate. First, temporal changes in the co-occurrence of "LGBT" and "danger*" words in Polish Twittersphere presented in Fig. 2, as well as changes in attitudes found in surveys conducted by recognized research companies (Bożewicz, 2019) suggest that society members were aware of the authorities' anti-minority rhetoric. Second, besides the state-sponsored anti-LGBT campaign, there was no other event that was likely to have intensified sexual prejudice (e.g., the outbreak of the global pandemic; see Golec de Zavala et al., 2021) in the period between measurement points. Our conceptual model (Fig. 1), which incorporated Hypotheses 1-5, was verified in the structural equations modeling framework.

Participants and Procedure
We conducted two representative 2 surveys of adult Poles, in September-October 2018 (N = 1000) and July-August 2019 (N = 1300), respectively. Each time, random sampling with the identity number (PESEL) database as the sampling frame was employed. 3 Data were collected by two external companies with the use of at-home computer-assisted personal interviews (CAPI). 4 The response rate was 74.46% in 2018 and 72.59% in 2019. After excluding participants who did not declare to be heterosexual (n = 179 and n = 60, respectively), the sample consisted of 2061 heterosexual respondents (976 men and 1085 women). Participants' age ranged from 18 to 91, M = 46.48, SD = 15.59. Distributions of gender, age, education, and settlement size did not differ between the samples (all ps ≥ .092).

Measures
Measures used in the analyses were embedded in larger questionnaires that assessed a range of constructs (e.g., life satisfaction; Diener et al., 1985). 5 Due to space constraints, we employed short versions of the original scales. Unless otherwise noted, all measures used a 1 (strongly disagree) to 7 (strongly agree) response scale. For multi-item measures, composite scores were calculated by averaging responses to respective items.
Modern homonegativity was measured with two items adopted from the Modern Homonegativity Scale Blatant dehumanization was assessed with the Ascent of Man measure (Kteily et al., 2015). Respondents were presented with the image depicting five phases of evolutionary progress from ape to human. The image was accompanied by a 9-point scale where 1 corresponded to the least and 9 to the most developed stage of evolution. Participants were asked to report how evolved they considered homosexual people to be. Responses were reverse-coded so that higher scores reflected greater dehumanization, M 2018 = 2.19, Covariates were gender (coded −0.5 for women and 0.5 for men), age, education (years of full-time education), 2 Each year, the samples were representative of the target population in terms of gender and settlement size. Sample collected in 2019 was also representative when the distribution of age was concerned. By contrast, the sample collected in 2018 was slightly younger than the population of adult Poles (for details, see the Online Supplement). 3 In both surveys, participants were selected randomly from a comprehensive register of adults (≥ 18) who had been assigned the identity number (PESEL). As such, although the same individuals could have been included in both samples, the probability of such an event was extremely low (i.e., 1.01 × 10 −15 ). 4 We acknowledge that the mode of data collection (i.e., interviewers visiting respondents' homes) could introduce social desirability bias to our results. However, since both studies used the same methodology, this problem is common for data collected in 2018 and 2019 and, as such, does not affect results' comparability. 5 Both questionnaires are available on the OSF companion website. 6 The full measure of modern homonegativity included four items, two of which were reverse-scored ("Homosexuals who are out of the closet should be admired for their courage" and "Homosexuals still need to protest for equal rights"). Due to the low reliability of the 4-item scale at T1 (α 2018 = .37), we decided to present the results obtained for the 2-item scale in the main text, and the results for the 4-item scale in the Online Supplement. 7 We caution the readers that our research used the absolute (as opposed to relative) blatant dehumanization scores. 8 We also measured economic (1 = I support welfare state, 7 = I support free market economy) and political orientation (1 = left, 7 = right). For the results of analyses accounting for (1) social and economic conservatism and (2) political conservatism, see the Online Supplement. and settlement size (1 = rural area, 2 = town up to 100,000 residents, 3 = town between 100,001 and 200,000 residents, 4 = town with more than 200,000 residents). Table 1 presents intercorrelations between the analyzed variables separately for 2018 and 2019. To verify Hypothesis 1, a series of two sample t-tests were performed. 9 In line with our expectations, blatant dehumanization, t(1851.52) = −9.04, p < .001, d = 0.42, modern homonegativity, t(1985) = −3.94, p < .001, d = 0.18, as well as antioutgroup, t(1771.43) = −5.71, p < .001, d = 0.27, and prooutgroup collective action intentions, t(1768.71) = −5.85, p < .001, d = 0.27, were higher after compared to before the state-sponsored anti-LGBT campaign. The levels of social conservatism and intergroup contact did not differ between the measurements, ps ≥ .091.

Results
Next, to test H2-H5, we estimated a saturated path model presented in Fig. 1. Due to the relatively low levels observed for some variables (e.g., intergroup contact), we tested the multivariate normality assumption (a requirement to use ML estimator). As this assumption was violated (Mardia's kurtosis = 57.36, p < .001; Mardia's skewness = 12240.72, p < .001), confidence intervals (CIs) for direct as well as indirect effects were obtained with bootstrapping (5000 samples). Missing data (5.39%) were handled using full information maximum-likelihood estimation (FIML), and the analyses were performed in R lavaan (Rosseel, 2012). Table 2 presents the results for the estimated model.
First, we verified our theorizing on the relative importance of blatant dehumanization and modern homonegativity in predicting one's willingness to engage anti-and prooutgroup collective action. Both blatant dehumanization and modern homonegativity were positive predictors of anti-outgroup collective action intentions. As the effect of blatant dehumanization was significantly stronger, Wald χ 2 (1) = 7.89, p = .005, H2a received firm support. At the same time, modern homonegativity but not blatant dehumanization was a negative predictor of participants' willingness to engage in pro-outgroup collective action (Wald χ 2 (1) = 68.29, p < .001), supporting H2b.
Next, we assessed the moderating effects of intergroup contact and social conservatism. In line with H4a, intergroup contact moderated the association between the anti-LGBT campaign and blatant dehumanization. As expected, the positive association between anti-LGBT campaign and blatant dehumanization was the strongest at low (−1 SD) levels of intergroup contact, B = 1.08, SE = 0.16, 95% CI [0.77, 1.40], and became weaker at medium (M), B = 0.87, SE = 0.10, 95% CI [0.67, 1.07], and high (+ 1 SD) levels of this variable, B = 0.65, SE = 0.14, 95% CI [0.38, 0.91]. Contrary to H4b, intergroup contact did not moderate the effect of the anti-LGBT campaign on modern homonegativity. Correlations obtained in 2018 and 2019 are presented below and above the diagonal, respectively. Correlation coefficients that differed in size (p < .05) between the years presented in bold *** p < .001; **p < .01; *p < .05 In accordance with H5a, social conservatism moderated the association between anti-LGBT campaign and blatant dehumanization, such that the positive relationship between the focal predictor and blatant dehumanization was stronger for individuals with a high (+ 1 SD) level of social conservatism, B = 1.14, SE = 0.17, 95% CI Importantly, for individuals who scored especially low (< −1.05 SD) on the measure of social conservatism, anti-LGBT campaign served as a significant negative predictor of modern homonegativity, suggesting that highly liberal participants were less likely to express modern homonegativity in response to the government campaign.

Discussion
In the last decade, illiberal parties have risen to power in many countries. They gained widespread support in large part by stoking up hatred of minority groups, such as immigrants, Jews, and LGBT + people, and presenting such groups as a threat to the nation (e.g., Applebaum, 2020). In the context of Law and Justice's anti-LGBT campaign in Poland, this research investigated the impact of such rhetoric on members of the majority group. Specifically, we aimed to identify factors that explain the extent to which majority group members embrace or reject these anti-minority narratives, focusing on levels of contact with the minority group and individual differences in ideological orientations. Our analysis of representative data collected just before and after the beginning of the anti-LGBT campaign suggests that, overall, the campaign resulted in a worsening of attitudes, as indicated by greater blatant dehumanization of, and modern prejudice against, LGBT people in 2019 compared to 2018. Interestingly, there was an increase in both anti-and pro-minority collective action intentions during this time. This highlights that, while anti-minority rhetoric has the intended effect of mobilizing the population against the minority, there is also resistance and an increased display of solidarity.
Our findings consistently demonstrate that this resistance is dependent upon individual differences in social conservatism. While we observed an increase in dehumanization and modern homonegativity as a function of the anti-LGBT campaign (i.e., time) at high levels of social conservatism, these effects became nonsignificant (or even reversed) at low levels of this variable. Moreover, willingness to engage in prominority collective action increased over this time period at lower levels of social conservatism. To further harness these findings, future research should investigate the psychological processes that underlie this resistance in parts of society.
Drawing on Radke et al.'s (2020) distinction between morality-and identity-based motivations of solidarity, we suggest that future work might focus on these two groups of processes. On one hand, it might be that resistance emerges from the fact that the attack on sexual and gender minorities violates liberal morality. One way to address this would involve including measures of moral foundations (i.e., fairness/reciprocity and harm/care; Graham et al., 2009) in future studies. On the other hand, it is plausible that the moderating role of social conservatism can be attributed to social identity processes. While the ideology and values represented by Law and Justice are likely to align with group understandings (i.e., the idea of "Polishness") among those high in social conservatism, resulting in a salient shared social identity and consequently social influence (Haslam & Reicher, 2012), the opposite is likely to be true among segments of the population low in social conservatism. Thus, individuals low in social conservatism may have been motivated to resist in part by a desire to distinguish their group from the authority and maintain positive distinctiveness in relation to them. Future work should include direct measures of identification and group-based motivations and could also explore how the above processes are interlinked (i.e., shared identity as a basis of morality; Haslam & Reicher, 2012).
There was less consistent support for intergroup contact as a moderator. As predicted, while the anti-LGBT campaign (i.e., time) was a strong positive predictor of blatant dehumanization at low levels of contact, its effect was weaker for participants who declared having numerous homosexual friends or acquaintances. However, unexpectedly, contact did not moderate the effects that the campaign exerted on modern homonegativity and, further, willingness to take pro-outgroup collective action. We believe that these null results may be at least partially attributed to the negative association between intergroup contact and social conservatism. It is possible that social conservatism emerged as a stronger moderator because of a sequential relation between intergroup contact and political ideology. The "cognitive liberalization hypothesis" (Hodson et al., 2018) suggests that contact does not only impact on specific outgroup attitudes but also shapes broader beliefs about how society should operate. Thus, one might expect that individuals who have numerous links with minority group members over time develop a more pluralist, socially liberal outlook, which in turn shapes their responses to attempts to stoke fear of minorities. Future research should employ longitudinal designs to establish this sequential relationship more conclusively and to assess how this liberalization process impacts on responses to illiberal rhetoric. Such work could additionally examine the emergence of a shared identity with the 1 3 minority group as a key psychological process that develops through contact and determines whether the majority group aligns with the minority against the authority (Cikara, 2021;Subašić et al., 2008).
A secondary aim of this research was to examine whether blatant dehumanization and modern homonegativity differentially predict pro-and anti-minority collective action. We expected that extremely hostile beliefs, such as blatant dehumanization, would be more predictive of anti-minority action than more subtle forms of prejudice, such as modern homonegativity, while the opposite should be true for prominority collective action. This is because actions harming an outgroup are often legitimized through the explicit denigration of that group (e.g., Kteily et al., 2015), while actions that support a minority are motivated by a sense of injustice (Leach et al., 2006). These ideas were supported by the data. Compared to modern homonegativity, blatant dehumanization was a stronger positive predictor of anti-outgroup collective action intentions and willingness to engage in pro-outgroup collective action was predicted negatively by modern homonegativity but not by blatant dehumanization.

Contributions and Limitations
The present research enhances our understanding of the psychological effects of anti-minority rhetoric currently observed in far-right populist regimes in general (see Applebaum, 2020;Snyder, 2018) and Polish politics in particular (e.g., Korolczuk & Graff, 2018;Żuk & Żuk, 2020). Importantly, our findings further underline that members of the majority group are not passive recipients who unquestioningly accept and act upon the authority's narrative (echoing Haslam & Reicher's, 2012 reinterpretation of classic studies on conformity and obedience). Rather, the extent to which individuals conform or resist the authorities is dependent on their ideological predispositions, which, as argued above, likely impact their sense of shared social identity with the authority. Resistance also depends on, though to a lesser extent, individuals' own experiences with the targeted minority group.
To our knowledge, no study has thus far determined the factors that make majority members receptive or immune to a threat-inducing rhetoric adopted by the authority in a realistic context, such as the current anti-LGBT campaign in Poland. Together these findings contribute to work on the role of political ideology in intergroup relations (e.g., Jost, 2017) and support emerging findings that intergroup contact may operate as a buffer that neutralizes established antecedents of prejudice (e.g., Visintin et al., 2020). Going beyond previous work, however, our study is the first to demonstrate that individuals low in social conservatism may engage in the defense of this outgroup when it is attacked by a third party (see also Osborne et al., 2019).
Finally, our findings contribute to our understanding of the antecedents of collective action aimed to harm disadvantaged groups, a question that has received little attention in the literature (e.g., Górska et al., 2022a, b). By adopting a more granular perspective on negative outgroup-directed attitudes, we demonstrated that the predictors of pro-and anti-outgroup collective action do not simply vary in direction but also in content. Specifically, while individuals manifesting high blatant dehumanization are likely to take anti-outgroup collective action, low levels of blatant dehumanization do not necessarily translate into minority support. This requires explicit recognition of inequality and unfairness (i.e., low modern homonegativity). Overall, these findings highlight the importance of examining pro-and anti-minority action separately, and suggest that the existing models of collective action, developed primarily in the context of disadvantaged groups and solidarity among advantaged groups, might not be fully transferable to anti-minority action.
We also acknowledge several limitations of our research. While our quasi-experimental approach allowed us to observe how these mechanisms unfold in a real-life context, this approach does not allow us to draw firm causal conclusions about the impact of threat-inducing rhetoric on prejudice and collective action. Thus, our approach needs to be supplemented with controlled experimental studies that manipulate exposure to threatening anti-minority rhetoric. In a similar vein, despite the causal order demonstrated in the previous research (Górska et al., 2022a, b), it is conceivable that participation in anti-or pro-outgroup collective action entails changes in the attitudes toward the outgroup. Further longitudinal or experimental studies would allow to test this possibility.
Furthermore, the high costs of conducting a representative survey prohibited the use of more comprehensive and reliable measures of our constructs of interest. This could contribute to our failure to obtain some of the expected effects.
In order to determine whether our findings generalize to other contexts, more research is needed to replicate our results in other political settings, including other illiberal democracies. Furthermore, it would be informative to examine these processes in contexts where an illiberal left-wing government utilizes anti-minority rhetoric to mobilize support, especially to disentangle different motivations underlying solidarity (Radke et al., 2020). Defense of the minority irrespective of authorities' political affiliation would suggest that liberals act in accordance with their moral principles (see Haidt et al., 2009). Joining the anti-minority campaign would mean that their behavior is determined mostly by ingroup concerns.

Conclusions
The current research shows that cis-heterosexual majority members' responses to the anti-LGBT campaigns inspired by far-right authority depend on social conservatism and intergroup contact. As some members of cis-heterosexual majority are particularly susceptible to the threat-inducing anti-LGBT narratives, right-wing politicians may cynically attack LGBT community to mobilize their voters, regardless of the psychological and physical costs paid by targeted minorities. A differentiated understanding of the psychological processes underlying anti-and pro-minority action in response to such anti-LGBT campaigns has implications for attempts to anticipate and interfere with attacks on minority groups (e.g., by monitoring, penalizing, and counter-arguing dehumanizing speech) and for efforts to effectively mobilize public support for minority groups (by appealing to moral principles and highlighting injustices). Moreover, the political efficiency of anti-LGBT narratives can be reduced by creating opportunities that facilitate positive contact between LGBT minority and cis-heterosexual majority members.
Funding Preparation of this manuscript and the research it reports were supported by the National Science Centre (Poland) 2017/26/M/ HS6/00689 and 2017/25/HS6/01116 grants conferred to the first author.

Availability of Data and Code
Data and code necessary to replicate present analyses are stored in the Open Science Framework repository: https:// osf. io/ 8c6me/? view_ only= 6edf5 51e4e 8c413 9a0dd 25475 f21af 4b. Neither the study nor the analyses have been preregistered.

Conflict of Interest The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.