The Impact of Masculinity Beliefs and Political Ideologies on Men’s Backlash Against Non-Traditional Men: The Moderating Role of Perceived Men’s Feminization

Traditional masculinity norms are still prevalent in our societies. As a result, men who deviate from these norms face adverse reactions (i.e., backlash), mainly from other men. The present research investigated whether the perceived threat to gender status quo accounts for this phenomenon. In two studies using a sample of heterosexual men (N total = 338), we measured male participants’ endorsement of traditional masculinity beliefs and their political orientation. As a means of examining the role of threat to the gender status quo, we then manipulated whether traditional masculinity norms remained stable (stability) or changed across time (men’s feminization). Finally, we assessed participants’ evaluation of men who deviate from traditional masculinity norms (i.e., a backlash against a non-traditional man). This target was either compared to a traditional man (Study 1) or a traditional woman (Study 2). The general hypothesis was that men who are strongly motivated to maintain the gender status quo (i.e., those who endorse traditional masculinity beliefs to a higher extent or support right-wing political ideologies) should show greater backlash, particularly when the gender status quo is threatened (i.e., in the men’s feminization condition). The results of a small-scale meta-analysis supported our hypothesis. We discuss the impact of these findings on the gender literature.


The Role of Perceived Men's Feminization
Contextual factors can also shape men's tendency to sanction non-traditional men. Since the second half of the 20th century, Western societies have seen an evolution towards more gender equality, which has been associated with changes in gender roles and norms that challenge the gender status quo. As an illustration, men have been increasingly involved in activities and behaviors that are traditionally associated with women (Champagne et al., 2015;Dotti Sani, 2014). Even though such a feminization of men appears to be relatively weak and may be unnoticed in some contexts (Barbieri et al., 2017;Haines et al., 2016), people generally perceive that masculinity is changing and men are becoming more feminine over time (Diekman & Eagly, 2000;Lopez Zafra & Garcia-Retamero, 2011). These normative changes have led some people to believe that masculinity is 'in crisis' (Dupuis-Déri, 2018).
Accordingly, one might expect that acceptance of nontraditional men increases as a result of such changes in masculinity norms. However, research has instead shown that perceived men's feminization can result in defensive reactions aimed at protecting the challenged norm, especially among men who strongly endorse these traditional norms. Indeed, the perceived feminization of men challenges the anti-femininity norm, which is the core definition of masculinity (Brannon, 1976;McCreary, 1994), and threatens men's positive social identity (Tajfel & Turner, 1979). Therefore, the perception that men are becoming more feminine erodes men's positive distinctiveness as a group. To reinforce the challenged anti-femininity norm, men can therefore react by either reaffirming their own masculinity or traditional masculinity norms in general.
Indeed, research has shown that men who are led to believe that their gender in-group (i.e., other men) is becoming more feminine over time report a greater likelihood of performing a range of stereotypically masculine behaviors (Bosson & Michniewicz, 2013;Study 5). This is particularly the case among men who strongly endorse stereotypically masculine characteristics (Babl, 1979). In the same vein, we contend that perceived men's feminization may also trigger derogation of those who are perceived as contributing to the decline of the anti-femininity norm and therefore threatening the gender status quo (see Marques et al., 2001). Existing research has indeed shown that perceived men's feminization increases heterosexual men's endorsement of the heterosexuality norm; that is, they distance themselves psychologically from gay men and express more negative attitudes towards homosexuality . Furthermore, this tendency is accounted for by heterosexual men's motivation to increase in-group distinctiveness (i.e., to psychologically exclude gay men from the male in-group; Iacoviello et al., 2019) and is greater among participants who strongly endorse traditional masculinity norms Valsecchi et al., 2020).
In line with these findings, we argue that men are not only likely to face backlash because of their sexual orientation (i.e., because they deviate from the heterosexuality norm) but also for their behaviors and career choices. Indeed, men who engage in stereotypically feminine behaviors (e.g., who display their emotions and vulnerability) or occupations (e.g., who work as a male nurse or secretary) could be seen as harming the anti-femininity norm. Therefore, we expect that men who are motivated to protect gender status quo (e.g., traditional and conservative men) would backlash more against non-traditional men, especially when the gender status quo is threatened by the perception that men (as a group) are becoming more feminine (i.e., there is a normative change).

Current Research
We aimed to test the interactive impact of dispositional and contextual factors in the emergence of backlash towards non-traditional men. More specifically, we examined the role of individual differences in men's motivation to maintain the gender status quo, especially when it is threatened by men's feminization. In two studies, we investigated backlash effects by comparing men's attitudes towards a non-traditional man to either attitudes towards a traditional man (Study 1) or towards a woman who conforms to gender norms (Study 2). To increase both the potential generalization of the results and ecological validity, we varied the dimension on which the male target was described as 'non-traditional'. In Study 1, the non-traditional man displayed his emotions and vulnerability, therefore deviating from the restricted emotionality norm of masculinity (see Costrich et al., 1975;Derlega & Chaikin, 1976). In Study 2, the non-traditional man was described as a candidate applying for a job that is traditionally occupied by women, thus deviating from the anti-femininity norm (see Moss-Racusin & Johnson, 2016). Moreover, we also varied the dimension on which the targets were evaluated. In Study 1, we measured desired social proximity with the target, trait ratings, and sympathy towards the target. In Study 2, we measured target likeability, intention to hire, and suggested salary. Finally, in each study, we manipulated the perceived gender norm (stability vs. men's feminization).
With these two studies, we aimed at testing the main hypothesis that backlash towards non-traditional men stems from men's motivation to protect the gender status quo. More specifically, Hypothesis 1: Backlash against the non-traditional male target (vs. the traditional target) should increase as male participants strongly endorse traditional masculinity norms, and this tendency should be particularly strong in the men's feminization (vs. stability) condition (H1).
Hypothesis 2: Backlash against the non-traditional male target (vs. the traditional target) should increase as male participants strongly endorse right-wing ideologies, and this tendency should be particularly strong in the men's feminization (vs. stability) condition (H2).
The two studies were conducted in the context of an undergraduate research seminar (practical work) in social psychology; as such, these two studies were not submitted for approval to the ethics committee of the UNIGE. As part of this research seminar, students assumed the role of experimenters after having been instructed about the ethical issues and the ethical guidelines applying to experimental research. The following ethical guidelines were applied: voluntary participation; anonymized data collection; informed consent, including the possibility of stopping participation at any time; full debriefing, including about deceptions in order to foresee and minimize unintended consequences. Databases for both Studies 1 and 2 can be found online on the Open Science Platform: https://osf.io/y4tb2/?view_only=3774def9e58f4837a71 cdb4240692330.

Study 1
Restrictive emotionality is a dimension of traditional masculinity norms (e.g., Levant et al., 2007;Wong et al., 2013) and is defined as the men's tendency to inhibit most of their feelings (Jansz, 2000). Therefore menand even boys-who openly show their emotions and vulnerability can face backlash (Costrich et al., 1975;Kane, 2006). Participants of the present study were asked to rate a man showing his emotions (i.e., the non-traditional man) or a man hiding them (i.e., the traditional man). We postulated that backlash against a man displaying his emotions (vs. a man hiding them) should increase as a function of men's endorsement of traditional masculinity norms or right-wing ideologies, particularly in the condition of men's feminization (vs. stability). The target's evaluation was assessed through three measures: desired social proximity, trait rating, and sympathy.

Participants
In the context of an undergraduate research seminar, students assumed the role of experimenters and were expected to recruit about 200 male participants. Participants were invited to complete an online questionnaire on 'several social issues, including the situation of men and women'. A total of 236 participants volunteered to fill in the questionnaire. Because some items of traditional masculinity are related to opinions towards sexual minorities, we removed participants that reported being not heterosexual (n = 46). More specifically, the dimension of 'negativity toward sexual minorities' is only relevant for heterosexual men. Being prejudiced against gay people implies motivations that could seem absurd for gay men or, at the very least, have a completely different meaning for them than for heterosexual men. The final sample was therefore composed of 190 French-speaking participants, all heterosexual males (M age = 22.94 years, SD = 6.26). Most of them (83.2%) were French (other frequently mentioned nationalities were Swiss and Belgian) and 51.1% were students. A sensitivity power analysis conducted on G*Power for a multiple linear regression model with seven predictors (i.e., including all main effects and interactions), assuming an α of 0.05 (two-tailed) and a power of 0.80, revealed that our final sample (N = 190) was powered enough to detect a small effect size (f 2 = 0.04).

Procedure
Participants first completed the traditional masculinity beliefs scale and then were presented with an excerpt of an article manipulating men's feminization. Thereafter, they answered the related comprehension-check questions and were asked to provide a real-life example that is in line with the text they had just read. They then read a vignette describing either a traditional or a nontraditional man. They evaluated the target using three different measures: social proximity, trait rating, and sympathy. Then, participants answered items assessing their own perception of the men's norm. Finally, they provided their demographic information, including their sexual orientation and their political orientation, and were fully debriefed and thanked for their participation. Unless otherwise mentioned, answers to all questions in this study were collected on seven-point scales, ranging from 1 (Completely disagree) to 7 (Completely agree).

Measure of Traditional Masculinity Beliefs
Traditional masculinity beliefs were measured using a French-adapted version of the 21-item Male Role Norm Inventory-Short Form (MRNI-SF; Levant et al., 2013). The scale encompasses seven dimensions (i.e., restrictive emotionality, self-reliance through mechanical skills, negativity toward sexual minorities, avoidance of femininity, importance of sex, dominance, and toughness). Sample items include the following: 'Men should be detached in emotionally charged situations' and 'Men should watch football games instead of soap operas.' A score was computed by averaging the responses of these 21 items (M = 2.11, SD = 1.00, α = 0.92).

Measure of Political Orientation
Political orientation was measured using a single item (from 1 = Very left-wing to 7 = Very right-wing; M = 3.37, SD = 1.44). Correlation with traditional masculinity beliefs was significant and positive, r(185) = 0.36, p < 0.001. 1

Men's Gender Norm Manipulation
Participants were presented with a short excerpt of an alleged news article describing the results of a longitudinal study that examined 'men's personalities and behaviors from 1957 to 2017'. In the feminization condition [stability condition in brackets], the article reported that this recent publication showed that 'men tend to become more feminine over time [men are just as masculine as ever]' and that, according to researchers, these results suggest that Men's way of being is changing and that there is a real feminization of men [men's way of being has not changed over time]. The distinction between 'being a man' and 'being a woman' tends to dis-appear [the distinction between 'being a man' and 'being a woman' remains clear].
To check that participants understood the text, they answered three statements about the content of the excerpt they had just read on a seven-point scale (1 = Not at all to 7 = Absolutely): 'Men's way of being has changed in recent years,' 'Today, men are more feminine than ever,' and 'Today, men are still as masculine as they were in the past (reversed score).' After recoding, a score was computed by averaging the responses of these three items (higher scores reflect greater men's feminization; M = 3.93, SD = 1.69, α = 0.80). To reinforce the manipulation, participants were further told, 'These findings certainly echo something you have already seen for yourself,' and they were asked to 'provide a short example that illustrates the study's findings' in an open-ended question.

Target
Participants were presented with a vignette describing a traffic accident that involved a heterosexual couple. The reaction of the man driving the car was manipulated so that, in one condition, his behavior was consistent with traditional masculinity norms (the traditional man condition), and in the other condition, it was at odds with traditional masculinity norms (the non-traditional man condition). More specifically, all participants read the following: Alerted by witnesses, the police arrive on the scene of the accident. Alexandre and his wife are still trying to come to their senses. His whole body is shaking, his heart beats fast and he feels nauseous.
In the traditional man condition, the vignette continued: Despite his emotional state, he tries not to show his emotions and tells his wife that he feels willing to get out of the vehicle and that she can stay seated. He gets out of the car and starts the conversation with the police in order to tell them the facts of the accident.
In the non-traditional man condition, the vignette instead continued: Overwhelmed by his emotions, he says to his wife that he feels unable to get out of the vehicle for now and asks her if she feels willing to do it. She accepts, gets out of the car and engages the conversation with the police to tell them the facts of the accident. 2

Social Proximity
Desired social proximity was assessed through a six-item scale adapted from Crandall's (1991) scale. Sample items included the following: 'To what extent … would you like to have Alexandre as close friend?' '… would you like to have Alexandre as your neighbor?' and '… would you try to avoid Alexandre? (reverse score).' After recoding, a score was computed by averaging the responses of these six items (higher scores reflect desired social proximity; M = 4.99, SD = 1.04, α = 0.87).

Trait Rating
Participants were presented with eight characteristics (i.e., welcoming, humble, naive, weak, self-assertive, independent, arrogant, and aggressive) and asked to rate the extent to which each one of them best described the man introduced in the vignette (i.e., Alexandre). As our hypotheses are focused on the judgments' valence, the measure was composed of four positive and four negative traits. Also, and since the literature has shown that the gender dimension (warmth vs. competence) can be relevant when examining backlash dynamics (Rudman & Mescher, 2013), we wanted to prevent any confound between the valence of the items and the gender dimension. This would also be useful to check for any impact of the gender dimension. Therefore, we adapted items from Rudman and Mescher (2013) to devise a measure that would keep these two dimensions independent from each other.
A principal component analysis provided three dimensions with eigenvalues higher than 1. The four negative characteristics loaded substantially and positively on the first dimension (loadings > 0.62). The two positive and agentic characteristics (i.e., independent and self-assertive) loaded substantially and positively on the second dimension (loadings > 0.58), and the two positive and communal characteristics (i.e., humble and welcoming) loaded substantially and positively on the third dimension (loadings > 0.59). This complexity about the positive characteristics reflected on their poor reliability (α = 0.42, M = 4.48, SD = 0.86), while the reliability of the negative characteristics was satisfying (α = 0.75, M = 2.33, SD = 1.10). Despite this, we computed the final rating score by subtracting the index of positive characteristics from the index of negative characteristics so that higher scores indicate positive ratings.

Sympathy
We measured sympathy for the target with a self-made six-item scale. Sample items included the following: 'I feel close to him,' 'I feel empathy for him,' and 'I feel respect for him' (M = 4.71, SD = 1.11, α = 0.83). The score of sympathy was positively and significantly correlated with both social proximity, r(189) = 0.76, p < 0.001, and trait rating, r(189) = 0.61, p < 0.001. Social proximity and trait rating also correlated positively and significantly, r(189) = 0.62, p < 0.001.

Perception of Men's Gender Norm
To check whether the men's norm manipulation influenced participants' personal opinion, we assessed their perception of the men's norm using the same three items as in the manipulation-check measure. The difference was the main instructions, which stated, 'Please indicate your personal agreement with the following statements.' After recoding, a score was computed by averaging the responses of the three items (higher scores reflect greater men's feminization; M = 4.29, SD = 1.24, α = 0.59).

Comprehension Check and Perception of Men's Gender Norm
First, a one-way analysis of variance (ANOVA) with men's gender norm (stability vs. feminization) as the independent variable was performed on the comprehension check. Results indicated that participants acknowledged a greater feminization in the feminization condition (M = 4.62, SE = 0.16) than in the stability condition (M = 3.20, SE = 0.16), F(1,187) = 40.50, p < 0.001, η p 2 = 0.18. The same analysis performed on participants' perception of the men's norm also revealed a main effect of the men's norm manipulation, F(1,188) = 5.21, p = 0.023, η p 2 = 0.03. Participants perceived greater men's feminization in the feminization condition (M = 4.48, SE = 0.13) than in the stability condition (M = 4.08, SE = 0.12). Together, these results indicated that the men's norm manipulation was effective in shaping participants' perception.

Traditional Masculinity as a Moderator
We ran a series of linear regression analyses on the three dependent variables (i.e., social proximity, trait rating, and sympathy) with gender norm (coded -1 for stability and +1 for feminization), target (coded -1 for non-traditional man and +1 for traditional man), traditional masculinity beliefs (centered continuous variable), and their interactions as predictors. Results of the regression analyses are presented in Table 1 On each outcome, this effect was qualified by a Target × Traditional masculinity beliefs interaction, indicating that the preference for the non-traditional man (over the traditional man) tended to decrease as a function of traditional masculinity beliefs. At odds with H1, this tendency did not differ as a function of men's norm, as indicated by the non-significant Norm × Target × Traditional masculinity beliefs.

Political Orientation as a Moderator
The same series of linear regressions were performed with political orientation instead of traditional masculinity beliefs as a predictor. Results are reported in Table 2. The analyses showed that the tendency to prefer the nontraditional man over the traditional man decreased as a function of participants being right-wing, even though this was not significant on the trait rating. Moreover, the Norm × Target × Political orientation interaction was significant on both trait rating and sympathy. As illustrated in Figures 1 and 2, and consistent with H2, the tendency for right-wing participants to show increased backlash towards the non-traditional man (vs. the traditional man)

Discussion
The present study first showed that the traditional man was rated more negatively than the non-traditional man. Therefore, these findings did not replicate the classic backlash effect observed in the literature. However, this preference for the non-traditional man decreased as participants endorsed traditional masculinity beliefs or right-wing ideologies. Second, results generally did not support H1, as the tendency for participants who strongly endorse traditional masculinity to be more severe toward the non-traditional man was not moderated by perceived men's feminization. There was no evidence of this predicted effect on any of the three dependent variables. Third, and finally, the findings spoke in favor of H2, as the tendency for right-wing participants to evaluate more negatively the non-traditional man (vs. the traditional man) was stronger in the men's feminization condition (vs. the stability condition). This general effect indeed was significant on two outcomes (i.e., trait rating and sympathy).

Study 2
Study 2 aimed at replicating Study 1's findings in a different setting and to overcome some of the potential limitations. Study 1 indeed compared evaluations of a non-traditional (i.e., more feminine) man to evaluations of a traditional (i.e., less feminine) man. Methodologically speaking, Study 1's design cannot rule out the alternative interpretation that men's backlash is directed towards femininity per se instead of deviance from traditional masculinity (as we argue here). Indeed, as non-traditional men are likely to be characterized by traditionally feminine characteristics, it could be argued that femininity per se, regardless of the target's gender, is despised by heterosexual men who highly endorse traditional masculinity or right-wing ideologies.   To address this potential issue, participants in the present study rated a man applying for a traditionally feminine position (i.e., a non-traditional man) or a woman applying for the same position (i.e., a traditional woman). We employed this design because only men (but not women) who apply for a job traditionally performed by women should face backlash (see Moss-Racusin & Johnson, 2016). We hypothesized that the preference for a traditional woman over a non-traditional man should increase as a function of men's endorsement of traditional masculinity norms or right-wing ideologies, particularly in the men's feminization condition (vs. stability).

Participants
As for the previous study, this study was conducted in the context of an undergraduate research seminar in which students were expected to recruit about 200 male participants. Despite this threshold not being met, the recruitment stopped after two weeks because of the seminar's schedule. Participants were invited to participate in a study on 'several social issues'. A total of 170 participants filled in the online questionnaire. We removed 20 participants from the analyses because they reported not being heterosexual and 2 participants because they failed the attention check. This attention-check question displayed a seven-point Likert scale and specifically asked participants not to answer the question and go to the following page. The final sample was therefore composed of 148 Frenchspeaking participants, all heterosexual male (52.7% of students, M age = 29.00 years, SD = 12.77). A sensitivity power analysis conducted on G*Power for a multiple linear regression model with seven predictors (i.e., including all main effects and interactions), assuming an α of 0.05 (two-tailed) and a power of 0.80, revealed that our final sample (N = 148) was powered enough to detect a small effect size (f 2 = 0.05).

Procedure
Participants first provided their demographic information (e.g., gender, age, sexual orientation, political orientation). This was necessary to filter participants for their gender and age, allowing only male participants above the age of 18 to participate. They then completed the traditional masculinity beliefs scale, read a short text manipulating men's norm, and answered the related comprehension-check measure. Afterwards, they read a vignette either describing a non-traditional man or a traditional woman and rated this target through three different measures: likeability, intention to hire, and suggested salary. They then answered the Traditional Gender Roles scale adapted from Morgan (1996) 4 and the attention-check item. At the end of the questionnaire, they were fully debriefed about the purpose of the study and thanked. Unless otherwise noted, responses to all questions were collected on seven-point scales, ranging from 1 (Completely disagree) to 7 (Completely agree).

Measures of Traditional Masculinity Beliefs
As in Study 1, traditional masculinity beliefs were assessed through the French-adapted version of the 21-item MRNI-SF scale (M = 2.48, SD = 1.03, α = 0.92).

Political Orientation
Political orientation was assessed using a Likert scale, going from Extremely left-wing (coded 1) to Extremely right-wing (coded 7). At odds with the preceding study, we added the option No political orientation, which was coded as 4, that is, the mid-scale point (see Aelenei et al., 2020; M = 3.84, SD = 1.47). 5 Political orientation and traditional masculinity beliefs were positively and significantly correlated, r(147) = 0.261, p < 001.

Men's Gender Norm Manipulation
In order to manipulate gender norm, participants read a one-page text (ostensibly published in a scientific journal of sociology) adapted from previous research (Borinca et al., 2020;Falomir-Pichastor et al., 2019;Iacoviello et al., 2019) that summarized the results of an international study conducted in Western countries about the evolution of men's masculinity in society. In the stability condition, the results showed that 'the general trend is a clear lack of change in the way of being a man. Indeed, in general, men tend to stay masculine on all these personal and social dimensions.' A few examples were then provided to illustrate this tendency, such as Men still give a lot of importance to their careers, while they do not invest much in traditionally female tasks, whether it is in the professional domain (for example, there are no more male nurses or cashiers than before) or at home (in household tasks, childcare, etc.).
The article concluded, 'In sum, there is no real "feminization of men", and the distinction between masculinity (being a man) and femininity (being a woman) is still essential.' In the feminization condition, the results showed that 'the general trend is towards a clear evolution in the way of being a man. Indeed, in general, men tend to feminize on all these personal and social dimensions.' A few examples were then provided, such as Men tend to give less importance to their careers, and spend more time in traditionally female tasks, whether it is in the professional domain (for example, there are more male nurses or cashiers than before) or at home (in household tasks, childcare, etc.).
The article concluded, 'In sum, we are facing a real "feminization of men", and the distinction between masculinity (being a man) and femininity (being a woman) tends to disappear.' As in Study 1, participants then answered the same three manipulation-check items (M = 4.09, SD = 1.69, α = 0.87) and provided a short example that illustrated the study's findings.

Target
Participants were informed that we were also interested in their evaluation of candidates applying for a specific position. They were thus presented with a vignette describing an available position as a secretary in the field of education. The vignette described the responsibilities of the position and the required skills. Participants were then told that the employer had received several applications, and they were presented with one of them, which consisted of the CV of the candidate. According to the experimental condition, the candidate was either a man (i.e., Alexandre; the non-traditional man) or a woman (i.e., Alexandra; the traditional woman). All other details of the CV were identical across experimental conditions.

Likeability
Participants were asked to rate how likable the candidate was on a seven-item scale. Sample items included 'Do you like this person?' 'Does this person look competent?' and 'Do you believe this person would be liked by his/her colleagues?' (M = 5.31, SD = 0.98, α = 0.89).

Intention to Hire
Participants answered the following item: 'Would you personally be ok to hire this person?' (1 = Not at all) to (7 = Totally).

Salary
Participants indicated the monthly wage (in Swiss francs) that they thought was appropriate for the target on the following scale: 1 = less than 2000; 2 = between 2000 and 3000; 3 = between 3000 and 4000; 4 = between 4000 and 5000; 5 = between 5000 and 6000; 6 = between 6000 and 7000; or 7 = more than 7000.

Comprehension Check
A one-way ANOVA with men's gender norm (stability vs. feminization) as the independent variable was performed on the comprehension-check score and indicated that participants acknowledged a greater feminization in the feminization condition (M = 4.61, SE = 0.19) than in the stability condition (M = 3.59, SE = 0.19), F(1,146) = 14.91, p < 0.001, η p 2 = 0.09.

Traditional Masculinity as a Moderator
We ran linear regression analyses on the three dependent variables (i.e., likeability, intention to hire, and suggested salary) with gender norm (coded -1 for stability and +1 for feminization), target (coded -1 for non-traditional man and +1 for traditional woman), traditional masculinity beliefs (centered continuous variable), and their interactions as predictors. Results of the regression analyses are presented in Table 3   The same series of regressions were performed with political orientation instead of traditional masculinity beliefs as a predictor. Results are reported in Table 4. Apart from the above-mentioned main effects of the target, the only notable effect was a marginal Norm × Target × Political orientation interaction on the likeability measure. In line with H2, the tendency for right-wing participants to show increased backlash towards the nontraditional man (vs. the traditional woman) was stronger in the feminization condition, B = 0.14, SE = 0.08, t = 1.82, p = 0.071, 95% CI [-0.01, 0.29] than in the stability condition, B = -0.05, SE = 0.09, t = -0.60, p = 0.549, 95% CI [-0.22, 0.12].

Discussion
Unlike the previous study, Study 2 highlighted an overall backlash against the non-traditional man. Indeed, this latter target tended to be evaluated more negatively than the traditional woman. This apparent inconsistency is discussed in the general discussion. Moreover, the findings provided partial support for H1 and H2, as the expected effects were only observed on the likeability measure. Indeed, the backlash effect increased as participants endorsed traditional masculinity norms or right-wing ideologies, and this was particularly true in the men's feminization condition. We have no explanation about why the expected pattern of results only appeared on the target likeability measure. We could however speculate that this was due to the order of presentation of the measures. As the likeability measure came first, backlash motives could have been expressed more eagerly.

Small-Scale Meta-Analyses
Across the two studies and the six outcomes, analyses showed mixed results. However, since the pattern of interaction was consistent across studies and outcomes (even though not significant in all instances), we conducted a random-effect meta-analysis in R, using the Metafor package (Viechtbauer, 2010) to investigate the reliability of the Target × Norm × Traditional masculinity interac- tion across the two studies. We used the Sidik-Jonkman method, which is preferred for meta-analyses including a small number of studies (Inthout et al., 2014). Results of the small-scale meta-analysis are displayed in Because endorsement of traditional masculinity and political orientation shared some common variance (correlations varied between 0.26 and 0.49 across studies), we did the same meta-analyses on the results while controlling for the other moderator (i.e., including it in the whole design). The analysis showed that the Target × Norm × Traditional masculinity interaction was globally not significant when controlling for political orientation, B = 0.07, SE = 0.04, z = 1.50, p = 0.133, 95% CI [-0.02, 0.16]; T² = 0, I² = 0.00%, Q (df = 5) = 3.91, p = 0.56. Conversely, the Target × Norm × Political orientation interaction remained globally reliable when controlling for the endorsement of masculinity beliefs, B = 0.11, SE = 0.03, z = 3.83, p < 0.001, 95% CI [0.05, 0.17]; T² = 0, I² = 0.00%, Q (df = 5) = 3.37, p = 0.64.

General Discussion
Across two studies, we examined the reasons underlying (heterosexual) men's backlash against men who deviate from traditional masculinity norms (i.e., nontraditional men). We relied on the status incongruity hypothesis (SIH; Rudman et al., 2012) to postulate that non-traditional men face backlash because they are perceived as a threat to the gender status quo. Accordingly, we ran two experimental studies to test the hypotheses that the preference for a traditional target over a nontraditional man should increase as male participants strongly endorse traditional masculinity norms or rightwing ideologies, and both these tendencies should be particularly strong in the men's feminization condition (H1 and H2). A small-scale meta-analysis on the results of the two studies showed that, overall, the findings provide empirical evidence in support of both H1 and H2 and that the observed variations across the measures in both studies were trivial (as indicated by the low levels of heterogeneity).

Theoretical Implications and Limitations
At odds with the exiting research on the backlash against non-traditional men (e.g., Costrich et al., 1975), Study 1 showed that the non-traditional man was overall evaluated more positively than the traditional man. However, Study 2 revealed the expected backlash pattern, the nontraditional man being overall more negatively evaluated than the traditional woman (e.g., Moss-Racusin & Johnson, 2016;Moss-Racusin et al., 2010). Several explanations could account for this apparent inconsistency. First, the discrepancy could be due to the specific sample of both studies. Indeed, Study 2's participants had a slightly higher score of traditional masculinity beliefs (M = 2.48 vs. 2.11), F(1,334) = 10.87, p = 0.001, η p 2 = 0.03, and were more right-wing (Ms = 3.84 vs. 3.37), F(1,329) = 8.50, p = 0.004, η p 2 = 0.03, than Study 1's participants. Second, it is also possible that occupying a traditionally feminine occupation is perceived as a more severe deviation from traditional masculinity than displaying one's own distress in a traumatic situation. Third, heterosexual men may be aware of the link between toxic masculinity and traditional masculinity in everyday life, and hence they may have tried to appear less biased by rating the non-traditional man better than their stereotypical in-group member (i.e., the traditional man). Finally, another possible explanation may come from the comparison target. Backlash towards a non-traditional man could appear more blatantly when compared to a traditional woman (Study 2) than when compared to a traditional man (Study 1). Indeed, research shows that within-gender comparison contexts decrease gender differences, while between-gender comparison contexts increase them (Guimond et al., 2006). Moreover, because of the within-gender comparison context, participants' gender identity may have been less salient. As a result, they may have felt more empathy for the nontraditional target. Future research should further examine how the comparison target is likely to influence backlash. For instance, it could be interesting to use a within-participant design to make the comparison between the traditional and the non-traditional target more salient.
In the present research, we considered traditional masculinity beliefs and political orientation as two related but independent moderators of backlash tendencies. However, we believe it is important to discuss the relationship between these two constructs. While they are clearly separate theoretical concepts, the former referring to the endorsement of traditional masculinity norms and the latter to a set of values encompassing conservatism and legitimacy of inequalities in general (see Thorisdottir et al., 2007), both share some common variance. We speculate that this interconnection is mainly due to the hegemonic nature of traditional masculinity norms. As traditional masculinity is related to the maintenance of men's privileges over women, the endorsement of this norm mirrors men's conservatism and justification of social inequalities. It is worth noting that the results of the small-scale meta-analysis showed that the overall interaction including political orientation remained significant when controlling for traditional masculinity beliefs, whereas the interaction including traditional masculinity beliefs dropped to non-significance when controlling for political orientation. We may therefore speculate that backlash in reaction to perceived men's feminization is primarily contingent on conservatism motives rather than on masculinity beliefs per se. In other words, men who strongly endorse masculinity beliefs could react to the perceived men's feminization specifically because of their willingness to protect the status quo. Further research is needed to investigate this issue.
Moreover, whereas there is no compelling evidence on the consequences of perceived men's feminization, the mechanism based on social identity threat and motivation to maintain positive in-group distinctiveness has never been adequately tested (see Iacoviello et al., 2019, for initial evidence on this purpose). Therefore, further research is needed to investigate this issue. For instance, future research could measure negative affect or the spontaneous activation of anxiety-related words (see Vandello et al., 2008, Study 4), and examine whether these factors mediate the impact of men's feminization on outcomes related to affirming one's masculinity, such as backlash towards non-traditional men.
The present findings further revealed that the tendency for men holding traditional masculinity beliefs or a rightwing orientation to be negatively inclined towards nontraditional men was particularly strong when they were led to believe that men as a group are becoming more feminine over time. In the present research, we argued that men's feminization is perceived as a threat to men's status and that therefore men who are most attached to the status quo tend to psychologically ostracize those who are contributing to the downfall of traditional masculinity. These results are consistent with subjective group dynamics model (Marques et al., 2001) and speak in favor of the idea that men's feminization is perceived as a threat to the gender status quo that needs to be handled by affirming one's own traditional masculinity (Babl, 1979;Bosson & Michniewicz, 2013) or by derogating gay men Iacoviello et al., 2019;Valsecchi et al., 2020). We may however wonder whether such a defensive strategy would also result in derogating the members of the out-group (i.e., women) who are perceived as contributing to men's loss of status, such as feminist women. Future research should therefore examine the consequences of men's feminization on out-group targets (e.g., by comparing feminist to non-feminist women).
As a further test of the SIH, future research should also compare male and female responses to men's feminization. Indeed, according to the SIH, the backlash against non-traditional men is driven by a motivation to protect the gender status quo. Though there are reasons to believe that even the low-status group (in this case, women) tends to legitimize the hierarchy and is therefore motivated to maintain the status quo (see Jost et al., 2004), such a motivation should be stronger in the high-status group (i.e., men; see Tajfel & Turner, 1979). Accordingly, future research should examine whether backlash against non-traditional men is stronger among men than among women, especially among those who endorse traditional masculinity beliefs and right-wing ideologies the most.
Finally, although the small-scale meta-analysis provides some confidence in the reliability of the present results, it must be acknowledged that the complexity of the design and the multiplicity of the DVs across the two studies may induce some form of caution when interpreting the strength of the effects. Further studies testing the SIH in the context of normative changes are welcome.

Conclusion
In an era where traditional gender norms are being increasingly challenged, it is crucial to understand how people react to these normative changes. It may indeed seem paradoxical that the progressive shift from traditional to modern forms of masculinities would reinforce some people's motivation to protect the gender status quo. The present research specifically focused on backlash tendencies, and it adds to the literature by highlighting who is most likely to react negatively to normative changes and why they do so.

Notes
1 In this and the following study, degrees of freedom can slightly vary across analyses due to missing values. 2 An English translation of the whole vignette can be found in the Appendix. 3 We conducted additional analyses by adding the type of trait (agentic vs. communal) as a within-participant variable in order to examine whether the pattern of interaction was contingent on the type of trait. Results showed that this was not the case, as indicated by the non-significant Norm × Target × Political orientation × Type of traits, F(1,182) = 0.03, p = 0.867, η p 2 = 0.00, and Norm × Target × Traditional masculinity beliefs × Type of traits, F(1,177) = 0.08, p = 0.783, η p 2 = 0.00. 4 In the context of the research seminar, the Traditional Gender Roles scale was added after the backlash DVs for explorative purposes. The aim was to investigate whether the interactions between the norm manipulation and both dispositional variables would affect this score. Results showed that neither the Norm × Traditional masculinity beliefs interaction, B = 0.04, SE = 0.05, t = 0.73, p = 0.471, 95% CI [-0.07, 0.15], nor the Norm × Political orientation interaction, B = -0.06, SE = 0.05, t = -1.33, p = 0.185, 95% CI [-0.15, 0.03], were significant. Since this measure is not directly related to the backlash effect, we do not report the results in the text. Further information can be obtained from the first author. 5 As the results did not differ substantially with and without excluding participants who answered No political orientation, we kept their data in the subsequent analysis with this coding.

Additional Files
The additional files for this article can be found as follows: • Appendix.