Women who consider themselves feminists are more likely to challenge gender-based discrimination and disadvantage through protest, organizing, and advocacy (i.e., collective action; Kelly and Breinlinger 1995; Nelson et al. 2008; Yoder et al. 2011; Zucker 2004; Zucker and Bay-Cheng 2010). Men who identify with feminists are more willing to join their cause (Subašić et al. 2018; White 2008; Wiley et al. 2013). These findings comport with a growing body of scholarship which shows that social identity is an important antecedent of allied activism (Dahling et al. 2016; Thomas et al. 2018; van Zomeren et al. 2011).

Male allies play an important role in the feminist movement,due to their position as “…the primary agents maintaining and supporting sexism and sexist oppression,” (hooks 1984, p. 83). In fact, a number of efforts to prevent violence against women have focused on men by targeting the norms that support gender-based violence (Flood 2011). One is the Mentors in Violence Prevention program, which teaches men to act when they see potential violence against women and encourages men to become peer mentors (Katz 1995; Katz et al. 2011). Another is the White Ribbon campaign, which mobilizes men to pledge their opposition to violence against women and girls (Kaufman 2001). A third effort addresses men’s perception that other men are not willing to confront violence (Berkowitz 2011; Fabiano et al. 2003). Thus, men can—and do—work together to combat sexism.

With their critical focus on men and masculinity norms, these efforts do not address how feminist men can support women when the two groups work together to challenge gender inequality. Male feminist scholars have articulated some guidelines for men who do feminist work with women: They can recognize the privilege afforded them by their gender; they can de-center their own experience and listen to women; they can speak up when they notice sexism—to name a few (Flood 2011; Katz et al. 2011; Kaufman 2001; Kimmel 2016). However, relatively little empirical research has examined how feminist women perceive support from feminist men. This omission is worth addressing because men’s involvement in the feminist movement presents challenges as well as opportunities for women (Flood 2011; but see Messner et al. 2015).

Social psychological research has shown that allies’ support may be unwanted or even harmful. When advantaged group allies work together with members of disadvantaged groups, they can reproduce broader inequalities within activist movements or de-mobilize the disadvantaged group to fight inequality (Becker et al. 2013; Droogendyk et al. 2016). Similarly, feminist scholars have pointed out the thorny situation of feminist men who bring their privilege as men to a movement that seeks to challenge gender inequality (Flood 2011; Messner et al. 2015). Male privilege can amplify men’s voices, recognition, and influence within the feminist movement and therefore reinforce gender inequalities even as feminist men seek to fight it. For these reasons, some members of disadvantaged groups may not want to work together with advantaged group allies or they may prefer some forms of support to others.

In the present studies, we examine how women encounter different styles of support from a feminist man seeking to work with them. Applying predictions from Intergroup Helping as Status Relations Model (Nadler 2002), we argue that feminist women prefer to work with feminist allies who take a backseat and offer partial support (i.e., autonomy-oriented help) over allies who attempt to solve the problem themselves and impose their will on the movement (i.e., dependency-oriented help). We argue that this is because autonomy-oriented support challenges the gender hierarchy by affirming women’s competence, whereas dependency-oriented support reinforces the hierarchy by implying that women are not capable of helping themselves. To paraphrase one feminist scholar, feminist women prefer not to be seen as damsels in distress and they do not see feminist men as the cavalry (Kimmel 2016).

Our studies move beyond the question of how to engage members of advantaged groups in activism. They begin to address what kinds of support members of disadvantaged groups prefer. In so doing, our studies extend research on feminist identification. They show that women who identify as feminists support gender equality within the feminist movement as much as in society-at-large. Our studies also add to the literature on intergroup helping. They test whether members of the disadvantaged group most motivated to achieve social change prefer autonomy- to dependency-oriented help from advantaged group allies. In the following, we review the Intergroup Helping as Status Relations Model. We then apply the model to the context of feminist identification before stating our predictions.

Intergroup Helping as Status Relations

The Intergroup Helping as Status Relations Model (IHSRM; Nadler 2002) argues that, “helping relations can be mechanisms through which groups create, maintain, and change status relations” (Nadler and Halabi 2006, p. 97). The model is based, in part, on the social identity theory of intergroup relations (SIT; Tajfel and Turner 1986). According to SIT, people derive meaning and value from the groups to which they belong and do so by comparing these groups to relevant outgroups. When that comparison is unfavorable, people can engage in a variety of strategies to achieve a positive identity. They can distance themselves from their ingroup (i.e., social mobility), they can change the parameters of the intergroup comparison (i.e., social creativity), or they can take action to improve their ingroup’s relative position (i.e., social competition). Which strategy people choose depends on their perception of the intergroup hierarchy. People are less likely to distance themselves from their ingroup if the boundaries between groups are impermeable or difficult to trespass. Under these conditions, people are more likely to engage in social competition if they perceive the intergroup hierarchy as both illegitimate, or unfair, and unstable, or ripe for change. A broad body of literature supports these claims (see Ellemers et al. 2002; Hornsey 2008, for reviews).

Nadler and colleagues (e.g., Nadler 2002; Nadler and Halabi 2006) recognize that intergroup helping is an important context in which these social identity processes unfold. They argue, first, that most helping involves unequal status relations. The helper has resources to provide and the recipient is in want of them. They argue, second, that not all styles of helping are equal in this regard. Dependency-oriented help, which solves a problem completely, affirms the intergroup hierarchy. It justifies the disadvantaged position of the receiving group by implying that they cannot solve problems on their own and it justifies the advantaged position of the helping group by implying that they are needed. In contrast, autonomy-oriented help, which offers only partial support to solve a problem, weakens the intergroup hierarchy. It challenges the disadvantaged position of the receiving group by implying that they can solve the problem with sufficient resources, and it challenges the advantaged position of the helping group by implying that their help is needed only because they have more resources than they deserve.

Thus, because justifying the intergroup hierarchy (for the advantaged group) and challenging it (for the disadvantaged group) represent forms of social competition, Nadler (2002), as well as Nadler and Halabi (2006), predict that members of advantaged groups should prefer to offer dependency-oriented help, especially when they perceive the intergroup hierarchy as illegitimate and unstable. They predict, in contrast, that members of disadvantaged groups under these same conditions should prefer to receive no help at all or else, “…try to mitigate the threat of dependency by seeking and accepting only autonomy-oriented assistance and/or by accepting the help only within a long term reciprocal relationship” (Nadler 2002, p. 492). IHSRM also predicts that the preference to offer or reject dependency-oriented help should be strongest for those who are most invested in their group, that is, high-identifiers (Ellemers et al. 2002).

Nadler (2002), as well as Nadler and Halabi (2006), have found support for much of their model. For example, high-identifying members of disadvantaged groups (Israeli Arabs) seek out less help from members of advantaged groups (Israeli Jews)—and reject unsolicited help more—when they perceive the intergroup hierarchy as unstable and illegitimate (Nadler and Halabi 2006). Similarly, Israeli Arabs react more negatively to help from Israeli Jews when they perceive the intergroup hierarchy as illegitimate, and they are less sanguine about the likelihood that help will improve relations between the groups the more they perceive the intergroup hierarchy as unstable (Halabi et al. 2012).

There is less evidence supporting the prediction that disadvantaged group members prefer autonomy-oriented help when they identify strongly with their group and perceive the intergroup hierarchy as unstable and illegitimate. However, Chernyak-Hai et al. (2014) observed that Israeli Arabs who scored low on a measure of system justification—and thus perceived the intergroup hierarchy as illegitimate—were less willing to receive dependency-oriented help from an Israeli Arab and somewhat more willing to receive autonomy-oriented help. Surprisingly, however, they observed the latter effect only when participants viewed the intergroup hierarchy as stable—not when they viewed it as unstable as Nadler (2002) would predict.

Taken together, both theory and evidence suggest that members of disadvantaged groups who view the intergroup hierarchy as less legitimate and, perhaps, less stable prefer autonomy-oriented help from members of high-status groups. To our knowledge, only one study has tested this prediction derived from IHSRM to date (Chernyak-Hai et al. 2014). Although scholars have suggested its application to allied activism by advantaged group members (Droogendyk et al. 2016), none is known to have tested it empirically.

Feminist Women’s Perception of Male Allies

We apply this model to the context of gender, similarly to other researchers before us (e.g., Shnabel et al. 2016; Wakefield et al. 2012). Within the existing gender hierarchy, men occupy an advantaged position, and women occupy a disadvantaged position. Based on the IHSRM, therefore, women who reject this hierarchy and believe it can change may not accept support from men, or they may only accept support if it challenges inequality (i.e. autonomy-oriented support). One group of men who challenge gender inequality is self-identified feminists; women, therefore, may be inclined to accept their support. However, they may view that support less positively if the manner in which feminist men give it reproduces women’s disadvantage in society-at-large. Men who work with women by positioning themselves as “the cavalry” and women as “damsels in distress” (i.e. dependency-oriented support) may be less desirable as allies than are men who attend to women’s priorities and offer the kinds of support for which women ask (i.e. autonomy-oriented support).

Which women are most likely to reject the gender hierarchy and believe that it can change? We argue that women who identify as feminists share these beliefs. Over the past 3 decades, psychologists have conceptualized feminist identity in a number of ways. Some consider it a set of beliefs regarding women and their position in society (e.g., Bargad and Hyde 1991; Downing and Roush 1985; Fischer et al. 2000; Hyde 2002; Rickard 1989). Others consider it the willingness to label oneself a feminist, publicly or privately (e.g., Leaper and Arias 2011; Nelson et al. 2008). Still others draw on both approaches; they categorize feminists as those who adopt the label and endorse gender equity (e.g., Yoder et al. 2012, 2011; Zucker 2004).

Comparing these approaches, it is increasingly clear that women who adopt the feminist label differ from those who do not—beyond whatever attitudes the latter may hold (Kelly and Breinlinger 1995; Liss and Erchull 2010; Nelson et al. 2008; Yoder et al. 2011; Zucker 2004; Zucker and Bay-Cheng 2010). For example, women who identify as feminists are more willing to engage in collective action to challenge gender inequality (Nelson et al. 2008; Zucker 2004; Zucker and Bay-Cheng 2010), and feminist identification is a much stronger predictor of collective action than identifying with women alone (Kelly and Breinlinger 1995; Swank and Fahs 2017; van Breen et al. 2017; Yoder et al. 2011). Further, women who identify as feminists reject sexist beliefs to a greater extent than do non-labelers (Zucker and Bay-Cheng 2010) and are more likely to believe that the gender hierarchy is unfair (Liss and Erchull 2010).

In sum, seen from the perspective of the IHRSRM, women who identify as feminists reject gender inequality and believe it can change (i.e., see it as illegitimate and unstable). Members of disadvantaged groups who hold these beliefs tend to prefer autonomy- to dependency-oriented help when working together with members of advantaged groups. Thus, we predict that women who identify with feminists will perceive a man who identifies himself as a feminist to be a better ally if he offers autonomy-oriented support as opposed to dependency-oriented support.

The Current Studies

We tested these ideas in two studies. In Study 1, we recruited U.S. undergraduate women who self-identified as feminists to read a brief description of a feminist man who offered either autonomy- or dependency-oriented support. We then asked women to rate whether they thought the man would be a good ally. We predicted that feminist women would rate the feminist man a better ally if he offered autonomy-oriented support as opposed to dependency-oriented support (Hypothesis 1).

In Study 2, we recruited an online sample of U.S. women and measured their feminist identification. We asked them to complete a similar protocol as in Study 1. Our goal was to establish a relationship between women’s feminist identification and ally perceptions in the context of autonomy-oriented support. We predicted that women who identified more strongly as feminists would rate the feminist man as a better ally when he offered autonomy-oriented support than women who identified less strongly as feminists. In contrast, we predicted that feminist identification would not be associated with women’s ratings of a feminist man who offered dependency-oriented support.

These studies contribute to the literature on intergroup helping as well as the research on feminist identification. With respect to IHSRM, the studies test an understudied prediction of the model: that disadvantaged group members who see the intergroup hierarchy as illegitimate (and perhaps unstable) prefer autonomy-oriented help from advantaged group members. With respect to the literature on feminist identification, the studies show how feminist women fight for gender equality within the feminist movement as well as in the broader U.S. society.

Study 1

Method

Participants

We calculated the necessary sample size for our study using G*Power (Faul et al. 2007). Based on a power level of .8, a medium effect size of .5, and an alpha level of .05, we estimated that we would need 102 participants for a one-tailed t-test. We recruited 102 self-identified feminists from feminist student organizations and the Psychology Department’s participant pool at a comprehensive college in the northeastern United States. Students were invited to participate in an online study run via Qualtrics (2017). The study was titled “Women’s Social Attitudes” and only students who considered themselves feminists could take part in the study. Based on a screening question embedded in the demographic items at the end of the survey, six participants who did not identify as feminists (“Do you identify as a feminist?” [Yes/No]) were removed from the analysis. The effective sample consisted of 96 self-identified feminist women (Mage = 21.77, SD = 2.51, range = 18–26). Fully 77 (80%) women identified as White, 11 (11%) as Asian or Pacific Islander, 5 (5%) as Hispanic or Latina, and 2 (2%) as Black or African-American. One participant declined to identify her race. In terms of sexual identity, 67 (70%) participants identified as straight or heterosexual, 19 (20%) as bisexual, 8 (8%) as lesbian and two (2%) as some other category.

Procedure

An undergraduate research assistant emailed interested students a link to participate in the study from a lab email address, the name of which is an acronym (SCCI: Social Change and Collective Identity lab). After completing an online informed consent form, the program directed women to the first page of instructions. The instructions read: “Please look at the image and then read the following description of a male college student, Josh, carefully. You will be asked to answer questions about him afterwards.” The program then randomly assigned women to one of two descriptions of “Josh.” Both descriptions were accompanied by a photograph of a moderately attractive White man in his early 20s and standing in front of a non-descript building. We generated the text in a pilot study by giving a separate sample of 20 undergraduate feminists the definitions of autonomy- and dependency-oriented support and asking them to describe a male feminist ally who offered each type of help. In both conditions, the opening text read:

Josh is a 20 year old college student who enjoys reading, listening to music, playing soccer, and going skiing. He spends his free time hanging out with friends and family. Josh considers himself a feminist and wants to help reduce gender inequality in his personal and professional life.

In the dependency-oriented support condition, the text continued: “He asserts his knowledge to women in the movement. He believes his role in the movement is to lead women towards their goals and protect them.” In the autonomy-oriented support condition, the text continued: “Josh listens empathetically to women in the movement. He believes his role is to support women in the fight for gender equality and provide resources where they are needed.” After reading the descriptions, the program directed women to a set of items in which they evaluated Josh as an ally. Following these items, women reported their demographic information, as well as whether they identified as a feminist.

Good Ally Measure

We generated seven items to assess whether women perceived Josh as a good ally. For each item, women responded on a scale from 1 (Strongly Disagree) to 7 (Strongly Agree), with the order of the items randomized across participants. The items were: “Josh seems like the ideal ally for women fighting for gender equity”; “Josh seems like exactly the right kind of man to fight for gender equity”; “The support Josh is willing to give is the kind that the movement for gender equity needs most”; “The kind of action Josh is willing to take is the most helpful for the movement for gender equity”; “Josh is an excellent example of how a man should support gender equity”; “How Josh supports gender equity is close to ideal”; and “There is no better way for a man to support gender equity than the way Josh plans to do so.” We submitted the items to Principal Axis Factoring with Promax Rotation and found that they loaded on a single factor accounting for 74.08% of the variance (Eigenvalue = 5.19; Factor loadings from .61 to .77). We averaged the seven items to create a single scale, which served as our dependent variable (M = 4.79, SD = 1.28, Cronbach’s α = .94).

Results

In order to test the hypothesis that women who identified as feminist would evaluate a feminist man who offered autonomy-oriented support as a better ally than a feminist man who offered dependency-oriented support, we conducted an independent samples t-test. As predicted, feminist women rated the feminist man as a better ally in the autonomy-oriented condition (M = 5.21, SD = .97) than in the dependency-oriented condition (M = 4.40, SD = 1.41), t(94) = 3.26, p = .002, d = .67. Our hypothesis was thus fully supported.

Discussion

The results of Study 1 supported our first hypothesis. U. S. undergraduate women who self-identified as feminists rated a feminist man as a better ally if he offered autonomy-oriented support as opposed to dependency-oriented support. We recognized an important limitation, however. In Study 1 we selected participants who identified as feminists. A stronger test of our claims would examine whether women who report higher feminist identification evaluate a feminist man who offers autonomy-oriented support to a greater degree than women who report lower feminist identification.

Study 2

In Study 2, therefore, we recruited an online sample of U.S. women and measured their level of feminist identification. We predicted that women who identified more strongly as feminists would rate the feminist man as a better ally when he offered autonomy-oriented support than women who identified less strongly as feminists. In contrast, we predicted that feminist identification would not be associated with women’s ratings of a feminist man who offered dependency-oriented support.

Study 2 also addresses an alternative explanation for our findings in Study 1. Women in Study 1 simply may have liked the man who offered autonomy-oriented support more than the man who offered dependency-oriented support regardless of his status as an ally. If women who identified themselves more strongly as feminists rated the autonomy-oriented man as a better ally than women who identified themselves less strongly as feminists—but there is no such relationship in the dependency-oriented condition—the results of Study 1 are unlikely to be due to liking alone.

Method

Participants

We calculated the necessary sample size for comparing the difference in slopes between two groups in linear regression using G*Power (Faul et al. 2007). Based on a power level of .8, an estimated slope difference of .25 (the smallest difference we deemed meaningful), and an alpha level of .05, we estimated that we would need 256 participants for Study 2. Oversampling slightly to account for potential attrition, we recruited 270 women (Mage = 24.73, SD = 3.22, range = 18–51) from Amazon’s MTurk for an online study titled “Women’s Opinions and Attitudes,” run via Qualtrics (2017). Online samples recruited from Amazon’s Mechanical Turk, like the sample in Study 2, are somewhat more diverse than college student samples like our sample in Study 1 (see Paolacci and Chandler 2014, for a review of the characteristics of participants recruited from MTurk). They also sample a much larger population (~7300) than is available in our department participant pool (Stewart et al. 2015) without sacrificing data quality (Paolacci and Chandler 2014). Fully 186 (69%) women identified their race as White, 25 (9.3%) as Asian or Pacific Islander, 23 (8.5%) as Black or African American, 20 (7.4%) as multiracial, 12 (4.4%) as Hispanic or Latina, and 4 (1.5%) as Indigenous American or Alaska Native. In terms of sexual identity, 215 (79.6%) participants identified as straight or heterosexual, 41 (15.2%) as bisexual, 6 (2.2%) as asexual, 4 (1.5%) as lesbian or gay, and 3 (1.1%) as pansexual. One woman declined to report her sexual identity.

Procedure

The procedure followed that of Study 1. In addition to the Good Ally dependent measure, women also completed a measure of feminist identification. We randomized the order of presentation of this measure so that approximately half completed it before reading about the ally Josh and approximately half completed it following the dependent measure (but before the demographic measures).

Measures

We averaged the same seven items to assess whether women perceived Josh as a good ally as in Study 1 (M = 4.72, SD = 1.55, Cronbach’s α = .96). We also assessed feminist identification on a 7-point scale from 1 (Strongly Disagree) to 7 (Strongly Agree) using eight items. We adapted one item from Postmes et al. (2013) single-item measure of social identification: “I identify with feminists.” We adapted remaining items from the solidarity (“I feel a bond with feminists”; “I feel solidarity with feminists”; “I feel committed to feminism.”), centrality (“Being a feminist is an important part of how I see myself”; “Being a feminist is an important part of my identity”), and individual self-stereotyping (“I am similar to the average feminist”; “I have a lot in common with the average feminist”) of Leach et al.’s (2008) Multicomponent model of In-group Identification. Women who agree with these items feel more strongly connected with feminists, similar to feminists, and think that being a feminist is an important part of who they are. In contrast, women who disagree with these items more strongly do not feel connected with or similar to feminists, and they do not consider being a feminist an important part of how they see themselves. We submitted the items to Principal Axis Factoring with Promax Rotation and found that they loaded on a single factor accounting for 85.02% of the variance (Eigenvalue = 6.80; Factor loadings from .78 to .87). We averaged the eight items to create a single scale, which served as our dependent variable (M = 4.35, SD = 1.80, Cronbach’s α = .98).

These measures have been shown to be valid and reliable across a wide number of different social categories and contexts (Leach et al. 2008; Postmes et al. 2013). Further, van Breen et al. (2017) have used a similar measure of feminist identification to predict critical attitudes toward gender stereotypes, attitudes about women’s position in the gender hierarchy, and support for activism (in combination with gender identification). Others have used the solidarity subscale of Leach et al.’s (2008) measure to assess identification with feminists among women and men (Subašić et al. 2018) or men alone (Wiley et al. 2013).

Results

We predicted that, in the context of autonomy-oriented support, women who identified more strongly as feminists would rate the feminist man as a better ally than women who identified less strongly as feminists. In contrast, we predicted that feminist identification would not be associated with ratings of the feminist man in the context of dependency-oriented support. In order to test this hypothesis, we conducted a moderation analysis in regression by using Hayes’ (2017) PROCESS macro for SPSS. First, we regressed women’s perceptions of the feminist man as a good ally on feminist identification, ally helping orientation, and their interaction term. As expected, we observed a significant interaction between feminist identification and ally helping orientation (B = .28, SE = .10, t = 2.88, p = .004, ΔR2 = .03), indicating that the relationship between feminist identification and ratings of the feminist man as a “good” ally varied by ally helping orientation. With the interaction term in the model, neither the main effect of feminist identification (B = .12, SE = .07, t = 1.82, p = .070) nor the main effect of ally helping orientation was significant (B = −.68, SE = .46, t = −1.48, p = .140.

In order to further, understand this statistically significant interaction, we further explored the relationship between the strength of women’s feminist identification (the independent variable) and their perceptions of the male ally (the dependent variables) under the two helping conditions (autonomy- and dependency-oriented support; the moderating variable). We present a plot of these slopes in Fig. 1. As predicted, in the autonomy-oriented support condition, the more strongly a woman identified as feminist, the more positive her rating of the male ally was (B = .40, SE = .07, t = 5.59, p < .001). In contrast, feminist identification was not associated with ratings of the feminist man in the dependency-oriented support condition (B = .17, SE = .07, t = 1.82, p = .070).

Fig. 1
figure 1

Interaction of reports of women’s reports of feminist man as a good ally by his helping orientation and by women’s self-reported feminist identification in study 2. Lower and higher feminist identification represent - 1 SD and + 1 SD from the mean, respectively

Full size image

Discussion

Thus, our hypothesis was fully supported. The more strongly a woman identified as a feminist, the more favorably she rated the male ally who offered autonomy-oriented support. Feminist identification was not significantly related to the evaluation of the feminist man who offered dependency-oriented support.

General Discussion

The present studies represent a shift in emphasis in research on allied activism from whether men will support women as they challenge gender inequality (Iyer and Ryan 2009; Subašić et al. 2018; Wiley et al. 2013) to what style of support feminist women prefer when working with feminist men. We found that women in our studies who perceived that the gender-based hierarchy is illegitimate and who are most motivated to change it—that is, women who identify more strongly as feminists—preferred the style of support that counters men’s privilege and women’s disadvantage. Specifically, women who identified more strongly as feminists reported that men are better allies when they take a backseat when addressing gender inequality (i.e., offered autonomy-oriented support) instead of the driver’s seat (i.e. offered dependency-oriented support). Based on the Intergroup Helping as Status Relations Model (Nadler 2002), we suggest that autonomy-oriented support challenges women’s disadvantage, implying that they are capable of addressing gender inequality on their own, whereas dependency-oriented support affirms it, implying that women need men to achieve their goals. Thus, women who identify more strongly as feminists challenge the gender-based hierarchy not only through their attitudes and activism, but also in how they construe a supportive feminist man. To be a “good” ally in the eyes of the feminist women in the present research, men must offer support in a way that resists the gender hierarchy. Calling oneself a feminist is only a start.

These studies are part of a broader shift in emphasis in social psychological research on intergroup relations from a focus on advantaged group members’ attitudes and behavior to the consequences of those attitudes and behavior for members of disadvantaged groups (Wright and Lubensky 2009). Scholars have shown for decades that positive intergroup contact can improve advantaged group members’ attitudes towards disadvantaged groups (see Pettigrew and Tropp 2006, for a meta-analysis). Only recently, however, have they shown that positive contact can also decrease the willingness of disadvantaged group members to challenge inequality (Becker et al. 2013; Dixon et al. 2010; Reimer et al. 2017; Saguy et al. 2009; Wright and Lubensky 2009). Emphasizing the perspective of the disadvantaged group reveals that apparently positive actions by members of advantaged groups, such as allied activism or positive contact, can be less than ideal (Droogendyk et al. 2016). This realization shifts the conversation from how to engage members of advantaged groups, like men, in social change to how to ensure they do so in ways that are accountable to disadvantaged groups, like women, and consistent with their goals.

These studies have implications for research on intergroup helping. To our knowledge, only one study has examined the IHSRM prediction that members of disadvantaged groups who are motivated to challenge the intergroup hierarchy, such as feminists, will prefer autonomy- to dependency-oriented help (Chernyak-Hai et al. 2014). Most studies have instead examined whether members of disadvantaged groups who challenge the intergroup hierarchy reject help altogether (Halabi et al. 2012; Nadler and Halabi 2006). In this sense, the present studies support an important, but understudied, prediction of the model.

Further, Nadler (2002, p. 492) has suggested that disadvantaged group members will prefer autonomy- to dependency-oriented help (as opposed to rejecting help altogether) “…only within a long-term reciprocal relationship.” The context in which we conducted the present studies, allied activism, may imply just such a relationship. Women may presume that men who identify with feminists are part of the movement. Therefore, they might not reject dependency-oriented support outright even if they would prefer autonomy-oriented support. Our findings are consistent with this reasoning. Although the stronger-identifying feminist women in our studies reported that feminist men were better allies when they offered autonomy-oriented support, they did not report that feminist men were bad allies in the dependency-oriented condition; in fact, their ratings in both studies were near the mid-point of the scale. It seems possible that feminist men who offer dependency-oriented support present something of a mixed message. On one hand, they identify with a movement that purports to challenge the gender hierarchy. On the other, they offer support in a way that affirms that hierarchy. In contrast, feminist men who offer autonomy-oriented support present no such contradiction. Their identification as feminists and the support they offer both challenge the gender hierarchy.

These studies also have implications for psychological research on feminist identification. There is robust evidence that women who identify more strongly as feminists challenge the gender hierarchy in broader society through their attitudes and activism (Kelly and Breinlinger 1995; Liss and Erchull 2010; Nelson et al. 2008; Yoder et al. 2011; Zucker 2004; Zucker and Bay-Cheng 2010). In the present studies, we observed that the women who identified more strongly as feminists also sought gender equality within the feminist movement. They saw some male allies as good, so long as the form of their support cut against the grain of the existing gender hierarchy. Strong-identifying feminist women were less sanguine about other allies—not because they rejected allied support altogether, but perhaps because some forms of support present a mixed message.

Finally, our studies have implications for feminist scholarship outside of psychology, which has examined whether men can and should be feminists and, if so, how they can contribute to the fight against sexism. Feminist separatists and some lesbian feminists have argued that working together with men to achieve equality under a patriarchal hierarchy leaves that hierarchy intact and that women should instead work on their own (e.g., Daly 1978; Dworkin 1987). Feminists working within these frameworks are unlikely to see men as allies, regardless of the forms of support men provide.

Other feminists have argued that men can and should be feminists (e.g., Flood 2015; Tarrant 2009). For example, intersectional feminists criticize the separatist perspective on the grounds that it alienates working-class women and Women of Color, who experience solidarity with the men with whom they share racial or economic disadvantage (e.g., hooks 1984). They argue that a focus on women alone divorces the struggle against sexism from struggles for racial and economic justice and obscures differences in power and privilege among women. Critical men’s studies scholars provide a different rationale for men’s role in feminism. They argue that men are the primary perpetrators of sexism and, as such, occupy a privileged position in changing their own attitudes and behaviors, as well as those of other men (e.g., Katz 1995; Kaufman 2001; Kimmel and Mossmiller 1992). At the same time, they recognize that many men—particularly poor and working class men, gay men, Men of Color, and trans men—are hurt by a gender-based hierarchy that either restricts their roles or positions them as marginal. For feminists working in intersectional and critical men’s frameworks, feminism benefits men as well as women and contributes to the fight against multiple, intersecting forms of oppression.

Within these frameworks, the question is less whether men can be allies and more how they can be allies. Feminist scholars have identified a number of characteristics of a “good” ally. These include taking responsibility for confronting sexism when it is apparent, adopting an intersectional perspective, and challenging the role of hegemonic masculinity in one’s own life and in the lives of other men (Flood 2015; hooks 1984; Katz 1995; Kaufman 2001; Kimmel and Mossmiller 1992; Tarrant 2009). However, there is also concern that in a society that privileges men’s voices and perspectives, feminist men may, perhaps inadvertently, reinforce gender inequality within the movement. Even well-intentioned men who have spent years fighting sexism experience what Peretz (2018, p. 2) has labeled the “pedestal effect,” for example, by which men receive disproportionate recognition for their work. This effect can escalate men into positions of influence and leadership and has the potential to undermine women’s voices and power (Messner et al. 2015). Our studies offer preliminary evidence that women who identify strongly as feminists view men who de-center themselves and remain accountable to the women with whom they work as better allies than those who are more willing to accept (or who seek out) an elevated position.

Limitations and Future Research Directions

Our studies have several important limitations. Our samples include mostly heterosexual, White, young women, and the man in our experimental materials is young, White, cisgendered, and (presumably) able-bodied. Existing studies show that the meaning of feminist identification differs among different age cohorts (Duncan 2010), sexual identities (DeBlaere et al. 2017; Szymanski 2004), and racial and ethnic groups (Robnett and Anderson 2017; Robnett et al. 2012). Perhaps more importantly, feminist perspectives recognize that women experience gendered oppression differently depending on their unique positions within the matrix of privilege and oppression (Crenshaw 1991; Collins 2002; Johnson 2006). They also have different relationships to the men in their lives with whom they might share one or more positions of privilege or disadvantage. These factors likely influence how feminist women perceive support from feminist men and represent an exciting avenue for future research. One intriguing possibility is that different groups of women and men may leverage their multiple identity positions to find “strategic intersectionalities” around which to organize (Belleau 2007, p. 51).

We designed the present experiments with the aim of testing a causal hypothesis derived from IHSRM. As such, we constructed materials and chose samples that would maximize experimental control and be as meaningful as possible to the women who participated in our studies. Our emphasis on internal validity and fidelity has the benefit of allowing us to make causal claims; it also represents some trade-offs. First, both studies employed convenience samples and are, therefore, not representative. It is our hope that future research might examine our hypotheses with an emphasis on external validity. Second, our experimental manipulations included a brief description of a feminist man offering either autonomy- or dependency-oriented support. Future studies, however, can extend the ecological validity of these findings by examining real-world interactions between feminist women and men or by examining the many other ways feminist men may provide support when working with feminist women.

We did not include a control condition in the present study. The pattern of results in Study 2, however, are strongly suggestive that the baseline rating of a male ally is neutral. Women rated the feminist man near the mid-point of the scale in three of the four conditions. Only highly identified feminist women in the autonomy-oriented condition rated the feminist man as a good ally. These results suggest that autonomy-oriented support increases feminist women’s ratings of a male ally. Dependency-oriented support may be what is expected.

Future studies might also examine the factors that encourage men not only to identify as feminists, but also to offer autonomy-oriented support. Some men may identify with feminists as a political group and may be oriented to challenge the gender hierarchy. They may also identify with their gender, however, and place value in the traits and roles that others find to be typical for men. (See van Breen et al. 2017 for a similar contrast between feminist identity and gender identity among women.) Such men may be motivated to take allied action, but also may be motivated to do so in a dependency-oriented manner that is gender-typical. In order to offer allied support, men may need to not only move toward feminism, but also move away from normative male traits and roles that reinforce the gender hierarchy.

Practice Implications

Keeping in mind these limitations, we venture a number of practical implications of our research. When working together with feminist women, men should take care to listen and learn. They should work so that their voices do not drown out the voices of women in the movement. And, they should ensure that their goals and priorities do not divert resources away from those of feminist women. In short, feminist men should thoughtfully consider how they are accountable to women in the feminist movement (Messner et al. 2015). Male privilege and power do not dissipate when men call themselves feminists or commit themselves to fighting gendered oppression. As more men do the hard work of feminism, they should model in their interactions with women the egalitarian society that they hope to build.

Conclusion

In the present studies we have shown that women perceive feminist men as better allies if they take a backseat in addressing gender inequality (i.e., offer autonomy-oriented support) as opposed to if they take the driver’s seat (i.e., offer dependency-oriented support). We have argued, based on the IHSRM (Nadler 2002), that this preference is because autonomy-oriented support challenges the gender hierarchy whereas dependency-oriented support affirms it. Consistent with this point of view, we observed that only women who identify more strongly as feminists—who are most motivated to challenge gender inequality—see feminist men who offer autonomy-oriented support as better allies than those who offer dependency-oriented support. Our studies show that allied activism is not only about whether men help, but how.