Feminist women’s online political participation: empowerment through feminist political attitudes or feminist identity?

ABSTRACT Citizens in modern democracies have a continuously expanding set of tools at their disposal through which they seek to exercise influence on politics, including digital modes of participation. However, the usage of these tools is still gendered to the disadvantage of women. Feminist attitudes have been shown to have a positive impact on women’s political participation, yet this effect is deeply interwoven with the empowering effect of a feminist identity. Based on an online survey of more than 300 German female Internet users self-labeling as feminists, we develop a comprehensive measure of a feminist identity and analyze the interplay of three distinct sets of feminist attitudes and a feminist identity on online political participation. To gain a fine-grained understanding of the impact of feminist cognitions on online political participation, we differentiate general political online behaviors from those geared toward women’s rights and feminist objectives. We find a feminist identity to be a strong predictor of both types of online political participation, with a stronger effect on feminist online participation. Our findings provide important insights into the empowering role of a feminist identity on women’s political behavior on the Internet.


Introduction
Gender differences with regard to political behavior have long been subject of public and academic discourse.While the Internet has bridged some sociodemographic participation gaps, women 1 still significantly lag behind men when it comes to political actions online (Abendschön & García-Albacete;Bode, 2017;Theocharis & van Deth, 2018).And although this gap does not manifest equally across all types of political behavior (Bode, 2017;Coffé & Bolzendahl, 2010;Heger, Leißner, Emmer, & Strippel, 2022;Lilleker, Koc-Michalska, & Bimber, 2021) nor across womanhood as a whole, the gender gap in political participation has important implications for democracy: The exercise of power in democratic political systems is legitimatized as an expression of the people's collective self-determination and democratic political systems gain legitimacy through the political participation of the governed (Scharpf, 1999).It is thus of critical importance to understand both facilitators of and obstacles to women's political participation.
One element that has been shown to facilitate women's political participation are feminist cognitions (Heger & Hoffmann, 2021;Schuster, 2013;Swank & Fahs, 2017).Two such cognitions widely discussed in the literature are feminist attitudes and a feminist identity.Previous research into the respective roles of these cognitions struggled to conceptually and empirically disentangle their effects (Rhodebeck, 1996).As a result, while there is some consensus that feminist cognitions can facilitate women's political participation, it is not yet clear if this is primarily due to feminist attitudes alone or the adoption of a feminist identity (Eisele & Stake, 2008).
This study sets out to disentangle the effects of feminist attitudes and identity on political participation in the online sphere.To that end, we develop a scale measuring a feminist identity building on established scales such as the Feminist Identity Development Scale (FIDS) by Bargad and Hyde (1991) as well as Social Identity Theory and concurrent findings (Ashmore, Deaux, & McLaughlin-Volpe, 2004;Tajfel, 1978).The novelty of this approach lies in the integration of both collectivist and individualist dimensions into one comprehensive measure of a feminist identity, hence reconciling competing scholarly perspectives.Second, we build on Heger and Hoffmann (2021) by examining the role of distinct feminist attitudes based on three paradigms or 'waves' of feminist thought.
Disentangling the respective influences of feminist attitudes and identity on online political participation is important to derive insights and recommendations for the political empowerment of women: While feminist attitudes toward specific topics are known to mobilize political participation, Social Identity Theory and the FIDS explicate how the development of a feminist identity engenders applying those attitudes to one's own life and role in society, which we expect to yield a distinct urge for action.
Finally, to derive a fine-grained understanding of the effects of feminist attitudes and identity on online participation, we distinguish general online political participation and feminist online political participation.We expect that feminist cognitions exert disparate influences on online engagement for women's causes specifically and political online engagement in general.This distinction is important as focusing solely on general online political participation may underestimate the empowering role of feminist cognitions -a feminist identity, in particular -on women's political engagement.Our analysis reveals that a feminist identity empowers political participation beyond issues specifically attached to this social identity.
The present study is therefore dedicated to answering the following research question: What is the relationship between feminist attitudes and a feminist identity in empowering both women's feminist and general online political participation?

Women's online political participation
The political participation concept has evolved with the transition of a multitude of political activities to the digital sphere (Gibson & Cantijoch, 2013;Ruess, Hoffmann, Boulianne, & Heger, 2021;Theocharis & van Deth, 2018).Initially, the opportunities provided by online platforms spawned hopes for more inclusive participation.Yet, to date, the Internet has not bridged the gender gap in political participation (Norris, 2002;Theocharis & van Deth, 2018).Some studies find the gender gap in political participation to be shrinking, at least with regard to specialized online platforms or topics (Bode, 2017;Lilleker et al., 2021;Oser, Hooghe, & Marien, 2013) or in very specific political and cultural contexts (Feezell, 2016;Oser et al., 2013).In the German context, where the present study was conducted, gender inequality patterns have been found to change to women's advantage with regard to some types of political participation, but to persist in the online realm (Heger & Hoffmann, 2021;Heger et al., 2022).
These findings advise a differentiated approach with regard to the measurement of political activities (Theocharis, 2015).An important distinction to make is that between general and feminist political engagement, as motivations can be expected to depend on specific goals and targets (Duncan, 1999;Simon et al., 1998).Szymanski (2004) argues that "[. ..] feminist [. ..] activism typically focusses on deconstructing oppression."Accordingly, previous research has found that a feminist identity relates to women's feminist or gender-specific political engagement.However, this relationship has primarily been explored in the offline realm (Duncan, 1999(Duncan, , 2010;;Szymanski, 2004).Building on this finding as well as Nancy Fraser's (1990) theory of counter-publics, new research discusses the possibility of (some) feminists avoiding mainstream online participation platforms such as comments sections on news sites or public digital spaces like Twitter to elude harassment or to organize in closed feminist groups (Mendes, Ringrose, & Keller, 2018). 2  Consequently, to systematically assess the role of feminist cognitions in women's online political participation, specifically feminist forms of online activism have to be taken into consideration alongside more widely adopted political online behaviors.We therefore define general online political participation as the usage of online platforms by private citizens to create, edit, transmit, and publish content, or to conduct transactions with the intent to influence government (cf., Ruess et al., 2021).In the case of feminist online political participation, these behaviors are focused specifically on feminist advocacy and women's rights.

Distinguishing feminist attitudes and a feminist identity
Feminism -just like any type of political orientation or ideology -provides a perspective through which politics can be perceived and evaluated.It defines "women's stake in the world outside the traditional, private homeplace" (Kay, 1985, p. 476).Core to the feminist analysis is the assumption that women are assigned an inferior societal status which arose from and is maintained by patriarchal structures that impose specific gender norms and expectations on social roles (Butler, 1991;De Beauvoir, 1949).As a theoretical paradigm, it explains how these structures and mechanisms come into existence, how they can be overcome and it entails mobilizing effects on women in a variety of regards, political participation among them (De Beauvoir, 1949;Fulenwider, 1981;Swank & Fahs, 2017).
However, studies do not agree on whether this empowerment is based on feminist attitudes or a feminist identity: While feminist attitudes are understood as attitudes toward the social role of women, consciousness of sex bias and beliefs in feminist goals, such as gender equality in social structures and practices (Brush et al., 1978;Eisele & Stake, 2008), a feminist identity evolves around a person's self-understanding as a feminist (Szymanski, 2004;Weis, Redford, Zucker, & Ratliff, 2018), based on their "personal beliefs about the definition of one's own role" (Brush et al., 1987, p. 875).Feminist attitudes can therefore be understood as attitudes toward gender as a political object while a feminist identity refers to the self-conception as a political subject of a specific gender.
Studies investigating the relationship of feminist attitudes and a feminist identity found the two phenomena to be empirically distinct (Rhodebeck, 1996).Bargad and Hyde developed the Feminist Identity Development Scale (FIDS) that distinguishes five stages in the development of a feminist identity ranging from "passive acceptance," in which there is denial or unawareness of sexism, to "active commitment," where social changes and "[. ..] the translation of the newly developed consolidated identity into meaningful and effective action" (Downing & Roush, 1984, p. 702) is located.However, the FIDS does not measure the self-labeling as a feminist and avoids the term altogether resulting in disagreement on whether it actually measures identity at all (Eisele & Stake, 2008;Siegel & Calogero, 2021;Zucker & Bay-Cheng, 2010).Given these theoretical arguments for the distinction between feminist attitudes and a feminist identity, we will analyze how both interrelate in informing women's online political participation -both general and feminist participation.

Feminist attitudes, feminist identity and women's online political participation
Various analyses have pointed out that feminist attitudes are not a homogenous concept, even though they may share core ideas (Henley, Meng, O'Brien, McCarthy, & Sockloskie, 1998;Siegel & Calogero, 2021).In fact, a number of distinct feminist attitudes can be distinguished, and different typologies of feminist thought have been proposed, among them the differentiation of three different 'waves' or paradigms of feminism (Baumgardner & Richards, 2000;Evans & Chamberlain, 2015;Gamble, 2006;Zhang & Rios, 2021).While the term 'wave' may denote a dimension of temporality, this is strictly limited to the chronology in which these schools of thought emerged and "had a mass base" (Mann & Huffman, 2005, p. 58).Scholars have highlighted the simultaneity and non-linearity in which these waves occurred and manifest today (Ewig & Ferree, 2013;Knappe & Lang, 2014).
The first wave, which nowadays translates into liberal feminism or feminism of equality (Donovan, 1992;Lindsey, 2014) is centered on the claim of equality of the sexes as both women and men are assumed to be equally capable of reason and therefore entitled to the same rights.Essentialist liberal feminist thinking, sometimes critically referred to as 'neo-liberalist feminism' or even anti-or postfeminism, locates the roots of (political) inequalities between women and men in the fact that women have not yet managed to prove themselves as equals in a male-dominated world (Dabrowski, 2021;Donovan, 1992;Fitz et al., 2012).The resultant strategy rests on women's appropriation of typically male-associated traits and behaviors (Plumwood, 1990).However, as the goals of liberal feminism lie in an individual gain of power rather than in all-encompassing political or societal transformation (Siegel & Calogero, 2021;Zhang & Rios, 2021), its primary arena might be the workplace or even the private sphere, rather than politics (Dabrowski, 2021;Lindsey, 2014).
The second wave of feminism, today materializing as standpoint or radical feminism or feminism of difference, postulates women and men to be essentially different and traces gender-based inequalities back to patriarchal structures based on the idea of manhood as the norm and womanhood as "the second sex" (De Beauvoir, 1949).It promotes opposition and resistance to the patriarchy as the main strategy to overcome the oppression of women (Henley et al., 1998)

which includes political participation
Finally, third-wave feminism is an umbrella term for a conglomerate of post-structuralist, post-colonialist, intersectional and queer feminist viewpoints and is "the most disparate of the waves" (Evans & Chamberlain, 2015, p. 399;McRobbie, 2009).It criticizes existing (patriarchal) power structures from differing theoretical angles albeit commonly rejecting both the dichotomization of gender as well as the homogenization of womanhood that are intrinsic to first-and secondwave feminism (Mann & Huffman, 2005;Snyder, 2008).The third wave also calls for activism and resistance against existing structures of genderbased discrimination with both a deconstructivist perspective on gender as a social and cultural category and a sensitivity toward intersecting discrimination based on ethnicity, sexual orientation, social class or other social or cultural characteristics (Butler, 1991;Cole, 2009;Mann & Huffman, 2005).Even though these three paradigms can be distinguished theoretically and empirically, they share historical and conceptual common ground (Mann & Huffman, 2005): In the US, the origins of feminism were deeply intertwined with the antislavery movement (Brown, 1978).Women of color continuously engaged with the women's movement, highlighting how intersectional elements were present in the first and second wave, until the US black and white women's movements split over the third-wave critique of the "essentialist woman" that primarily white feminisms had established (Breines, 2006;Mann & Huffman, 2005;Spelman, 1988;Springer, 2002;Vogel, 1991).In Germany, issues like diversity and intersectionality found their way into the feminist movement during the 1990s, triggered by the German reunification and the process of European integration (Ferree, 2012).Cyberfeminism and hashtag activism have also gained salience in Germany with the 2013 hashtag #aufschrei ("outcry") that feminists employed to highlight issues such as gendered power relations and everyday sexism.The hashtag received over 50.000 retweets in only a few days, mirroring the success of the global #MeToo movement.Szymanski (2004) as well as Heger and Hoffmann (2021) show that attitudes reflecting all three paradigms of feminism positively correlate with feminist self-labeling thereby confirming numerous studies in their claim that embracing feminist attitudes is related to developing a feminist identity.It is important to note, however, that the study of a feminist identity goes beyond mere feminist self-labeling (Ashmore et al., 2004;Duncan, 2010).The literature distinguishes two levels of a feminist identity: an individual identity, centered around the individual's personal role as a feminist, and the collective character that is expressed by identifying with a group (Brewer & Gardner, 1996;Simon & Klandermans, 2001).While Social Identity Theory captures the notion of a collective identity (Tajfel, 1978), Ashmore et al. (2004) emphasize the need to distinguish an individual identity compounded by "characteristics of the self that one believes [. ..] to be unique to the self [and], [. . .][to] set [. ..] one apart from all others" (Ashmore et al., 2004, p. 82).In the present study, we therefore develop a new, comprehensive measure of a feminist identity that explicitly includes both the self-perception as part of a feminist group, and a personal or individual feminist identity that is framed as a personal selfunderstanding as a feminist.

Research model
We expect a feminist identity to empower women to participate politically on the Internet.It is assumed to empower both general online political participation (GOPP) and feminist online political participation (FOPP).However, building on prior research on the association of personal cognitions and issue-specific behavior (Vraga, Anderson, Kotcher, & Maibach, 2015), we also expect a feminist identity to yield more empowering potential for feminist online political participation than participation of less issue-specific character (Zucker, 2004).Women's issues serve as a focal point for both the development of a feminist identity as well as their advocacy in terms of feminist online political participation (Duncan, 1999(Duncan, , 2010;;Szymanski, 2004).Hypothesis 1a: A feminist identity has a significant positive effect on both GOPP and FOPP.

Hypothesis 1b:
The effect of a feminist identity on FOPP is stronger than its effect on GOPP.
As first-wave feminism is focused on adaption to existing structures and as it is known to have a negative direct effect on GOPP (Heger & Hoffmann, 2021), we also expect it to exert a negative direct effect on FOPP.Yet, since this type of liberal feminism focuses on individual strategies other than political participation to advocate women's interests, we do not expect explicitly feminist political participation on the Internet to be its primary instrument (Donovan, 1992;Zhang & Rios, 2021).
Prior research has shown second-wave feminism's positive effect on women's GOPP, tracing this finding back to the oppositional impetus inherent to radical feminism which builds on the idea that women are fundamentally different than men because of socialization (De Beauvoir, 1949;Heger & Hoffmann, 2021).It is therefore plausible for this radical feminist paradigm of difference to have a stronger positive effect on FOPP than GOPP.Given post-structural, post-colonial and queer feminists' preference for more group-specific, directional activities, we also expect a stronger effect of this feminist attitude on FOPP than on GOPP (cf., Fraser, 1990;Henley et al., 1998;Snyder, 2008).

Hypothesis 2a: Liberal feminist attitudes of equality (first wave) exert a negative direct effect on FOPP and GOPP while radical feminist attitudes of difference (second wave) and post-structural feminist attitudes (third wave) exert a positive direct effect on FOPP and GOPP.
Hypothesis 2b: The direct effects of feminist attitudes on FOPP are stronger than their direct effects on GOPP.
Finally, as earlier studies have shown that a feminist identity mediates the effects of more general political attitudes on women's political activism offline (Duncan, 1999), we expect a feminist identity to partially mediate some of the effects of feminist attitudes on online political participation.
In accordance with prior research on the varying relationship of a feminist identity with different feminist attitudes (Szymanski, 2004), we hypothesize liberal feminism of equality to be less interrelated with a feminist identity than the other two paradigms.This is based on the liberal paradigm focusing on every woman's personal advancement rather than wholesale societal change (Szymanski, 2004).We propose radical feminism of difference to be strongly associated with a feminist identity for it is not only strongly influenced by women's collective movements but also a political strategy of opposition to patriarchy.Post-structural feminism is considered an antecedent of a feminist identity, possibly even more so than radical feminism, as it embraces the concepts of diversity and intersectionality.Its political strategy of 'strategic alliances' empowers inter-group collaboration.As a feminist paradigm, post-structural feminism focusses on the solidarity among several groups sharing interests and the importance of collective action.Hypothesis 3: A feminist identity partially mediates the effect of radical (second-wave) and poststructural (third-wave) feminist attitudes on both GOPP and FOPP but not the effect of liberal feminist attitudes.

Research design
We use data gathered in a survey in 2018 among 1078 German Internet users from a non-probability sample.A certified market research institute provided access to the participants who were recruited via e-mail from a panel and received a small monetary compensation.The recruitment procedure included a quota for women 3 (70%) and quotas for age equivalent to the German population.We employed feminist self-labeling as a sampling criterion for this study as we focus on the effects of feminist attitudes and identity, which can only be meaningfully assessed among those with a feminist self-conceptualization in the broadest sense.In the framework of social identity theory, selfcategorization theory has established self-labeling as a necessary condition for the adoption and practice of any social identity (Ashmore et al., 2004;Duncan, 2010;Turner, Hogg, Oakes, Reicher, & Wetherell, 1987).Self-labeling was assessed with at least partial agreement with (i.e.non-rejection of) the statement "I am a feminist."Of all participants, 43,58% at least somewhat agreed with this statement, resulting in a sample size of 329 (see Appendix A).We conduct step-wise linear regression analyses to assess the hypothesized mediation employing SPSS v.27.

Measures
We developed a novel measure of a feminist identity explicitly including both an individual and a collective feminist dimension.Items included "Being a feminist shapes my identity" and "Being part of a community of feminists is important for me," among others (see Table 1) and are based on Bargad and Hyde's (1991) FIDS.However, the items include explicit references to a feminist identity, accommodating the criticism toward the ambiguity of the FIDS items.The first three items refer more strongly to an individual identity, stressing how feminism plays an important role in a woman's life and how it impacts their identity while the last four items emphasize the importance of membership in a collective of feminists.The  On a scale from 1 "I completely disagree" to 5 "I completely agree," how much do you agree with the following statements? .904 Feminism plays an important role in my life.
.843 Being a feminist shapes my identity.
.864 The situation of women in this country has an effect on my situation.
.753 Being member of a feminist community is important to me.
.850 I spend most of my time with friends who share my feminist values.
.801 Spending time with people who share my political orientations is important to me.
.728 As feminists we must demand legislation to support women. .741 scale was tested with regard to its validity and found conclusive, with a Cronbach's Alpha of .904.
Our measures of liberal/first-wave, radical/second-wave and post-structural/third-wave feminist attitudes were derived from Heger and Hoffmann (2021, see Appendix B).
Our measure of GOPP is based on measures applied in previous studies (Gibson & Cantijoch, 2013;Russo & Amnå, 2016;Theocharis & van Deth, 2018) and includes questions about the frequency (ranging from 1 "never" to 6 "several times a day") with which nine political activities are carried out in the Internet, with Cronbach's Alpha of .933(see Appendix D).
Our operationalization of FOPP is novel and partly derived from Szymanski's ( 2004) and Duncan's (1999Duncan's ( , 2010) ) typologies of feminist participation enabling us to tie in with their findings on the effects of a feminist identity on women's political participation (see Table 2).We incorporated their typologies in the context of more recent research on types of online political participation and attempted to reflect our measures of GOPP.All eight items were assessed on a scale of frequencies from 1 "never" to 6 "several times a day."The scale, built as a summative index, was validated through exploratory factor analysis (PCA), Cronbach's Alpha is .951.Appendix D presents some details on the average frequency with which GOPP and FOPP are practiced, confirming previous findings that the average level of political online engagement is low in Germany.
Prior studies have shown how socioeconomic resources, such as education and income, political resources, such as interest and knowledge, age and gendered socialization experiences stratify women's approach to political participation (Anduiza, Gallego, & Cantijoch, 2010;Coffé & Bolzendahl, 2010;Schlozman, Burns, & Verba, 1999).Additionally, some political dispositions like self-efficacy positively impact political engagement (Almond & Verba, 1963).We therefore included a number of control variables which are presented in detail in Appendix E: To measure age, participants were asked for their year of birth.Internet skills were assessed based on Hargittai's (2005) index of participants' familiarity with online-related terms such as "forum" or "download."Time spent online was measured on a 6-point scale ranging from 1 "never" to 6 "more than three hours per day."Diversity of Internet usage was measured as an index of activities such as e-mailing and online shopping.Political knowledge was assessed as an index of correct answers by asking participants to correctly ascribe political positions to four leading German politicians, such as the chancellor or president (based on Delli Carpini & Keeter, 1996, reduced for reasons of parsimony).Political interest was measured by a single question asking for self-assessment.For political orientation, participants were asked to identify their preference for a selection of seven German political parties on a left-to right-wing spectrum.We employed established measures for internal (Hayes & Bean, 1993) and external political self-efficacy (Niemi, Craig, & Mattei, 1991).On a scale from 1 "never" to 6 "several times a day," how often do you practice the following activities? .951 I sign petitions on the Internet that specifically invest in women's rights (via e-mail or social media) .720I share petitions on the Internet that specifically invest in women's rights (via e-mail or social media) .847 I share political postings, articles or videos on the Internet that contain feminist or gender-specific messages (via e-mail or social media) .860 I join feminist groups on the Internet and discuss with other members (via e-mail or social media) .878I contact politicians or public representatives on social media and call for feminist engagement .925I contact politicians or public representatives via e-mail and call for feminist engagement .918I write articles or postings with a feminist or gender-specific message and share them on the Internet (via e-mail or social media) .919 I produce videos about women-or gender-specific issues with a feminist message and share them on the Internet (via e-mail or social media) .908

Results
Table 3 shows how a feminist identity contributes to both the general and feminist online political participation among women self-labeling as feminists (models 2 and 4) -beyond the contribution of feminist attitudes (models 1 and 3).All models control for socio-demographic variables (age, education, income), political and online resources as well as general political attitudes.In fact, a feminist identity is by far the strongest positive predictor of both forms of online participation in our study, while the effect on FOPP is stronger than the effect on GOPP.We therefore find support for H1a and H1b.Liberal feminist attitudes of equality have a significant negative impact on both forms of OPP, yet the difference in effect size is not substantial.Radical feminist attitudes of difference, found to have a positive effect on GOPP before (Heger & Hoffmann, 2021), shows no direct effect in our analysis, which indicates that a feminist identity explains the empowering effect for this group more than their specific radical feminist convictions do.Similarly, post-structural feminist attitudes do not appear to empower feminist women to participate in politics online beyond their feminist identity, regardless of the participation mode.Hence, we find only partial support for H2a as it relates to the negative effect of liberal feminist attitudes.We reject H2b.
In terms of the hypothesized mediation, we reject H3: As radical and post-structural feminist attitudes do not exert a significant direct effect on GOPP or FOPP, we also do not find a mediation through a feminist identity.There is, however, some evidence for a small, unsubstantial partial mediation of the effect of liberal feminist attitudes on GOPP and FOPP through a feminist identity.
Finally, with the inclusion of feminist identity, some further effects can be observed: The effect sizes of other relevant predictors of (feminist) online political participation such as age and internal political self-efficacy are diminished when feminist identity is introduced to the model.On the other hand, the importance of income increases.Political interest plays a more sizable role for general OPP than for feminist OPP in feminist women while interestingly, our measure of political knowledge appears to be hampering both types of participation: The more feminist women know about politics, the less they participate, and more strongly so with regard to feminist online political activities.

Discussion
Building on previous research on feminist cognitions as well as the gender gap in political participation, we a develop a theoretical framework to explain women's political participation on the Internet based on the interplay of feminist attitudes and a feminist identity.Our new, comprehensive measure of feminist identity confirms that both the notion of feminist group membership and the notion of feminist singularity and distinction contribute to women's self-concept as feminists and should explicitly be included in examinations of a feminist identity (Ashmore et al., 2004).We also find a feminist identity to be the more important feminist cognition with regard to both feminist and general online participation as compared to feminist attitudes.In addition, our analysis substantiates the expectation that a feminist identity has a stronger positive effect on feminist online political participation than on general online political participation.
We interpret the finding that a feminist identity outperforms feminist attitudes in the framework of feminist identity building as the FIDS conceptualizes involvement in feminist activism as the last stage of the development of a feminist identity (Bargad & Hyde, 1991;Downing & Roush, 1984).Thereby, while the mere adoption of feminist beliefs can already yield an empowering effect, this mechanism solidifies when those beliefs are converted from general feminist reflections of society to a feminist self-reflection in society, altering the perception of gender-based inequalities into inequalities that affect the self, which, ultimately, culminates in a stronger effect on political activism.The positive association of a strong identification as group member with participation in political activities linked to the group's interests has been shown for other groups as well, such as members of trade unions (Kelly & Kelly, 1994;Veenstra & Haslam, 2000) and gun owners (Lacombe, Howat, & Rothschild, 2019;Schwartz, 2021).This finding, in conjunction with the especially pronounced effect of feminist identity on FOPP, might, to a certain extent, substantiate theorizing on feminist identity politics: Feminists engage in feminist OPP more than in less issue-specific OPP -and they do so because of their feminist identity and not so much because of their feminist standpoints, even though these two feminist cognitions are interrelated.While this effect is stronger for FOPP, it is also significant for general OPP -a finding that is especially relevant in the light of contemporary, critical debates of identity politics.
We find that the motivating effect of a feminist identity is not a restrictive or segregating one as its effect stretches from specifically feminist participation modes to more 'mainstream' ones.Hence, feminist women participate more in politics online because they identify as feminists -but this does not stop them from engaging in online debates, activities and interactions that reach beyond gender-specific or feminist issues.This finding could hint at a socially integrating instead of separating effect of a feminist identity, contrary to what some opponents of identity-driven political participation propose.Future research should investigate the extent to which issue-specific participation and general participation inform each other and whether there is a 'chain of empowerment' at play here: The accumulation of resources in the context of issue-specific participation (Szymanski, 2004) may serve to bolster more general forms of political participation.
Our finding on the negative relationship between liberal feminism of equality and feminist women's online political participation -which remains largely untouched by a feminist identity -signifies that liberal feminists' abstaining from digital modes of participation is a choice based on their feminist standpoints rather than their feminist selfconception.In our sample of Internet users selflabeling as feminists, neither radical nor poststructural feminist attitudes exert a positive influence on online political participation.Interestingly, though, a correlation analysis (Appendix C) reveals that post-structural feminism correlates positively with a feminist identity, which substantiates prior thinking on the relevance of identities for poststructural and intersectional feminist attitudes and politics.Those who apply the third-wave idea of intersectionality to feminism have even been found to be most prone to identify as feminists (Hoskin, Jenson, & Blair, 2017).We conclude that the differential role of feminist self-labeling and a feminist identity deserves further investigation for liberal and radical feminist attitudes, as self-labeling appears to capture most of the empowering impetus of these attitudes.
The present study has a number of limitations: First, due to our focus on feminist female Internet users, our analysis cannot speak to the relationship of feminist self-labeling and the feminist identity.Given our sample size as well as our reliance on a market research institute for sample access, it is likely that a part of the feminist spectrum is not represented in our sample.A different sampling approach would likely be necessary to explore the full breadth of (feminist) women, such as contacting local feminist groups.
In addition, feminism, like any other political paradigm or movement, is subject to overarching cultural, political and social characteristics of the society investigated, as illustrated by our comparison of U.S. to German feminist history, and, thus, not universal in its interpretation and application (Banaszak & Plutzer, 1993).Against this background, we consider our findings most applicable to other Western European societies as feminism in the United States or Eastern Europe developed quite differently (Molyneux, 1996).Moreover, the plurality of typologies that describe the historical and contemporary varieties of feminism points to the complexity of these political theories and cultures which poses a challenge to grasping them empirically: While some measurement instruments assess feminism unidimensionally (see Brodsky, Elmore, & Naffziger, 1976;Fassinger, 1994;Morgan, 1996;Zucker, 2004), others promote an understanding of feminism as liberal, radical, Marxist, cultural, Women of color and lesbian feminism to reflect perspectives that target different issues and audiences (see Henley et al., 1998;Henley, Spalding, & Kosta, 2000;Simoni, Henley, & Christie, 1999).And while the first approach appears rather simplistic and monolithic, the latter tends to not be feasible for much of quantitative research for pragmatic reasons.The 'wave' metaphor that is used in this contribution has been met with criticism for a variety of drawbacks, especially for its exclusive focus on white feminists' achievements and its oversimplification of feminist movements (Gillis & Munford, 2004;Springer, 2002).However, it has conceptual recognition value and empirical rooting (Evans & Chamberlain, 2015;Laughlin et al., 2010;Reger, 2017), especially in the European context (Dean & Aune, 2015), it enables the integration of more specialized attitudes like liberal and radical feminism into a broader framework (Mann & Huffman, 2005) and it is being carried forward by feminist scholars debating the possibility of a fourth wave (Munro, 2013).
In line with the FIDS, our research model builds on the assumption that feminist attitudes constitute a precondition for the development of a feminist identity and therefore precede it.However, we did not find that among our sample, a feminist identity mediates feminist attitudinal effects on OPP.Furthermore, the differing effects of distinct feminist attitudes indicate that researching their relationship with differing stages of feminist identity development, rather than the final product captured in this study, might offer interesting insights into whether and when these two cognitions develop and diverge, and how their varying relationship affects political participation.After all, the linearity implied in the model by Downing and Roush may not be experienced by all women (Fischer et al., 2000;Liss & Erchull, 2010).To fully capture the relationships between feminist attitudes and the adoption of a feminist identity, longitudinal data would be required.Additionally, investigating a moderating rather than a mediating mechanism would be a next step to better understand the relationship of feminist attitudes, a feminist identity and women's political engagement; however, exploring such moderating mechanisms would require larger sample sizes than the present study.As political participation is a multifaceted phenomenon that is structured by a multitude of socio-economic, socio-demographic and attitudinal factors, our model includes a variety of control variables which somewhat stretches the limits of our sample size.However, this study can serve as starting point for future research on the potential of feminist cognitions to close the gender gap in political participation, particularly in the digital sphere.
Some interesting model changes could be observed when introducing feminist identity that go beyond political attitudes, implying a relevance of the concept for other participatory determinants such as age and political self-efficacy.We find an initially puzzling negative relationship between political knowledge and OPP, which may be due to our measure of political knowledge being focused on established political institutions, while online participation (especially feminist OPP) may be geared more toward non-conventional forms of political action and, thus, political knowledge.The measurement of political knowledge remains contested (Bullock & Rader, 2021;Robison, 2015).While we used a reduced variant of the original scale developed by Delli Carpini and Keeter (1996) to measure this control variable in a parsimonious way, given this interesting finding we suggest further investigating the impact different types of political knowledge might exert on the political engagement for a range of social and political groups, especially (feminist) women.Similarly, against the background of increasing hashtag activism, an updated measure of Internet skills would help in explaining differences in online engagement.
The results established in this study are relevant to of both scientific and nonscientific audiences.Not only do they point to the urgent need of distinguishing between different types of political cognitions, especially attitudes, identity building and selfcategorization, when researching their impact on political behavior, they also demonstrate the role of feminist cognitions when it comes to empowering women to engage politically on the Internet.As our contribution shows, a feminist identity is the dominant cognition in doing so, as it relates the issues feminist attitude revolve around directly to one's own life and role in society, thereby urging the individual more strongly to engage.Hence, fostering feminist identity building in (young) women can be considered a tool in overcoming the gender gap in political engagement in the long run.However, the spreading and commercializing of feminist self-labeling, as it has been observed lately, is not the key to success: Identity-building must be undergirded by the adoption of feminist attitudes.Therefore, sensitizing society to gender-based inequalities and topics, and helping young women in relating those inequalities and topics to their own lives, would help with the development of such an identity.Policies such as gender mainstreaming can be considered a step in that direction.
In an (online) political discourse in which women are typically underrepresented, this finding can be the starting point for the development of both political and civil initiatives and campaigns toward equality, focusing on feminist attitudes, and relating them to the social identity of women.Finally, in times where the aforementioned online discourse increasingly spills over to the offline public debate (Abendschön & García-Albacete, 2021), working toward a less gender-biased Internet also implies working toward increased democratic legitimacy.

Notes
1.While we fully acknowledge that the gender-based inequalities discussed in this article do not only affect persons who are women by their biological sex, most empirical studies in the field, including our own, refer to this group.More information on gender in our sample can be found in the methods section.2. Arguably the most prominent recent example of feminist online engagement is the global #MeToo movement, which encompassed activism in both closed and public digital spaces.
3. The procedure with which quota were applied was based on a binary concept of sex and gender.However, we included a more diverse measure in our survey from Döring (2013) based on which individuals who did not consider themselves women were excluded from this analysis.

Table 1 .
Exploratory Factor Analysis: Feminist Identity

Table 2 .
Exploratory Factor Analysis of Feminist Online Political Participation (FOPP)

Table 3 .
Stepwise linear regression analyses of general and feminist OPP in feminist women mediated through a feminist identity