Dimensions of Transnational Feminism: Autonomous Organizing, Multilateralism and Agenda-Setting in Global Civil Society

The importance and impact of feminist mobilization across borders is well documented, but the impact of autonomy as an aspect of such organizing has not been explored in the transnational context. We argue that to understand the impact of transnational feminist mobilization, at least two distinct types of feminist mobilization require further conceptual development and empirical exploration in the transnational context, namely, autonomous as contrasted with multilateral mobilization. We o ﬀ er a conceptual framework for distinguishing and studying these two forms. Further, using a mixed-methods study design, we empirically distinguish domestic and transnational dimensions of feminist activism and illuminate the impact of both multilateral feminist organizing and autonomous feminist organizing in the transnational space. Our analysis reveals that domestic and transnational organizing are distinct but related phenomena. We also ﬁ nd that in online organizing spaces, autonomous feminist campaigns amplify the messaging of geographically dispersed grassroots and individual activists more than multilateral ones. It further suggests that autonomous movements may o ﬀ er more potential for representing marginalized groups of women, though this potential may not always be realized. The paper o ﬀ ers new


F
eminists working inside multilateral institutions from the World Bank to the UN have successfully pushed for initiatives to advance gender equality and women's empowerment. 1 Yet these international organizations present challenging terrain for feminists, with obstacles stemming from bureaucratic processes, the agendas of member governments, and vested economic interests (Chappell and MacKay 2017;Hozic and True 2016;MacKay 2013;Prügl 2017).These constraints may be reflected in the results of feminist activism in these organizations.For example, gender-focused reports by the World Bank rarely criticize specific member governments, preferring instead to offer examples of best practices as incentives (e.g."Appointing a gender-equal cabinet is good for Canada")(World Bank Group 2015a).In contrast, Women Living Under Muslim Law (WLUML), an independent transnational feminist network, regularly criticizes governments, calling for "stopping gender apartheid in Afghanistan" and criticizing the Iranian government for its attack on education for Iranian schoolgirls (WLUML n.d.).The fact that WLUML is autonomous enables it to make such statements without compromising its primary mission or alienating supporters.Indeed, its very purpose is to make such statements.
Distinguishing the different dimensions of transnational feminist organizing helps us understand the impact of transnational feminism as a political phenomenon.A form of social movement organization, transnational feminism is an important dimension of global civil society and democracy (Clark et al. 1998;Moghadam 2005;Smith 2008;Smith et al. 2021).Such movements are important avenues for representation for a wide range of groups and issues in both domestic and global civil society (Clark et al. 1998;Keck and Sikkink 1998;Teivanen and Trommer, 2017;Weldon 2011).In this paper, we argue that the role transnational feminism plays in representing women varies across different organizational forms, contrasting multilateral feminisms (as in the example of feminists organizing inside international organizations like the World Bank) with autonomous feminisms (as in the case of WLUML).
We differentiate these forms of transnational feminism by examining the differences in their organizational arrangements, examining whether they are free-standing or intra-institutional.We theorize the concept of movement autonomy in the transnational context and analyze the participation of organizations from the Global South in a series of online campaigns. 2By way of contrast, we use the concept of multilateral feminism as a mode of nonautonomous feminist organizing in international spaces that is akin to the concept of state feminism employed in comparative feminist analysis (McBride and Mazur 2010;Smith et al. 2021).We focus on participants from the Global South to get at issues of representation that confront marginalized groups in social movements.We follow Williams (1998, 16) in defining marginalized groups as those for whom 1) social and political inequality is structured along the lines of group membership, 2) group membership is not experienced as voluntary or mutable, and 3) negative meanings are assigned to group identity.These patterns of inequality persist over time, continuing into the present.Following this definition, women from the Global South are marginalized in transnational feminist settings (Ferree and Pudrovska 2006;Hughes et al. 2018).
The first part of the analysis, a factor analysis, establishes that transnational feminism is indeed a distinct arena of action, justifying the need for further conceptual development to facilitate analysis in this context.Next, we examine the differences between autonomous and multilateral transnational feminist action by comparing network analyses of three digital transnational feminist campaigns, #WomensDay, #WhyIStrike, and #NiUnaMas.In our analysis we pay particular attention to who participates and who is setting the agenda in these digital campaigns.We find that grassroots organizations from the Global South have more influence in online autonomous feminist networks than in multilateral ones.We employ a secondary, embedded case comparison of two Spanish-language campaigns to show that the differences we observe in our main comparisons are present on a more granular level (Yin 2012).
This analysis provides proof-of-concept by illuminating the difference that organizational forms make for women's participation in transnational feminism.Autonomy is a key dimension of feminist mobilization and must be addressed to fully understand the effects of transnational feminist mobilization on global politics and policy.Further, our analysis suggests that measures of feminist mobilization that reflect only established organizations (such as WINGOs), and that fail to incorporate more diffuse forms of activism (such as digital networks), may limit our understanding of how activists from the Global South participate in transnational feminism (cf.Hughes et al. 2018).In this way, our exploratory study motivates a new research agenda on transnational feminism.

Transnational Feminist Mobilization as a Distinct Dimension of Feminist Activism
Transnational feminism is at once a kind of movement, a network, an idea, and a mode of feminist mobilization.Feminism has often been defined conceptually as an idea or set of discourses (Mansbridge 1995;McBride and Mazur 2010).Movements are diffuse phenomena, that involve both individuals and organizations connected through both formal and informal ties around a shared idea and sense of purpose, and they involve political action of some kind (Forester et al. 2022;Mansbridge 1995;Mazur et al. 2016).
Transnational social movements coordinate political action around a shared ideas and values across borders (Smith 2005, 229; see also : Moghadam 2009;Smith 2008;Smith and Wiest 2012;Keck and Sikkink 1998).
Transnational feminist movements are identifiable by their commitment to addressing gender injustice, whether this is conceptualized by movements as male bias, sexual inequality, patriarchy, or gender hierarchy more generally (Forester et al. 2022;Moghadam 2000).A transnational feminist movement allows actors to organize under unifying issues rather than under the banner of their national origin or identity, opening new bases for organization and new pathways to political action (Moghadam 2000;Tripp 2006).Transnational feminist networks (TFNs) connect feminist activists from three or more countries.They are delineated by a shared goal or agenda that can be as expansive as the goal of addressing the gendered impacts of globalization, or as limited as a specific event or issue (Moghadam 2005, 4).Moghadam (2005) argues that TFNs are a distinctive organizational form spawned by global feminism: They transcend national boundaries and have the potential to be more detached from national level institutions and identities (Moghadam 2005).As such, transnational spaces offer the possibility of greater autonomy from established practices and domestic political institutions, such as parties, government agencies and national, regional, and ethnic identities.While actors may bring these identities and relationships into transnational spaces, they are also able to forge new relationships and identities.This is not to suggest that this space is free from power.Global politics including feminist movements are as structured by inequalities as their domestic counterparts (Mohanty 2003;Moghadam 2005).These contexts are nevertheless different from domestic ones.Transnational contexts enable (but do not require) greater organizational fluidity and experimentation, facilitating a wider variety of organizational forms.
For example, domestic feminist organizations often adopt a structure that mirrors federal structures or other geographic and ethnic divisions of the polity in which they reside.They may find themselves required to conform to registration and other legal requirements of particular states.TFNs have a greater ability to choose a favorable political regime and to organize in ways that incorporate multiple ethnic and geographic identities.Feminists may organize transnationally to establish autonomous feminist organizations when they do not have such opportunities in their home country.Such networks may be able to provide an avenue of participation for a wider group of women.We hypothesize that these features of transnational feminist mobilization make it distinct from domestic feminist mobilization, and that the various forms of transnational feminism will reflect their shared foci and/or organizational fields to some degree, when contrasted with their domestic counterparts.Thus, H1: Transnational mobilization constitutes a distinct dimension of feminist mobilization.
Such transnational feminist mobilization matters because it enables different forms of organization, each of which offers different advantages for representing women and for women's participation.This is especially true for historically marginalized and excluded groups, whose participation is complicated by power differentials.Though all avenues of representation are riven by power, the fluid and informal nature of political mobilization in movements makes it more accessible compared to elections or formal lobbying.Norms of inclusion enable better representation for marginalized women in transnational movements (Weldon 2006a;2011).
Distinguishing between transnational feminism's autonomous and multilateral dimensions deepens our understanding of the role that transnational feminist movements play in representing marginalized women.Given that feminist movements continue to struggle to ensure that women from the Global South are not dominated by women from the Global North in activist settings, the analysis pays particular attention to the degree of participation by women from the North and South in each form of mobilization as a way of examining the degree to which these different forms reflect and reinforce power differentials.
autonomous movements as having an organizational basis that is not only independent of the state but is also free of male-dominated organizations and authority structures, whether state-based or not.Autonomous organizations are neither subordinate to nor controlled by larger institutions that have overriding goals that prevent making gender equality or justice a fundamental priority.In domestic politics, this means autonomous feminist movements are independent of the state, political parties, and other associations that do not center the subordination of women or some subsection of women (Forester et al. 2022;Molyneux 1998;Ray and Korteweg 1999;Weldon 2002;2011).
Autonomy is central to understanding how feminist movements represent women.Autonomy matters because it enables activists to set a feminist agenda without having to compete with a set of priorities that does not include gender justice.Autonomous movements do face constraints of their own: They are answerable to their constituents-activists who may abandon the movement if they appear to move too far from their principles.But they are not subordinate to so-called "political bosses" who may be calling the shots in political parties, government agencies, and unions (McBride and Mazur 2013;Weldon 2011).Autonomous organizations also tend to be more accessible to intersectionally-marginalized groups of women, such as groups defined not only by gender but also by race, ethnicity, religion, sexuality, gender identity, and national origin, making autonomous movements particularly important for representation of marginalized women (Weldon 2002;2011).
To say that autonomous organizations have some distinct advantages is not to say that women, and women's movements, cannot be represented by other movement forms.State feminism is another prominent avenue of influence for women and women's movements in domestic politics.The concept of state feminism grew out of the political phenomenon of the proliferation of government agencies, aimed at promoting women's rights and gender equality.What makes state feminism influential is its lack of autonomy from political bosses: while ties to organized feminism can be important, being closer to power enables feminist actors to propose and influence policies (Mazur 2002;McBride and Mazur 2010;McBride and Mazur 2013).As with autonomy, state feminism is reflected in the organizational locus and context of mobilization.
Conceptually, the idea of autonomy is opposed to the concept of state feminism or integration, in the sense that autonomy signifies independence from established nonfeminist institutions, while state feminism works through integration into organizations.However, in practice, the organizational strategies of autonomous feminism and state feminism are matters of degree, and the participants and activists who engage in these forms of political mobilization often overlap.Further, politically, the strategies of organizing autonomously and internally to institutions are often pursued in tandem, with the combination of insideroutsider partnerships touted as being particularly influential in the domestic arena (Ewig and Ferree 2013;Gelb and Palley 1982;Prujt and Roggeband 2014;Weldon 2002).
How important is movement autonomy in the transnational arena?Do autonomous forms of feminist mobilization offer the same benefits of representation for marginalized women that they offer in domestic politics?What does it even mean to be "autonomous" in a transnational space?How might feminists pursue nonautonomous, or insider strategies in a transnational space?We cannot answer these questions without at least briefly considering the ways these concepts might work differently in the transnational context.
In its transnational form, autonomy implies that, in addition to being independent of national institutions such as domestic political parties and government agencies, feminist organizations are also organizationally independent of intergovernmental and multilateral agencies (like the UN or WTO).This does not mean that autonomous organizations do not engage with these institutions, but rather suggests that they have their own organizational bases outside of these spaces. 3Autonomous organizations often operate in multilateral settings and are not necessarily separatist or "rejectionist" organizations (Smith et al. 2021); They may well hold consultative status for one or more intergovernmental organizations.But they are distinguished by the fact that their primary focus is conceptualized as an independent one, and the multilateral setting is only one of many possible spheres in which they intervene, and they may well reject it altogether.Autonomy is discerned by examining the organizational form of movements.Some of the most important and pathbreaking transnational mobilization has been through autonomous forms of transnational feminist networks, from the Encuentros in Latin America to the Association of Women of the Mediterranean Region (Alvarez 2000;Moghadam 2000).One of the first feminist transnational meetings, the 1976 Tribunal on Crimes against Women, was formed in opposition to the UN meetings, precisely because of the limitations of the UN setting.Indeed, this autonomous meeting on violence against women resulted in some important conceptual advances, advances that were deepened at the autonomous NGO Fora run alongside the official UN meetings.Access to the official meetings was far more circumscribed, and the NGO Fora served as spaces where feminist activists could meet and formulate autonomous demands that they later pressed governments to adopt (Weldon 2006b).Similarly, the Latin American Encuentros enabled the development of regional connections across borders and the development of agendas for feminist action without being restricted to the extant agenda of, for example, the UN (Alvarez 2000).
On the international stage, there is no world government to which state feminist strategies might correspond, but there are strategies of integration into or opposition to international institutions.Smith et al. (2021) call feminist organizations that are more tightly tied to officially defined agendas, structures, and issues "multilateral" organizations.Without taking on all the elements of the concept (for example, leaving aside questions about the ideological orientation of those engaging in this form of organizing) this concept captures a form of feminist mobilization that is more oriented towards participating in official activities and less focused on articulating independent priorities.These purposes are made explicit in the stated goals of these organizations and are reflected in their organizational form.Multilateral feminist networks, which might be seen as an analog to state feminism on the international stage, seek to work with and within and transform multilateral institutions, and their organizational locus, foci and activities draw from and contribute to the agendas and activities of such institutions.
UN Women, the women's policy apparatus of the United Nations (UN), is a paradigmatic site of multilateral feminism.The UN has been an important locus for feminist mobilization, even an "Unlikely Godmother" of the international women's movement, as feminist conferences and agencies of the UN have driven feminist mobilization (Snyder 2006; see also Forester et al. 2022;Friedman 1999;Sandler and Goetz 2020).The feminist agencies of the UN helped to drive the idea of women's human rights, from violence against women to developing a new international economic order (Friedman 1995;Snyder 2006;Weldon 2006b).In comparison to multilateral institutions such as the WTO or NATO, the UN has a broader focus (going beyond economic development) and a long association with democracy and human rights.The combination of four UN agencies into the single powerhouse of UN Women in 2010 created a single agency with gender justice as its mission.This effort to advance feminist aims by organizing within the UN, pursuing insider strategies that capitalize on the proximity to power, is emblematic of a multilateral feminist strategy.
Distinguishing autonomous transnational feminist mobilization from its multilateral counterpart provides a clearer view of the distinct nature of these two interrelated concepts.We expect that the greater accessibility of autonomous TFNs will lead to different set of participants in comparison with multilateral networks: Autonomous networks will feature a more central role for individual activists and autonomous feminist organizations while multilateral transnational feminist campaigns will centrally involve more government officials, formal bodies of intergovernmental organizations, and representatives of multilateral organizations.This suggests the following expectations: H2a: In autonomous feminist networks, individual activists and autonomous feminist organizations will be more likely to be influential in shaping the agenda and discourse than in multilateral feminist networks.
H2b: The discussions in multilateral networks will be more driven by intergovernmental agencies and national government actors than in autonomous networks.

H3:
We expect autonomous networks to be better avenues of representation for women of the Global South than multilateral networks.

Multi-Method Approach: Using Factor and Network Analysis to characterize TFNs
We use a combination of analytic approaches to explore these questions.Through these analyses we demonstrate there are observable differences between multilateral and autonomous transnational feminist networks in the online sphere.We began with the question of whether domestic and transnational feminism are distinct.To address this question, we begin by establishing the distinctiveness of transnational feminist activism using factor analysis, showing that working out the concept of autonomy in this new area is both necessary and important.Our factor analysis also shows that even if they represent a different type of organizing, the digital campaigns we study are sufficiently similar to the traditional forms of activism that they warrant inclusion in the same conceptual field.We then turn to an analysis of three cases of online transnational feminist organizingthe multilaterally focused International Women's Day campaign, and the autonomous campaigns of International Women's Strike and Ni Una Más.We select these cases to explore the differences between autonomous and multilateral transnational feminist action, focusing on questions of who participates and who is setting the agenda in these digital campaigns.We also employ a secondary, embedded case comparison of two predominantly Spanish-language hashtags to demonstrate that the differences we observe in our overall analysis are also visible at this more granular level, and even when comparing two Spanish-language cases.In our study, we treat the North-South dimension as a critical axis of marginalization, but we do not seek to show that autonomous TFNs represent all of the Global South.Rather, we seek to show that organizations from the Global South have more influence in autonomous feminist networks than in multilateral ones, other things being equal.
Our focus on digital networks contributes to the research on transnational social movements by adding insight into another dimension of movement activity in online spaces, one that has been found to be critical for understanding contemporary movements (Jackson et al. 2020;Wright et al. 2022). 4A particular advantage of social media data, especially Twitter data, is that it allows us to "see" online communication, capturing traces of social movement behavior that might be missed by studies focusing on organizations and relationships captured by, for example, the Yearbook of International Organizations (YIO), which tends to capture more formal and betterresourced organizations (Smith et al. 2021).

Methods: Factor Analysis
We use factor analysis to demonstrate that transnational feminism and domestic feminism are distinct.Factor analysis shows the degree to which transnational and domestic feminism reflect the same underlying concept, the same shared variability.We also use it demonstrate the utility of using online campaigns to capture transnational feminist networks.
Scholars focused on transnational feminist movements primarily look for evidence of transnational organizing where activities involve connections across national boundaries, such as UN conferences (Moghadam 2000;Tripp and Ferree 2006), transnational feminist networks (Moghadam 2000), and WINGOs (e.g., Paxton et al. 2006;Hughes et al. 2015;Hughes et al. 2018).In the following section we give descriptions of the data used in the factor analysis.Domestic Feminist Mobilization.To measure domestic feminist mobilization, we employ a newly available measure of feminist mobilization, the Feminist Mobilization Index, a measure based on a database of 126 countries from 1975 to 2015 (Forester et al. 2022).This database encompasses a wide range of types of feminist mobilization over an unprecedented temporal and geographic scope.For example, this database offers unprecedented coverage of the Global South, especially Africa.This measure reflects the emphasis on strength and autonomy in the literature (Fallon and Rademacher 2018;Htun and Weldon 2018;Mazur et al. 2016;Tripp 2006) and is novel in its emphasis on feminist autonomy (Forester et al. 2022).It also reflects an effort to take the diversity of feminist mobilization into account (Cohen et al. 2018;Irvine et al. 2019).
The feminist mobilization index ranges from 0 to 3, where 0 means there is no feminist mobilization and 3 indicates a high level of feminist mobilization.For more on the definitions, sources of data and other questions, please see the published research note covering such questions about this publicly-available dataset (Forester et al. 2022).
WINGOs.The most used measure of transnational feminist mobilization is participation in women's international non-governmental organizations (or WINGOs) (e.g., Paxton et al. 2006;Hughes et al. 2015;Hughes 2018).Indeed, transnational social movement organizations are seen as the best way to get at international social movement activity, for transnational collective action across a wide range of areas (Smith 2008).We employ participation in WINGOs as a primary measure of transnational feminist activity.
Some scholars have looked at the proportion of feminist activity in each country as a way of capturing the relative strength of transnational feminism, a measure that controls for the growth of such organizations over time.Because we want to see the effect of this growth in strength, our measure uses the absolute value of these connections (counts) rather than the proportional measure.Organizations were identified from the YIO using a computer algorithm.This measure correlates well with the measure developed by Paxton et al. (2006), using the data archived at ICPSR.See appendix A for more details.
Women's NGOs Participating in the UN Process.Participation in state-oriented processes may be distinct from participation in relatively autonomous organizational activities.To examine a state-oriented dimension of transnational feminism, we compiled an original database of the number of NGOs officially participating in the various UN World Conferences on Women, or for follow-up meetings in 2009.These meetings are important examples of global civil society (Clark et al. 1998).These are counts of organizations based on the country in which they are headquartered.Please see appendix B for more details.By focusing on activism that is connected to UN processes and multilateral forums this measure enables us to capture this state-oriented dimension of transnational feminist mobilization.
Participation in International Women's Day Twitter Campaigns.We use a dataset of Tweets from the UN-led International Women's Day to measure engagement with institutional feminist campaigns.Counts of participation in International Women's Day Twitter campaigns were drawn from a larger database of feminist Twitter campaigns (see appendix C), purchased directly from Twitter through Gnip, an authorized seller.The count data used covers all tweets using the following hashtags #IWD, #InternationalWomensDay, #Interna-tionalWomensDay2015, and #IWD2015, over the week of March 3 to March 11 th , 2015 (thus covering the lead up to and discussions that follow International Women's Day on March 8 th ).Counts were recorded for each hashtag in our 2015 database by using the application MongoDB Compass to query a specific hashtag (such as #IWD) and the ISO country code included within the metadata for self-disclosed location (such as CL for Chile).The returning number of tweets were recorded in an Excel spreadsheet organized by country-year and hashtag (See appendix C for more information).The measure included in the analysis is a count of the total number of tweets using all the listed hashtags during the week of International Women's Day.

Analyzing and Distinguishing Domestic and Transnational Feminism
Domestic and transnational feminism are distinct but related phenomena as reflected in the pairwise correlations between our measures of these various types of activism (table 1) and our factor analysis (table 2).The Feminist Mobilization Index (FMI) and our count of national participation in WINGOs are correlated at 0.311 (table 1).The FMI is similarly correlated to the two other measures of transnational feminist activity (NGOs Participating in the UN Process and IWD Tweets).Interestingly, participation in online campaigns is tightly, significantly correlated with both participation in WINGOs (0.872) and women's NGOs participating in the UN process (0.913).As expected, this indicates that participation in the online UN-sponsored campaign (IWD Twitter campaign) is closely related to NGOs participation in the UN process and to organizing in WINGOs.These three variables seem to share variation not tapped by our measure of domestic mobilization (the FMI).
Our factor analysis (table 2) lends credence to the idea that transnational and domestic feminist activism, while similar, represent two different dimensions of the same phenomenon.Factor analysis of four variables measuring feminist organization reveals two dimensions, with the three transnational variables loading most strongly onto the first factor, while the domestic feminist variable loading roughly equally onto both factors.In addition, one of the transnational feminist variables loads negatively onto the second factor.The first factor may represent the transnational dimension (perhaps including more general feminist consciousness) while the second factor may represent elements more closely related to domestic activism.In the correlational analysis and (to some degree) the factor analysis, the IWD tweets and the NGO UN process variables appear to vary together, suggesting a close relationship between the IWD Twitter campaigns and participation in the UN Process.

Transnational Feminist Networks in Cyberspace: Feminist Twitter Campaigns as TFNs
In this section, we use online feminist campaigns to differentiate between autonomous and multilateral feminism.Twitter hashtags offer some advantages for social movement scholars as sites of digital activism, even as the future of Twitter itself is unknown.Participation in any Twitter discussion is fluid, constrained only by the number of characters-240 characters per post for each user during the period of study (Costanza-Chock 2020; Jackson et al. 2020).Hashtags on Twitter organize discussions by topic or theme, and in some sense, they can be seen as constituting public spheresdiscussions where citizens gather to discuss matters of interest relatively free from the interference of the state (Costanza-Chock 2020; Habermas 1991; Jackson et al. 2020).Social movements contribute to democracy by building civil society, by creating and participating in public discussions.Online activism is no exception.

Hashtags as Organizing Transnational Publics and Counterpublics
Democracy includes multiple, overlapping, and decentered discussions, including both dominant publics and   (Fraser 1992, 123).Social media is an important way of creating contemporary counterpublics (Jackson et al. 2020).Activists were early adopters of the use of hashtags and are strategic in their use of hashtags to amplify messages and connect Twitter users interested in similar topics of discussion (Costanza-Chock 2020; Jackson et al. 2020).The use of hashtags on Twitter is one way that individuals and organizations indicate their identity as a movement participant, their support for the framing of an issue, their participation in a campaign, or to challenge the frames used by their peers (Stewart et al. 2017).Social movements of marginalized groups sometimes use hashtags to constitute counterpublics that serve as spaces for developing oppositional consciousness and for developing constituencies (Jackson et al. 2020).Activists can use these oppositional spaces as jumping off points from which to address and intervene in dominant public spheres (Jackson et al. 2020).Therefore, hashtags enable us to map out the online networks involved in a specific publicor counterpublicand develop insights into how transnational feminist organizing functions in cyberspace.
We use hashtags to identify our examples of digital publics created by multilateral and autonomous feminist networks.International Women's Day, as represented by the hashtag #WomensDay, represents the public, the discussion or deliberation, created by the multilateral transnational feminist network on Twitter.The Women's Strike, as represented by #WhyIStrike, is an example of a counterpublic created by an autonomous transnational feminist network.The strike consciously used IWD as the day of their action to drive attention to a range of feminist issues including, but not limited to: reproductive justice, violence against women, women's care work, and women's labor.To explore the difference that language makes, we further examined the case of #NiUnaMas, an example of an autonomous and predominantly Spanish hashtag.Last, using an embedded case study design, we examined two Spanish hashtags with parallel foci to IWD and the Women's Strike: #YoParo8M and #díadelamujer to compare across our categories of autonomous and multilateral feminism while controlling for dominant language (see table 3).These latter cases constitute embedded case studies that offer more detailed, controlled comparisons within the broader conversations or campaigns (Yin 2012).
Twitter data enables us to analyze the structure of the discussion that constitutes the public sphere.Tracing how social media users interact with one another's interventions or "tweets" allows us both to delineate the network and to track the movement of ideas through the network (Jackson et al. 2020).Online actions such as "retweeting" serve to amplify the messages presented by fellow users, and those individuals or organizations which receive more attention in the network hold discursive power within this space.Examining who is participating in these hashtags, and who is retweeting whom, enables us to map out the online networks involved in a specific campaign and develop insights into how transnational feminist organizing functions in cyberspace.
Social media data provide observable traces of social movement activity (Jackson et al. 2020;Wright et al. 2022).Social media data is only one way to observe social movements.Like all forms of data, it has limits, for example, reflecting differential access to technology and savvy required to access digital networks such as cellphones.By the same token, however, traditional measures such as counts of formal organizations suffer from a focus on formal organization that tends to overlook informal, diffuse forms of activism.Analysis of social media data offers a different source of data that has the potential to reveal new insight.
We proposed that TFNs would be particularly important for marginalized groups.Women from the Global South have historically been blocked from full participation in transnational settings, and their exclusion appears to continue into the present.Do women from the Global South participate in digital TFNs?Does this participation vary across autonomous and multilateral TFNs in the way we expect (recall H2a and H2b above)?Twitter data can show whether organizations and activists from the Global South who aim to represent such women participate in some TFNs more than others.Again, our aim is not to show that these participants speak for all women from the Global South.Rather, we seek to demonstrate that some TFNs give a more central role to women from the Global South participating than other TFNs (H3).This is a relative, not absolute, judgement.In our study these women come primarily from Latin America, and we do not claim that women from Latin America speak for or represent women from every region or part of the Global South.Indeed, whether TFNs are equally open to women from every region of the Global South is an area for further scholarship in the research agenda we lay out in this paper.were chosen to represent autonomous transnational feminist networks as these latter hashtags were defined by activists organizing outside male-dominated organizations.International Women's Day has a complex history but today it represents a campaign that is squarely tied to official activities of the UN, making it a good example of multilateral feminism.Originally, International Women's Day was grounded in socialist feminist efforts to make the issue of woman suffrage a priority for the international socialist movement (Kaplan 1985).After World War II events were scattered throughout the world until the UN established an official International Women's Day in 1977 (Boxer 2009).In the 1970s and 1980s, the association of International Women's Day with the UN led to its association with the re-emergence of a strong international feminist movement with close ties to the UN conferences (Çağatay et al. 2021).As our factor analysis demonstrates, online campaigns represented by International Women's Day hashtags are closely associated with the UN.International Women's Day's close association with multilateral processes, and the willingness of corporations and governments to use the day for celebration of women has led some scholars to argue that it has lost its radical potential (Çağatay et al. 2021).Yet, there are still concrete efforts by feminists from the local to the transnational level to center autonomous feminist organizing on International Women's Day (Çağatay et al. 2021;English and Campbell 2020;Kelly-Thompson 2020).
The International Women's Strike of 2017 is an autonomous transnational feminist action that strategically used International Women's Day as a day of transnational protest for gender equality with an organizational base outside of the UN.It arose out of autonomous transnational organizing between Polish and Argentinian feminists who were using strikes to bring attention to the problem of violence against women in Argentina and reproductive rights in Poland in late 2016 (International Women's Strike/Paro Internacional de Mujeres 2020; Kelly-Thompson n.d.).After their second strike, on October 24, Polish organizers began reaching out to those who showed transnational solidarity with their strike.This led to the development of a group of organizers from seven countries-Poland, South Korea, Russia, Argentina, Ireland, Italy, and Israel -working for a global strike.The strike was planned for International Women's Day 2017 and included feminists from 35 countries. 5In conjunction with the on-the-ground organizing for the Strike, organizers used the hashtags #WomensStrike and #WhyIStrike to publicize and promote the day of action.
A third TFN is Ni Una Más, an autonomous TFN that focusses on a different issue (gender-based violence), in a (predominantly) different language (Spanish), and representing an older, more established campaign.The movement's name comes from the phrase "Ni una muerta más, ni una menos" (Not one death more, not one (woman) less) coined by Susana Chávez, a Mexican poet and femicide activist, in 1995.The campaign links shared struggles by Latin American feminists to address femicide (Blanco 2019).The campaign, originating in the 2002 Mexican feminist anti-femicide campaign called Ni Una Más, came to encompass the Ni Una Menos campaign.In 2015, Argentinian feminists organized a collective to address patriarchal violence under this name and held protests across Argentina (Blanco 2019;Ni Una Menos n.d.;Wright 2006).As a campaign, Ni Una Más reflects the prevalence of autonomous feminist organizing in Latin America (Blanco 2019).Interviews with participants indicate that the online activism associated with Ni Una Menos is linked to on-the-ground actions by femicide activists (Sjoberg 2019).

Network Analysis
Network analysis enables us to examine how the individuals and groups in these online feminist networks are connected (Jackson et al. 2020;Ferree and Pudrovska 2006).As a method, network analysis draws inferences from social connections.Network analysis has roots in graph theory and is used to depict the relationships between any number of units in the observable world, such as including people, organizations, or communications (Coleman 1964;Scott 1988).Network analysis uses the terminology of nodes or vertices to represent individual actors in a network and edges or ties to describe how those individual actors are connected to each other (Wasserman and Faust 1994).This approach allows the researcher to measure aspects of a network such as its density, and who are the most central actors.
The ties or the relationships in a network tell a story about what is happening in the network.Nodes in the networks have ties to one another, sometimes the ties are reflected in actions that have a direction from one node to another.Mapping these actions provides the researcher the ability to study the structure of the overall network, and also enables the examination of individual relationships between people in this network (Granovetter 1973).In a Twitter network, for example, we can say a user (or node) with a high centrality score in a Twitter network is being retweeted and responded to more frequently than other users in the network.In contrast, a user with a low centrality score is being retweeted and responded to less frequently than other users in the same network.
Network analysis is particularly useful for analyzing social movement activity on social media, the ability of network analysis to identify central users illuminates how information moves through the wider network (González-Bailón and Wang 2016; Jackson et al. 2020;Wright et al. 2022).Network analysis enables addressing the influence of a particular user or idea in ways that analyzing individual tweets cannot (Jackson et al. 2020;Wright et al. 2022).We use in-degree centrality, which can be read as an indicator that a particular node's message is being spread further than other nodes (González-Bailón and Wang 2016).
The following network analyses are based on snapshots of the Twitter activity from the following Twitter campaigns: #WhyIStrike, #WomensDay, and #NiUnaMas.The #WhyIStrike and #WomensDay campaigns occurred on the same day, enabling us to directly compare the networks to each other. 6For this comparison, a maximum of 2,000 tweets for each hashtag was pulled from Twitter's API using the program NodeXL on March 6, 2017. 7To address whether or not our findings are generalizable beyond the Women's Strike, a coordinated one-day campaign and address the limitations of using only English language hashtags, we include an analysis of Twitter discourse around Ni Una Más-an ongoing transnational feminist network organized around a shared commitment to ending violence against women.The network of tweets using the hashtag #NiUnaMas was pulled from Twitter using the program NodeXL on February 6, 2020. 8Furthermore, #NiUnaMas is a Spanish language hashtag linked to a movement led by women in Latin America, enabling us some leverage on the limitations presented by using the English language hashtags #WomensDay and #WhyIStrike for the events on International Women's Day 2017.There are a total of 1,554 tweets in the network using the hashtag #NiUnaMas from February 4, 2020, to February 6, 2020.Excluded from the networks are Tweets that use the hashtag without mentioning, retweeting, or replying to a Twitter user.
While the hashtags used for these analyses are in English, #WhyIStrike and #WomensDay, and Spanish, #NiUnaMas, these are not the only languages present in the dataset (See appendix F).In the case of #WhyIStrike and #WomensDay, the top used co-occurring hashtags in both networks include hashtags in English and Spanish (see appendix G).Likewise, in both campaigns, users employ English hashtags with non-English tweets (see appendix F).
When comparing these networks, we expect to see differences between campaigns in terms of the kinds of users who are most central to the networks.In the multilateral transnational feminist campaign, #Womens-Day, we expect that the users with the highest centrality scores will be more likely to be multilateral-linked organizations, intergovernmental organizations, or national organizations.In autonomous feminist campaigns, #WhyIStrike and #NiUnaMas, we expect that the users with the highest centrality score will be more likely to be individual feminists and autonomous feminist organizations.Likewise, we expect to see a wider range of participants in terms of their geographic location (with more users from the Global South) in autonomous campaigns (#WhyIStrike and #NiUnaMas), when compared to a multilateral campaign, #WomensDay.

Analyzing Autonomous and Multilateral-Oriented Networks: Transnational Feminist Twitter Campaigns
We address the differences between autonomous and multilateral feminisms by comparing the network characteristics for three Twitter campaigns: International Women's Day, Ni Una Más, and the Women's Strike.Ni Una Más and the Women's Strike of 2017 both arise out of autonomous feminist mobilization developed by feminist organizers working across borders.As we have established, International Women's Day tweets by country are highly correlated with NGO's involved in the UN process.Therefore, we use International Women's Day tweets to represent multilateral transnational feminist activism online.Importantly, as demonstrated through our factor analyses, we should expect that these transnational campaigns are distinct from domestic feminist organizing.For example, the 2017 Women's Strike is distinct from the strikes organized by only Polish women in 2020 and 2021 against Poland's abortion bans.Polish women's participation in the 2017 Women's Strike thus represents involvement in a transnational action for gender justice.
Table 4 provides a summary of the three primary networks we analyze, the multilateral network, #Womens-Day, and the two autonomous networks, #WhyIStrike and #NiUnaMas.The network for #WhyIStrike is denser (0.003) than both the networks for #WomensDay and #NiUnaMas (0.001).
First, there are fewer Twitter users engaging with the #WhyIStrike hashtag, but there are more connections between these users (table 4), making it a smaller, more densely connected network: The network created for #WhyIStrike has a higher score for network density (0.003), than the network developed for #WomensDay (0.001).#NiUnaMas' graphs total density is 0.001, a similar level of density when compared to #Womens-Day in 2017, and less dense than the #WhyIStrike network in 2017.Users engaged with #WhyIStrike are more likely to engage with each other than those involved in the International Women's Day campaign or in Ni Una Más.
We see differences in the number of unique edges within the networks.Edges indicate the connections between the users, through retweets and responses to each other's content, while unique edges indicate the number of unique connections between users.When comparing #WhyIStrike with #WomensDay, we see that despite the smaller number of users, #WhyIStrike has 202 more unique connections when compared to #WomensDay. 9There are 750 Twitter users represented within the #NiU-naMas network (table 4).The majority of the edges (1409 of 1554) are unique, indicating that, on average, users have approximately 2 unique connections.Self-loops refer to the number of tweets where the user who made the tweet responds or retweets their own content, this captures common Twitter behavior such as threading where a user connects multiple tweets together through replies to create a longer narrative.#WomensDay tweets have the highest number of self-loops (182), followed by #NiUnaMas (175), and #WhyIStrike (112).This could indicate that when compared to #WhyIStrike more users in #Womens-Day and #NiUnaMas are threading their tweets to develop a longer narrative or are retweeting their own tweets to further amplify their original messages.

Women's Strike versus International Women's Day 2017
In the following section we compare who sets the agenda in the multilateral network of #WomensDay and in the autonomous International Women's Strike hashtag, #WhyIStrike.Some differences emerge in our analysis of the top users by in-degree centrality in these two campaigns.Through this analysis we find strong support for our second hypothesis, that autonomous networks are more likely to center autonomous organizations (H2a) and that multilateral networks center multilateral organizations (H2b).
Beyond the differences between these two networks in terms of their density, and the connection between the users engaged in these two campaigns, there are also distinct differences between the users who are best able to promote their message and generate engagement.The top users for #WhyIStrike by in-degree centrality (red triangles in figure 1; table 5) are mostly autonomous organizations, such as the official International Women's Strike account while #WomensDay is dominated by UN-affiliated accounts (red triangles in figure 2; table 6).The users that gain the most attention from other users within #WhyIStrike include the official accounts for the International Women's Strike and the Belgium Women's Strike and other autonomous feminist organizations such as the English Collective of Prostitutes and FAB! Feminisms Against Borders (see table 5).Many of these organizations are smaller, and many are explicitly transnational and autonomous (for example, FAB!).Even though the same user, womensmarch, appears to be the most successful in terms of amplifying its message in both campaigns (ranking highly in in-degree centrality), closer examination reveals that womensmarch does not play the same role in both networks.Specifically, in #WomensDay, womensmarch is the top user developing connections between users within the network (betweenness centrality) while in #WhyIStrike it does not play this connective role (see appendix D for more details).This suggests that one of our expectations for autonomous versus multilateral networks seems to be borne out by this comparison: As expected (H2), users representing official accounts and agencies play a more central role in driving the discussion in multilateral agencies, while in autonomous networks, individual activists and grassroots, independent feminist organizations played the central role.
The comparison is less clearly supportive of our expectations related to the accessibility of these networks to women from the Global South (H3).Comparing tables 5, and 6, we see some differences in where the top users for each campaign are located.For #WhyIStrike, many of the top users who share their location represent the Global North, except for the International Women's Strike.The user represented by the International Women's Strike represents multiple countries and activists from both the Global North and South.The other top users represent organizations in Belgium (Women's Strike Belgium), Italy (Non Una Di Meno chapters), Spain (Juventud Comunista en Madrid), the United States (Women's March, Inc.), and the United Kingdom (English Collective of Prostitutes).In terms of region, most of the top users are located in the Global North-specifically Europe.The network represented by this set of tweets, however, may be limited by the use of the English language hashtag: all tweets pulled had #WhyIStrike in the tweet.However, it is important to note that not all tweets are fully in English or only use English hashtags (see below and appendices F and G).
The importance of organizations within the Global North may reflect, at least to some degree, the use of the English hashtag by these users.
A majority of the top users for #WomensDay are located in the United States (table 6), which is likely in part due to the UN's headquarters in New York (and thus is the home of UN Women and the UN Global Compact).Beyond the United States, two users (Agustnmilln and RumboaGaza) indicate their location as being in Spain, while two (unpeacekeeping and womensday) indicate a global position.This indicates that, despite their differences, in these two campaigns the Twitter users who are dominating the conversation are mostly located in the Global North.This is consistent with prior work on transnational feminist mobilization (Ferree and Pudrovska 2006;Hughes et al. 2018).

Twitter Networks for International Women's Day versus #NiUnaMas
Another autonomous network, #NiUnaMas, contrasts more sharply with the online IWD campaign, and has some expected similarities to the Women's Strike campaign.Much like the International Women's Strike, the network for the Twitter campaign #NiUnaMas is dominated not by UN Women or WINGOs, but instead by individual feminists (as anticipated by our Hypotheses H2a and b).However, unlike #WhyIStrike, the network for this online campaign reflects the geography of the larger transnational feminist network of Ni Una Más, a predominantly Latin American transnational feminist network.Though there are many differences between #NiU-naMas (with language and topical focus being obvious ones) the multilateral-autonomous comparison here may lend some support to our theoretical expectations as outlined by H3.
Most of the top 10 users for #NiUnaMas (as defined by their in-degree centrality) are individual activists using the hashtag to promote addressing violence against women (table 7).This indicates support for our second hypothesis that autonomous feminist networks are potentially more accessible for individual feminists than multilateral networks (H2a).
As indicated in table 7, of the top 10 users by in-degree centrality, the majority who indicate their location are in Mexico, with only two of the top 10 indicating their location as in the Global North (the United States and Spain).Furthermore, most of the users engaging with the hashtag are in the Global South (67%), primarily in Latin America (see appendix D).However, many users from Spain (26%) also engaged with the hashtag.This provides some support for our expectations about autonomous networks greater openness to participation by women from the Global South (H3), though the many differences between #WomensDay and #NiUnaMas make it difficult to attribute the differences solely to the autonomous or multilateral nature of the network.To further explore whether these findings comparing #NiUnaMas to #WhyIStrike and #WomensDay are a function of the difference of the language of the hashtag (Spanish), we also develop two embedded cases.Our embedded case study lends further support for our hypotheses.
Embedded Cases: Spanish-Language Hashtags for Women's Strike (#YoParo8M) Compared to IWD (#díadelamujer) In the following section we focus on our embedded cases, #YoParo8M and #díadelamujer.Twitter users use multiple hashtags to create links between conversations within the larger social network, in this section we focus on the hashtags used to bridge conversations in multiple languages (e.g., Wang et al. 2017).#YoParo8M (I strike on 8 th of March) is the top used Spanish language hashtag with #WhyIStrike.#Díadelamujer (Women's Day) is the top used Spanish hashtag with #WomensDay (For more details on the top used hashtags in all networks see appendix G).Within these embedded networks we have overall similar findings to the larger networks for #WhyIStrike and #WomensDay.
There are differences with respect to their size and density as these are smaller networks.The network for #YoParo8M is denser (0.01) than the larger network of #WhyIStrike (0.003), as is the network for #díadelamujer (.08) when compared to #WomensDay (0.001) (table 8).The smaller network, #díadelamujer, is denser than the larger network, #YoParo8M.
The network for #YoParo8M is larger than the network for #díadelamujer.#YoParo8M includes 100 Twitter users with 121 unique edges (or tweets) while #díadelamujer includes only 13 users and 13 tweets.Even though #WhyIStrike is a much smaller network than #Womens-Day, a larger percent of the #WhyIStrike network combines both English and Spanish hashtags.Perhaps this campaign does more to bridge linguistic divides by connecting Twitter users using different languages.
One of the key differences between #YoParo8M and #díadelamujer is in the percent of the larger network they   This embedded comparison between both #YoParo8M and #díadelamujer reveals patterns similar to those we observed in our comparison of the primary networks.The top users for the autonomous hashtag, #YoParo8M, by in-degree centrality are mostly autonomous organizations (such as Non Una di Meno) and individual feminists (table 9 and indicated as red triangles in figure 4).This is similar to the networks for the autonomous networks #WhyIStrike and #NiUnaMas, where autonomous feminist organizations and individual feminists dominate the network.Likewise, we find that for the more multilaterally-oriented #díadelamujer the majority of users who are retweeted and responded to within the network represent the United Nations (table 10 and indicated as red triangles in figure 5). 11As in our main case comparisons, we find support for our second hypothesis: The multilaterally-oriented network #WomensDay is driven by users affiliated with the United Nations (H2b) while the autonomous networks reflected more grassroots and independent actors and activists (H2a).
In #YoParo8M, all top users (by in-degree centrality) represent organizations linked to Ni Una Más/Menos.These top users are primarily Italian despite the hashtag being in Spanish.This is explained by the fact that the most retweeted and responded to tweets using the Spanish hashtag used terms in multiple languages including Italian. 13Non Una di Meno is the Italian translation of Ni Una Menos and reflects how the transnational feminist network has expanded to Europe.In this way, the users NonUnaDiMeno, NonUnaDiMenoMI, and NUDMCagliari represent Italian feminists involved in a transnational feminist campaign that originated in Latin America.
Importantly, there is overlap between our top users (measured by in-degree centrality) in both #YoParo8M and #WhyIStrike.The national Italian and the local Milanese chapters of Ni Una Di Meno are top users by in-degree centrality in both networks.In fact, the Milanese chapter NiUnaDiMenoMI, has the same score for in-degree centrality in both networks.This means that NiUnaDiMenoMI uses both #YoParo8M and #WhyIStrike in every tweet retweeted or responded to within the network.This indicates the centrality of the multilingual conversations within the larger network for #WhyIStrike, and these autonomous feminist organizations' explicit effort to engage with each other across potential language barriers.
Looking at the top users by in-degree centrality in #díadelamujer, there are only four users with any measure of in-degree centrality, meaning that only four users are retweeted or replied to within the network.Of these users three represent the United Nations, with the top two users being the official accounts for the UN and UN Women in Spanish.Furthermore, none of these users are top users within the larger #WomensDay network.Therefore, while #WomensDay has more linguistic diversity (appendix F) these conversations are not as central to the larger network when compared to #WhyIStrike.This finding may indicate a key difference between how multilateral organizations and autonomous feminist organizations use Twitter.UN Women has multiple accounts in multiple languages.The UN Women Spanish language account tweets using both the English and Spanish hashtags, while the English language UN Women account does not.
Overall, these embedded cases provide further support for our claim that there are unique types of transnational feminism that can be distinguished through an analysis of online campaigns.These differences can be distinguished in both English language and Spanish language online transnational feminist campaigns.Multilateral organizations are the central users for both #diadelamujer and #WomensDay, and autonomous organizations are the central users for both #YoParo8M and #WhyIStrike.
Interestingly, the nested cases also add nuance to our finding that #WhyIStrike is dominated by users from the Global North: Though the most central users in #WhyIStrike are from the Global North, these activists appear to be taking their inspiration from activists and ideas from the Global South.In the #WhyIStrike network, the conversation is being driven by feminists of the Global North but these feminists are using frames developed by Mexican activist Susana Chávez and popularized by Argentinian feminists.
This analysis helps to show how distinguishing between multilateral and autonomous transnational feminism might inform debates about the degree to which activists from the Global North dominate transnational feminist organizing.Though both autonomous and multilateral campaigns tend to be dominated by activists based in the Global North, autonomous campaigns appear to be less North-dominated than multilateral ones.Especially when they originate in the Global South, autonomous campaigns may do more than multilateral campaigns to center the perspectives of Southern women.Sometimes, this may be due to the way they evolve, as in those cases where autonomous campaigns, such as Ni Una Más, develop connections across regions.
Importantly, there are differences in the location of the users driving the conversations in International Women's Day, the Women's Strike, and Ni Una Más.In contrast to #WomensDay and #WhyIStrike, the main drivers of the conversation in #NiUnaMas come from the Global South.The most central users in #WomensDay and #WhyIStrike come from the Global North (primarily the United States, Spain, the United Kingdom, Belgium, Italy, etc.), and most of the top users in #NiUnaMas come from Mexico.Furthermore, when looking at the overall number of users (not just those that are most central) we see a similar trend with most of the users for #WomensDay and #WhyIStrike coming from the Global North, and most of the users for #NiUnaMas coming from the Global South.The more multilaterally-tied campaign of #WomensDay has a broader geographic scope (more countries) but conversations in different languages may be more siloed, indicating both the continued strength of multilaterally-oriented transnational feminism and its potential limitations.Overall, these analyses provide mixed evidence about whether   autonomous campaigns are more likely to engage users from the Global South: Our autonomous Spanish-language network is much more reflective of Southern women while #WhyIStrike, English language and autonomous, was still slightly more inclusive of Southern frames, but still dominated by users from the Global North.Our embedded case suggests there is more to this difference than language.Further it suggests that  autonomy alone is not a silver bullet for countering marginalization.Other considerations-the origin and locus of organizing may determine the legibility, accessibility and openness of a network.Other norms of inclusivity-descriptive representation of leadership, symbolic and discursive openness, and other strategies for intersectional solidarity-may also be necessary (Weldon 2006b;Einwohner et al. 2021).Furthermore, this presents opportunities for future empirical work on these questions.Beyond the differences in regional scope, there are also differences in terms of which users are driving the conversation and setting the agenda.Where the conversation in International Women's Day hashtags is dominated by UN organizations (e.g., UN Women and UN Peacekeeping), the Women's Strike message was primarily driven by smaller feminist organizations (e.g., Women's Strike and FAB! Feminisms Against Borders), and within #NiUna-Mas individual activists were the main drivers of the message.

Conclusion and Implications
This exploratory set of comparisons lends some support to our main hypotheses, though the nature of the cases and the comparisons makes these conclusions more suggestive than definitive.We find strong support for our first hypothesis that transnational and domestic feminist mobilization are distinct but related phenomena.Our factor  analysis demonstrates that while domestic and transnational mobilization are distinct phenomena, they are related to each other, suggesting they reflect some aspect of the same underlying phenomenon.Furthermore, through this factor analysis, we find that the number of International Women's Day tweets by country are closely linked to the number of NGOs involved in the UN process by country, indicating that we can use Twitter participation as one indicator of transnational feminism, a strategy we pursue in our analyses of multilateral and autonomous transnational feminisms.
Our findings on the exact nature and implications of autonomous and multilateral feminisms are more preliminary.Our comparative network analyses provide support for our second hypothesis that the users involved in autonomous and multilateral oriented transnational feminist digital campaigns are different.Autonomous feminist campaigns are more likely to amplify the messages of individuals and grassroots activists (H2a), when compared to multilateral campaigns which are more likely to amplify the messages of states and intergovernmental agencies, such as UN Women (H2b), even if both forms of organization comprise voluntary, civil society actors who originate and act outside of governments and IGOs.We find evidence of increased presence of the Global South in the #NiUnaMas campaign, but do not find the same presence in the #WhyIStrike campaign (H3).These findings, while preliminary, provide evidence that there is value in distinguishing between autonomous and multilateral forms of transnational feminist activism, though further study of dominant language and topical focus, and how these relate to regional feminist networks (and not just North or South), would be necessary to draw firm conclusions.Overall, this analysis reveals differences in terms of who drives the conversation in these different types of networks, with autonomous feminist groups and activists having greater influence in freestanding networks than multilateral ones.Our comparisons also revealed differences in the global locations from which key activists drive the conversation, though these did not fall as neatly along the lines of our theoretical expectations.While one autonomous network differed from the multilateral network in terms of its representativeness of women from the Global South, the other did not, pointing to the need to think more about what additional factors might magnify the impact of autonomy in transnational contexts.We suggest examining whether the strategies of intersectional solidarity can enhance the impact of autonomy on accessibility to marginalized women (Einwohner et al. 2021).
Within the different online conversations, we find evidence for distinct threads of transnational feminist organizing, a more multilaterally-oriented transnational feminism tied to the UN process (#WomensDay) and autonomous transnational feminisms within both the Global North and the Global South (#WhyIStrike and #NiUnaMas).In the future, examining Twitter in languages other than English and Spanish, and examining cases with purchase in other areas of the Global Southsuch as a campaign by WLUML-would be particularly interesting for examining the reach of transnational feminist campaigns online.Finally, we reiterate that we do not claim that all parts of the Global South are equally wellrepresented by the grassroots organizations we identify, nor do we claim that all parts of the Global South have equal access to TFNs.We leave these questions for future research.
This analysis shows the utility of parsing the different dimensions of feminist organizing.Through our factor analyses we demonstrate that domestic feminist mobilization, WINGOs, and transnational feminist online campaigns measure various aspects of feminist mobilization.Furthermore, we have preliminary evidence that demonstrates some of the ways these forms of mobilization intersect and interact.For example, the Women's March Twitter account, an autonomous national feminist organization, engages with both autonomous and multilateral transnational feminist campaigns.This account plays a key role in connecting users within the multilateral transnational campaign of #WomensDay (see appendix E).Similarly, regional, and local chapters of Non Una Di Meno play key roles in amplifying the message of the movement within the online campaign of the Women's Strike in multiple languages.These linkages between the local, national, regional, and transnational provide avenues for connecting autonomous feminist organizing between these different political spheres.By identifying the distinct dimensions of transnational and national feminist organizing we can better unpack the role of feminist movements in policy change.
Future research should work to develop cross-national measures that capture autonomy as a dimension of transnational feminism: When women set their own agendas, those agendas will reflect women's distinctive perspectives and concerns (Molyneux 1998;Weldon 2002).Identifying and measuring these different types of transnational feminist networksthat is, multilateral feminism and autonomous feminismmay enable us to develop more precise research questions.Instead of asking whether transnational pressure advances women's rights or whether it undermines them, we can ask questions about the conditions under which transnational feminism advances or undermines these rights.Adding to questions about frame alignment (Swiss and Fallon 2017) and differences across issue-types (Htun and Weldon 2018), we can add questions about the type of transnational feminist mobilization we are examining.When do multilateral and autonomous transnational feminism advance women's rights?Do these types of feminism have different degrees or kinds of Figure 1 Network for #WhyIStrike, 2017 Figure 2 Network for #WomensDay, 2017

Table 3
Logic of Case Selection Perspectives on PoliticsCase Selection: Multilateral and Autonomous Online Transnational Feminist Campaigns Comparing network characteristics for three Twitter campaigns-namely, International Women's Day, Ni Una Más, and the Women's Strike-allows us to explore the differences between autonomous and multilateral campaigns.International Women's Day represents multilateral feminism while Ni Una Más and the Women's Strike

Table 6
Top Users for #WomensDay by In-Degree Centrality

Table 7 Top
Users for #NiUnaMas by In-Degree Centrality

Table 9
Top Users by In-Degree Centrality #YoParo8M

Table 10 Top
Users by In-Degree #díadelamujer