How might one reconcile feminist international relations theorizing with quantitative methods, if they can be reconciled at all? This question has occupied a central place in the field of feminist international relations theoryFootnote 1 since its inception, and while it is not my intention to rehash it in its entirety here,Footnote 2 my story sheds light on the difficulties that feminist IR scholars may face in choosing the quantitative route. My essay highlights these conundrums from different stages of the research process: formulation of the research question, selection of a research method, usage of quantitative methods, and the dissemination of the findings. I conclude with questions for further consideration.
My discussion is informed by a number of different vantage points. My Ph.D. is in political science and I have been trained in international relations—both mainstream and feminist or conventional feminist, as Mary Caprioli (Reference Caprioli2004) refers to it—as well as in quantitative methods. Nevertheless, my institutional home for much of my career has been in Women's Studies departments, where I regularly work with colleagues and students from many different disciplines, such as english, biology, sociology, history, and film making. I am rarely recognized in Women's Studies as a “feminist IR scholar” since this identification really has theoretical and perhaps methodological meaning only in the field of international relations. Working in this context has allowed me to reflect more broadly on what feminist scholarship is and should be beyond the field of feminist IR, and it has also led me to look for answers to my dilemmas in other fields, especially feminist economics and development studies.
My own journey as a feminist IR scholar is definitively marked by a shift away from the question over whether or not we ought to engage in and utilize positivist research methods (for a long time my answer was no) to an attempt to understand the limits and potential promise of these methods and the ways that feminist IR theorizing might improve the data collection needed for quantitative modeling. While it is certainly important and necessary that we continue to be self-reflective about the question of whether or not the master's methodological tools can dismantle or transform the master's positivist house, it is equally important, in my mind, that feminist IR scholars also recognize the necessity of being able to dictate the terms of our participation in positivist research agendas.
What I mean by this is that while rejecting positivist frameworks and corresponding research methods might be thought of as one way to dictate the terms of our (non)participation in, or acceptance of, prevailing mainstream positivist models, the rejection of positivist ideology and quantitative models does not make them go away anymore than rejecting patriarchy or racism makes these ideologies and systems go away. Because these research methods serve to “normalize” mainstream IR's perspectives on neoliberalism, militarized conflict and security, and democratization, we also need to disrupt these narratives from within to show how mainstream IR is incomplete even on its own terms. Therefore, what we should be rejecting are not all positivist models merely because they are positivist, regardless of how they are theoretically informed, but rather positivist models that do not draw on the considerable theoretical insights of feminist IR scholars. As Cecile Jackson (Reference Jackson2002, 499) notes, while the use of sex disaggregated data in quantitative modeling is a welcome development, unless it is “harnessed to an understanding of gender relations, it is analytically impoverished.” If we do not employ both inside and outside methodological strategies, not only do we lose a “seat at the table” (Barker Reference Barker, Kuiper and Barker2006) but we also run the risk of “privileging another master's house with its own compromising perils” (Staudt Reference Staudt and Saunders2002, 58).
In my mind, quantitative methods are but one of many useful tools for supporting an overarching feminist methodology. Indeed, these are not the only research methods that I use. Due to the demands of many feminist scholars and activists in multiple fields, sex disaggregated data collection is expanding, not shrinking, and this development virtually guarantees its inclusion in future quantitative modeling. Much of the increased data collection, such as data about time use, is actually due to, and informed by, the qualitative and theoretical work of feminist scholars. We should support these efforts and use the data in our work, however we see fit to maintain the urgency of collecting such data. The question becomes how we most effectively utilize these methods and data toward the goal of transforming hierarchical gender, racial, ethnic, class, and sexual relations that feminize and marginalize many, including men. In the following, I tackle this question by describing and reflecting on the intellectual journey that was my dissertation research.
Formulating the Research Question
Brooke A. Ackerly, Maria Stern, and Jacqui True (Reference Ackerly, Stern, True, Ackerly, Stern and True2006, 5) argue that what makes scholarship feminist is the “research question and the theoretical methodology and not the tool or particular method used.” Gwendolyn Beetham and Justina Demetraides (Reference Beetham and Demetriades2007, 200) underscore this point by suggesting that many research tools can be used in ways that are consistent with feminist goals and ideology. These perspectives are useful for reflecting on how I came to reconcile my initial reluctance to employ quantitative methods in my own research.
My research questions were based at the systemic level and informed by insights from a lot of case studies and qualitative work produced by feminist IR, economics, and development scholars who provided compelling evidence that economic globalization is a profoundly gendered, racialized, sexualized, and classed process. Because of this, I needed a way to investigate how globalization processes affected states' abilities to meet their human rights obligations and whether or not this had differential effects on the achievement of socioeconomic rights for women and men. I was interested in what global and temporal patterns we could see with regard to globalization (measured as trade openness, foreign direct investment, and economic liberalization), economic development, and democratization. The best way to accomplish this was to use quantitative models, which gave me the ability to approach my research question cross-nationally and longitudinally.
With few exceptions, many of the quantitative models that have been used to examine the relationships among globalization, economic development, democratization, and the achievement of human rights were not informed by feminist IR theory, and I wanted to rectify this problem. I came up with the strategy of contextualizing and informing my study using Dianne Otto's notion (Reference Otto, Merali and Oosterveld2001, 54) of “structural indivisibility,” which “stresses interconnections between the political, economic, environmental, and security priorities of the international order and violations of human rights.” I infused this framework with feminist IR theorizing about the political, economic, environmental, and security priorities that had become “normalized” in mainstream IR in order to inform the hypotheses to be tested, the operationalization of my variables, the models, and the analysis of the results.
Sexing/Gendering Human Rights Data and Models
Scholars and policymakers working in the human rights arena have not just debated how to define human rights and development but also struggled with a way to measure human rights achievement and violations. There are two primary ways that scholars have attempted to operationalize rights—as outcomes and as laws. In my study, the outcomes-based measures were employed, but in so doing a number of issues came to light.Footnote 3
First, there are many provisions in both women's rights and human rights documents that guarantee a wide range of civil and political liberties and socioeconomic rights, but there is actually little data to measure these particular rights. Feminists conceptualize human rights as something far more complex than the equitable distribution of the presumed benefits and resources of economic development, globalization, and democratization. They also see human rights, for example, in terms of individual empowerment and capacity building, which are difficult to quantify.
Second, because socioeconomic rights data is outcomes based and reflects the performance of states, it is actually defined by the public sphere. One of the key insights of feminist IR perspectives on globalization and gender inequality has been its insistence on the interaction of the public and private spheres, and the rejection of this binary as mutually exclusive. Thus, what happens in the public sphere has ramifications for gender ideology and roles in the private sphere. As such, these measures simply cannot capture the gendered dynamics of the private sphere, which have ideological, physical, and material consequences for women with regards to socioeconomic rights.
For example, in my study, infant mortality rates and life expectancy rates were used to capture health and adequate standard of living. These are popular measures because the data have consistently been collected over a long period of time, and they are thought to tap into a multitude of issues related to health and well-being, such as poverty, environmental degradation, and access to resources. They also capture how well government expenditures are spent with regard to securing basic needs. Finally, these data are also sex differentiated, allowing a comparison between females and males with regard to quality of life.
However appealing these data may be, there are also limitations.Footnote 4 Traditional poverty measures are typically based on measures of household consumption and/or income. Therefore, the question becomes how to understand infant mortality and life expectancy rates also as indicators of poverty, rather than only as indicators of good (or bad) government practices (Morris Reference Morris1979). Is good health merely a matter of consumption, and if so, consumption of what? How is this consumption gendered? An additional problem with using infant mortality and life expectancy rates as proxies for poverty is that they have different units of analysis. Infant mortality and life expectancy rates use the individual as the unit of analysis, whereas traditional poverty measures often use the household, conceptualized as unitary and cooperative.Footnote 5 Feminist economists have been instrumental in deconstructing the household to show how distribution of resources within households is often highly contested (and not cooperative) and, as such, affects women and men differently (albeit relationally) (Benería Reference Benería2003; Chant Reference Chant, Jaquette and Summerfield2006). For the purposes of my study, this presents difficulty for understanding how household consumption decision making and resource distribution affect individual well-being and health. Furthermore, this also cannot be captured by the use of sex disaggregated data since these capture comparisons rather than relational dynamics.
Given the oft-cited statistic that 70% of the world's 1.7 billion people in poverty are in fact women and children, and the subsequent coining of the term “the feminization of poverty” due to this statistic (UNDP 1995, iii), perhaps a sex-differentiated individual poverty statistic would have been more ideal for this study than using infant mortality and life expectancy rates as indicators for health and adequate standard of living. In order to calculate this figure cited by the United Nations Development Program, poverty data would need to be disaggregated by sex. However, as far as I can tell, this type of indicator simply does not exist or is unavailable to the public. After having combed World Bank, International Monetary Fund, and UN sources, I discovered that while general poverty statistics, such as income-based measures (living on less than $1 a day)Footnote 6 and the human poverty index (HPI)Footnote 7 are available, there is no sex differentiation of these statistics, and that the figure quoted by the UNDP is an estimate that seems to have taken on the status of “fact” with little population data to support such a finding (World Bank 2001, 63).
This brings me to my final point about the use of this type of output data. Because of the inability to capture gendered interactions of the public and private spheres as well as gendered relations within those spheres, the data are at best capturing sex discrimination within the context of the neoliberal global economy, rather than the structural feature of gender oppression, even though sex discrimination is an important part of maintaining gender oppression.Footnote 8 Many human rights indicators (though not all) use male experience as the norm, and the achievement of women's human rights is seen as relative to the rights that men have already achieved. Thus, the data show that women are discriminated against insofar as they have not achieved the same rights as men, despite the efforts put forth by many feminists to expand and reframe notions of rights that take into account the difference of women and men's lived realities. Sex disaggregated data captures material dimensions of the achievement of rights, but not the overarching ideological dimensions of gender oppression that privileges and normalizes a particular understanding of such rights. This is why I deliberately chose to label the sex variables in my work as “sex” rather than “gender” and discussed sex differential effects rather than gender differential effects.
In this sense, sex is an empirical category and gender is an analytical one, because as Jackson (Reference Jackson2002, 499) notes, sex disaggregation of statistics in itself does not amount to gender analysis or, I would add, feminist analysis. However, as Helen Kinsella (Reference Kinsella2003, 296) argues, to identify sex as an empirical category is deeply problematic because it presumes the naturalization of the biological sex binary and obscures how the categorization of sex (as well as gender) is politically produced and regulated. One way to respond to Kinsella's position would be to reject the use of sex disaggregated data (and associated quantitative models) altogether. Another way to look at this position is that it also allows for interrogation of the process of data construction, collection, and analysis that is not informed by feminist IR (and other feminist) theorizing.
It is this tension raised by feminist IR scholars that enables me to conclude that due to the parameters of such data, my findings do not shed much light on how globalization, economic development, and democratization help shape and sustain gender ideologies that serve to oppress all those identified with the feminine, whether they be women or men. I concur with V. Eudine Barriteau (Reference Barriteau, Kuiper and Barker2006), who argues that much of the sex disaggregated data, gender indexes, and models place too much emphasis on the material relations of power to the exclusion of the social and ideological relations of power. In this sense, there is indeed a formidable disjuncture between IR feminist theorizing and the currently available data. Despite this problem, we can use feminist IR analysis to explain why this is so and to frame interpretation of statistical findings.
Dissemination of Results and Concluding Thoughts
While my dissertation research did, in fact, show that there is a significant difference between the achievement of women's and men's socioeconomic rights in the context of globalization, and that there are significant differential effects of globalization on female and male socioeconomic rights, the results both affirmed and contradicted feminist IR theorizing. This is perhaps where I have faced the biggest challenge in locating myself as a feminist IR scholar. What do I do with results that actually affirm the arguments of pro-globalization and neoliberal economic development advocates? I have been wary of trying to publish the results for fear that they may be used to support arguments that I do not wish to support or may be misinterpreted altogether, even though I am quite clear about the limitations of the data and models in my study and use feminist IR theorizing as a guide to my analysis.
For example, I share the concerns of feminist scholars who worry about what happens when women “catch up” with or exceed men on these indicators that are currently being used to frame public policy at the both the international and domestic levels (Barriteau Reference Barriteau, Kuiper and Barker2006; McIlwaine and Datta Reference McIlwaine and Datta2003). As Diane Elson (2006), notes, we must also be attentive to whether or not men are “equalizing down” to women. Regardless of the direction of the movement toward equality (perhaps better understood as parity in this case) on these indicators, many feminists are worried that this outcome would “create an impression that women no longer require assistance and that men are now much more needy beneficiaries,” and that this would “remasculinize” the human rights and development discourses (McIlwaine and Datta Reference McIlwaine and Datta2003, 375). But perhaps this fear is precisely the point: When quantitative models produce results that are inconsistent with feminist IR theorizing, we need to be able to provide context for, and influence the discourse around, the interpretation of those findings. In this sense, the use of quantitative data and methodologies serve as an entry point rather than as an end point, thereby challenging positivist claims about what “realities” these models depict.
In addition, should I wish to publish the results of my work, the issue of placement has loomed large for me. Given the divide in the feminist IR field about the use of econometric models, it is not immediately obvious where my work would find a home especially since I consider feminist IR scholars to be my audience. Most feminist journals rarely, if ever, publish quantitative work, and should I decide to send my work to one of them, who might the editors find to review it? Would anyone in feminist IR be willing, and if so, would they have the methodological training to be able to assess the work fairly? Would feminist IR scholars consider the work feminist enough? If I send my work to mainstream IR journals, I would not have to worry about the methods; rather, the concern here is over the openness to feminist IR theorizing with regards to the models and data used. Finally, this leads to a further dilemma, given my institutional home of Women's Studies. Should the work be published, I also have to find enough qualified people to be able to assess it for tenure and promotion, which may or may not be feasible, especially in Canada, where IR scholarship, for the most part, has taken a different path than that which is produced in the United States.
Writing this critical reflection piece, however, has persuaded me that I should in fact try to publish the work despite the fears that I have identified. We need to know about patterns in the international system in order to guide deeper qualitative analysis, and we need this deeper qualitative analysis to build better data. By viewing both quantitative and qualitative research tools as crucial to an overall feminist IR methodology, we will, hopefully, be able to bridge and redefine the disjunctures between feminist IR theorizing and quantitatively driven studies of the international system, not just in the field of IR but also in other disciplines, such as economics. It is the contradictions in politics that make what we study interesting, and we should not be afraid to explore them theoretically and methodologically if this can help move us toward feminist goals of eradicating social inequality.
