Abstract
This article evaluates a Participatory Action Research (PAR) approach with mixed methods including concept mapping, q-sorting and deliberative dialogue in the context of a research project on young people’s experiences with digital communications technologies, and addresses some of the central insights of intersectionality theory and praxis. Our approach seeks to ensure that, insofar as possible, the gathered data provide a rich and layered window into the experiences of young people from a range of marginalized communities served by our project partners. The article revisits some key insights and contestations relating to intersectionality and addresses their relationship to our approach. We evaluate whether these methods enhance understandings of the interactions of structures of subordination with other factors identified in intersectionality scholarship, as well as the extent to which they centre the knowledge and expertise of those subordinated by matrices of domination as discussed by authors such as Crenshaw and Hill Collins. Our approach is just one of many that social science researchers interested in advancing intersectionality’s key insights could deploy. While it falls short of full consistency with these insights, its mixed methods work toward our partners’ social justice objectives while facilitating exploration of intersecting axes of subordination. Our approach can also help our project recapture the politic at the heart of many intersectional feminist critiques, such as those of Crenshaw and Hill Collins - that reconceptualizing knowledge requires centring the knowledge and expertise of those traditionally excluded due to interlocking systems of subordination.
Keywords
Black and other feminist of color scholarship, activism, and organizing have a long global history of discourses illuminating the multiple ways in which “oppression, discrimination and violence…structure conditions in which women of color live” (Carastathis, 2016, p. 16). In the late 1980s, Crenshaw (1989, 1991) used the term “intersectionality” to illustrate how, despite the ongoing interventions of feminist and anti-racism activism, law reform efforts continued to marginalize the experiences of Black women in the United States. Crenshaw argued that single-axis considerations of race or gender or double-axis approaches that simply added race to gender (Harris, 1990) provided a distorted understanding of Black women’s lives precisely because they failed to consider the inseparability of race and gender in Black women’s experiences, including the many ways that the interactions between gender and race shape Black women’s realities. At around the same time, Hill Collins (1990, p. 225) described the complex interaction of multiple factors and contexts—or “interlocking systems of oppression”—that worked to marginalize Black women in a “matrix of domination.”
Since that time, scholars from a wide variety of disciplines, including education, health, history, sociology, feminist studies, ethnic studies, and queer studies, have engaged with, expanded upon, and debated the parameters and merits of intersectionality (Cho, Crenshaw, & McCall, 2013, p. 785; Dhamoon & Hankivsky, 2011). The incorporation of intersectionality into theory and praxis has led to important developments with respect to health, economic, and legal justice reforms for members of marginalized communities (Cho et al., 2013, p. 786; Hankivsky, 2011). Nonetheless, its incorporation into social science methodology has been particularly challenging and controversial (Chadwick, 2017).
In this article, we review some 1 intersectionality scholarship, with a view to evaluating the potential of using participatory action research (PAR) with mixed methods to investigate the entanglement of intersecting and interlocking systems of oppression and how they shape lived experience and knowledge. We situate our investigation in the context of an ongoing research project that focuses on experiences of youth from marginalized communities with digital communications technologies. Our project includes a number of partner organizations 2 serving people from a diverse range of marginalized communities, with various objectives relating to policy-making and production of educational materials. For the purposes of our project, we propose to use a PAR approach that includes concept mapping, q-sorting, and deliberative dialogues. Adopting a mixed method approach is consistent with intersectionality literature in the field of health research, where Dhamoon and Hankivsky (2011, p. 44) favor experimentation with “a range of existing methodologies.” We explore the degree to which research conducted using this approach, when guided by intersectional feminist principles, has the potential to provide a nuanced reflection of how systems of oppression interact and to enrich our understanding of the relationship between theory and praxis (Hill Collins & Bilge, 2016, p. 192). At the same time, we recognize that the complexity of intersectionality makes it difficult for any single approach (even one deploying multiple methods) to fully address all of the complications and contradictions between individuals and within any social group. Recognizing this challenge as well as the diversity of objectives of the wealth of community, institutional, governmental, and academic partners within our project, we discuss our proposed methodology as one possible approach for extending and/or supplementing traditionally siloed qualitative and quantitative methods to better take some of these complexities into account. We reflect critically upon the difficulties we encountered in making the various methods we propose fit with the overarching objectives of intersectionality by assessing the advantages and disadvantages of each method. Ultimately, we conclude that our proposed approach responds to the social justice objectives of our partners for more inclusive evidence-based educational materials and policies, while at least partially addressing the politic at the heart of many intersectional feminist critiques, that is, that centering the knowledge of those traditionally excluded is essential to better understanding the processes by which domination and subordination occur and thus to achieving a more just society (Crenshaw, 1989; Hill Collins, 1990; May, 2015, p. 53).
Situating Ourselves and Our Project
We are the principal investigator and five coinvestigators on The eQuality Project, a 7-year partnership initiative funded by the Social Sciences and Humanities Research Council of Canada. Our project focuses on young people’s experiences with privacy and equality 3 in networked spaces. In particular, we focus on the impacts of commercial online behavioral targeting and tech-facilitated violence on the experiences of young people aged 15–22 from communities marginalized due to transphobia, homophobia, racism, colonialism, sexism, classism, and intersections thereof. Our partnership is founded on collaboration between researchers, community organizations, educators, research and policy institutes, and young people who serve these communities in various ways. It has multiple aims that reflect the diverse constituencies our partner organizations serve and the particular objectives of those constituencies in relation to online behavioral targeting and tech-facilitated violence. We aim to create new knowledge about commercial data practices and their impact on young people, as well as how young people from marginalized communities conceptualize privacy and equality. In addition to contributing to digital policy-making, our partnership aims to develop school-based and public educational materials designed to fill knowledge gaps identified by our partner organizations.
Our partnership is committed to conducting research and to developing and disseminating knowledge and educational material using methods that reflect and are grounded in the political and equality-affirming ethics set out in the rich body of intersectional theory, praxis, and research that has gone before us. Honoring our joint social justice commitment to intersectionality is unlikely to be achieved through selection of a single method given the diversity in our partner organizations, the constituent communities they serve, and their objectives. For example, it seems unlikely that the policy-making goals of some of our partners can be fully achieved through purely qualitative methods, while qualitative methods are likely to be particularly useful in developing educational materials that showcase the diverse ways in which intersecting axes of subordination operate in the lives of young people from marginalized communities. As a result, rather than opting for a purely qualitative or a purely quantitative approach, we propose to use PAR with mixed methods. While we recognize that this approach offers both advantages and disadvantages for achieving the key insights of intersectionality, we are both conscious of the importance of attending to the needs and objectives of the diverse communities our partners serve and hopeful that our approach offers some prospect for getting at the social justice objectives underpinning intersectionality theory and praxis.
We also recognize that our social positions as socioeconomically privileged individuals who self-identify as women who are differentially affected by intersecting axes of subordination and privilege relating to race, age, sexual orientation, and ability undoubtedly affect our understanding of and approach to intersectionality. These factors will raise issues of imbalances in power and privilege throughout the research process that will have to be addressed. As a result, it will be necessary for us to take steps to minimize these imbalances and their potentially negative effects as much as possible, such as engaging in reflexive awareness exercises throughout the research process, as suggested by Raheim et al. (2016), working to form relationships of trust with communities in which research participants reside and retaining in-community researchers to conduct data collection exercises in situations where the power imbalance is particularly pronounced and, in any event, whenever the communities with whom we are working prefer it (Marin et al., 2015). In our project, we will also rely heavily on our partners for direction and advice since those working in community will have excellent on-the-ground knowledge about the expectations and areas of concern related to power and privilege that are held by the members of the communities that they serve.
Intersectionality
Recognizing the “tremendous heterogeneity that currently characterizes how people understand and use intersectionality” (p. 2), Hill Collins and Bilge (2016) describe it as follows: Intersectionality is a way of understanding and analyzing the complexity in the world, in people, and in human experiences. The events and conditions of social and political life and the self can seldom be understood as shaped by one factor. They are generally shaped by many factors in diverse and mutually influencing ways. When it comes to social inequality, people’s lives and the organization of power in a given society are better understood as being shaped not by a single axis of social division, be it race or gender or class, but by many axes that work together and influence each other. Intersectionality as an analytic tool gives people better access to the complexity of the world and of themselves. (p. 2) social-movement discourses that identified the manifold manifestations of oppression, discrimination and violence that structure conditions in which women of color live in the United States, Britain and other white settler and imperial states. (p. 16)
Hill Collins (1990, p. 225) posited that categories like race, gender, and class should be thought of as systems of oppression that interlock within a matrix of domination. She argued that placing those marginalized by the matrix at the center of analysis “opens up possibilities for a both/and conceptual stance, one in which all groups possess varying amounts of penalty and privilege in one historically created system.” Taking a both/and approach (e.g., examining the co-constituting effects of race and gender), situated in relevant historical and social context, also assists in understanding how communities of resistance are formed (Hill Collins, 1990).
Since the experiences of all people are influenced by multiple systems within the matrix of domination (May, 2015, p. 6), the study of these systems and their intersections will not necessarily produce or be aimed at producing social justice outcomes (Carastathis, 2016). Working to insure a link between intersectionality and social justice outcomes requires those engaging with intersectionality to focus on its anti-oppression origins (May, 2015, p. 229). Hill Collins and Bilge (2016) suggest that six core ideas of intersectional frameworks are present in critical inquiry and praxis: social inequality should be understood as a product of interactions between categories rather than of a single factor (p. 26); power should be understood as mutually constructed by many diverse factors and power relations (e.g., racism and sexism) analyzed by their intersections and across domains of structural, disciplinary, cultural and interpersonal power (p. 27); analysis should be focused on relationality, that is, the interconnections between entities such as race and gender, rather than the differences between them (p. 27); an awareness of social context is needed to take into account the ways in which “historical, intellectual, and political contexts shape what we think and do” (p. 28); complexity should be recognized as arising from the interactions of social inequality, power, relationality and social context, thus complicating use of intersectional analysis (p. 29); and social justice should be understood as mandating critical analysis of the status quo and looking beyond rules stated to be superficially equally applicable to all, in order to see whether they are practiced and/or applied in unequal ways (pp. 29–30).
Centering the knowledge and expertise of “disremembered subjects” who have traditionally been marginalized (May, 2015, p. 53, quoting Foreman, 2013, p. 316) can also strengthen the connection between intersectionality and critical social science inquiry. Members of marginalized communities not only have lived expertise in identifying the negative outcomes of interactions between socially constructed categories and context, they also have lived expertise in resisting them. Ignorance and discounting of marginalized community members’ knowledge and agency have led to exclusionary policy and other negative social justice outcomes (Dhamoon & Hankivsky, 2011).
Applying Intersectionality to Quantitative and Qualitative Research: Strengths and Weaknesses
Since the 1990s, the intersections between a myriad of socially constructed categories and systems of domination have been addressed in a wide variety of social science disciplines using both quantitative and qualitative methods (Cho et al., 2013). However, the transition from social justice advocacy to social science methods has been anything but smooth.
Notwithstanding the compelling and insightful arguments advanced in intersectional theory, as Hankivsky et al. (2014) note, conducting research or developing policy that seeks to advance its insights can be “intimidating” at least in part due to a need for “explicit and user-friendly methods that can effectively translate intersectional theory into practical approaches that can be understood by decision-makers and policy researchers” (Hankivsky et al., 2014, p. 2). This situation is beginning to change in a number of areas including in the health policy field where Hankivsky et al. (2014) have proposed guiding principles and an Intersectionality-Based Policy Analysis tool. These guiding principles provide helpful signposts for intersectional research. Ultimately, however, selecting a method for data collection intended to be used in policy development in a diverse range of fields involving a diverse range of communities and intersecting axes of subordination demands careful tailoring in each case. As a result, no single social science method or group of methods is likely to prove to be the approach for every situation.
Certainly, a number of scholars have argued that there are both strengths and weaknesses in quantitative and qualitative approaches when it comes to addressing the key messages of intersectionality theory and praxis. As McCall (2005) notes, quantitative methods have been used to help advance our understanding of intersectionality precisely because they enable researchers to explore individual and interactive effects of multiple factors, such as race and gender, across a population. By focusing analysis on the relationships between groups who are advantaged and disadvantaged by these factors, we can examine how they interact to produce different outcomes. For example, multivariate analysis enables us to examine not only the effect of race on income but also how the effect differs for men and women (McCall, 2005, pp. 1786, 1788). Accordingly, quantitative methods allow us to mobilize socially constructed categories like race and gender strategically, in order to examine relationships of inequality between social groups and to examine changing configurations of inequality along multiple dimensions (Dhamoon & Hankivsky, 2011, p. 44; McCall, 2005, pp. 1773, 1774). However, as Bowleg (2008, p. 322) explains, strategic deployment of categories requires careful fashioning of questions and data analysis with a clear understanding that “variables such as race, sex, sexual orientation, class, and disability” are not explanatory constructs in and of themselves.
A full commitment to intersectionality significantly complicates quantitative analysis because intersectionality calls upon us to study multiple groups along multiple axes (McCall, 2005, p. 1790), and the more categories we seek to include, the more challenging it is to measure individual and interactive effects. This may lead to falling back on a limited set of predetermined categories (e.g., gender and race but not gender, race, and ability), potentially losing the elasticity we need to fully understand how systems of domination interact to constitute one another and to socially locate individuals. Moreover, where quantitative approaches seek to disentangle categories in order to address their impacts and interactions, they fail to appreciate one of intersectionality’s central messages that race, gender, and class oppressions create a synthetic experience that can’t be understood by decompositional analysis of the interplay between discrete factors (Fernandes, 1997; Hancock, 2013).
Quantitative analysis often involves comparing the experience of groups of interest against a “norm” in order to measure whether or not they have been treated differently. This comparison can reiterate and recenter the experiences of dominant communities, thereby undermining the political underpinnings of Crenshaw’s initial conceptualization of intersectionality as a method for centering the experiences of the marginalized (Alexander-Floyd, 2012, p. 12).
In contrast, qualitative methods interrogate the boundary-making process, often focusing on the experiences of those detrimentally affected by the interactions of multiple factors (e.g., poor women of color; Choo & Ferree, 2010, p. 129; McCall, 2005, p. 1773). As such, these methods center on the narratives of those subordinated by these interactions to explore the richness of their lived experience. The data collected make “experiences visible in complex ways” (Alexander-Floyd, 2012, p. 20) and can be used to extrapolate about the broader social location that each individual embodies (McCall, 2005, p. 1781), as well as the ways in which underlying power structures shape identities and opportunities (Alexander-Floyd, 2012, pp. 4, 5; Morris, 2006, p. 4).
Importantly, qualitative methods place the perspectives of multiple marginalized groups at the center of research (Choo & Ferree, 2010, p. 133) but are limited in their capacity to advocate for collective goals in the policy arena. Despite the richness of qualitative research in illustrating lived realities, the results are not generalizable, and generalizability is often considered a prerequisite for evidence-based policy-making. Moreover, like quantitative studies, qualitative investigations often rely on comparisons to the usual (dominant) groups to show how the experiences compare or contrast. Not only does this recenter the usual groups as discussed above (Alexander-Floyd, 2012, p. 9), it also can devolve into “us” and “them” language, especially in reports aimed at mainstream audiences (Choo & Ferree, 2010, p. 138). Once again, the danger is that we compromise one of intersectionality’s key political objectives—reinvigorating collective political action that promotes the perspectives and needs of the most subordinated.
Intersectionality introduces complexity into research in at least three ways. First, intersectionality expands the subject of analysis to include multiple dimensions of social life and categories of analysis rendering it more challenging to achieve intelligibility (McCall, 2005, p. 1774). Second, research that aims solely to ensure that the voices of those traditionally excluded are heard does not necessarily address the process by which categories are produced, experienced, reproduced, and resisted in daily life (Choo & Ferree, 2010, p. 133; McCall, 2005, p. 1783). Third, because intersectionality can be understood as a method for creating new knowledge and new democratic institutions, identities, and practices (Chun, Lipsitz, & Shin, 2013, p. 924), concerns can arise around methods that may examine intersections but fail to address the fundamental commitment to social justice and equality articulated in Crenshaw’s and Hill Collins’ articulation of the concept (Alexander-Floyd, 2012).
Some scholars have suggested categories should be eschewed completely because they are socially constructed fictions (Butler, 1993). Indeed, intersectionality can be used to buttress this position because it challenges the completeness of all categorical groups (e.g., the gender binary) and raises the concern that no number of categories is enough. However, this approach flies in the face of the basic premise underlying Black women and women of color feminists’ insight that intersectionality is not a tool to reinforce simplistic socially constructed categories. Instead, it is a method to expose social and political processes of subordination in order to form a basis for developing justice-seeking group politics (Crenshaw, 1991, pp. 1296, 1298; Hill Collins, 1990).
We turn now to discuss the strengths and weaknesses of PAR with mixed methods as one of many approaches for both serving the diverse objectives of our project partners and addressing two key insights of intersectionality as an analytic tool: (1) the importance of starting with the perspectives and experiences of those whose views and knowledge are often marginalized or neglected (Morris, 2006, p. 2) and (2) the need to move beyond “business as usual” mainstream research approaches that describe categories of people (e.g., low income women) toward approaches that both critically explore the structural causes of inequality and enable the liberatory potential of intersectionally informed group politics (Alexander-Floyd, 2012, p. 4; Morris, 2006, pp. 2, 4). In particular, we suggest that concept mapping and q-sorting assist in exploring structural categories that can contribute to inequality. In addition, PAR (especially in the form of deliberative dialogue in which traditionally excluded voices are centered on group-identified policy issues of concern) is well positioned to allow for articulation of group needs and aspirations related to policy-making and political action, as well as to support community members in gaining access to information about and developing skills related to policy and political processes.
Moving Forward: Methods That Center Intersectional Subjects in the Research Process
PAR as a Guiding Philosophy
At the heart of much intersectional feminist critique is a call to center the perspectives and experiences of those subordinated by matrices of domination, in order to develop policies that take the needs and aspirations of these communities fully into account and seek to dismantle the power structures that contribute to their subordination. PAR is useful in this regard precisely because it: is not so much a set of procedures to follow to gather information as it is a philosophy and approach to gathering and using information. It is also a tool for building and strengthening communities and our understanding of ourselves, each other, and our relationships. It can also be a powerful outreach, base building and organizing tool to help bring people together to build movements for change. (Incite!, 2014)
As in any community, there are power dynamics and differential relationships within subordinated communities that will need to be acknowledged and taken into account during PAR research. Without such efforts, there is a risk that certain members of the community may feel disenfranchised or excluded from the project design. Simply inviting an assortment of community actors to participate in a study will not be sufficient to ensure the intersectional objectives of the project will be met. Instead, in keeping with PAR principles, researchers must recognize and consider how they will negotiate inevitable differences between and among group members.
PAR practiced without a feminist, critical race, 5 or intersectional perspective will not necessarily achieve the goals of intersectionality. As such, social justice principles must be consciously incorporated into the PAR framework in order to reach the social justice goals imbedded in an intersectional approach. Much PAR research is concerned with creation of social justice outcomes. Further, PAR has been specifically adopted as a mechanism for addressing intersectionality, yielding results in one example that not only generated new knowledge about the lived experiences of Asian women garment workers in California but also instigated new community-led democratic organizations, social justice policy reform, and positive shifts in intrafamilial relationships for the women involved (Chun et al., 2013).
Adopting PAR as an overarching principle for intersectional research helps embody intersectional praxis by centering the knowledge of those subordinated by systems of domination to “see how power works,” allowing for coextensive creation and elaboration on the concept itself (Chun et al., 2013, p. 918). As Chun, Lipsitz, and Shin (2013) note, it is an ideal approach for understanding how intersections of multiple oppressions shape participants’ experiences, leading to original and grounded insights, evidence, ideas, and ambitions for responding to social problems (p. 919). In recognizing and building upon this lived expertise, we can advance the goals of intersectionality by revealing the way that systems of domination set in motion processes that link people together, revealing common problems that require redress (p. 923). In this way, PAR can help contribute to development of the sort of collective political action envisioned in intersectionality theory.
By taking this approach, participatory methods build alliances around common beliefs and experiences, while also recognizing difference within the group, rather than relying on presumptions that the members of any particular socially constructed group share a common identity or perspective (Chun et al., 2013, p. 937). Further, participatory methods can challenge myths and stereotypes about marginalized groups (Chun et al., 2013, p. 935) by emphasizing members’ knowledge, strength and resiliency, and countering the mischaracterization of subordinated groups as weak, powerless, or politically apathetic. Finally, by shaping the agenda around the “questions and issues that are significant for those who participate as co-researchers” (Chun et al., 2013, p. 930), participatory methods create opportunities for skill and capacity building in areas that have meaning for marginalized community members themselves (Cammarota & Fine, 2008). We recognize that by preidentifying the methods to be adopted, our proposed approach falls short of fully incorporating participants as coresearchers. However, we hope that this shortcoming is attenuated to some extent by the fact that the methods we have chosen are informed by the needs identified by community partners who directly serve the communities from which our participants will be drawn.
PAR objectives involving youth “extend from the exposition of local inequities with contextual specificity, to broader coalition building with similarly situated youth nationally and globally” (The Public Science Project [PSP], n.d. a; PSP, n.d. b). One example of a PAR project related to media, technology, and youth illustrates this research ethos. Castellanos, Bach, and Kulick (2010), in their report on the collaborative creation of a youth-led community needs assessment for a youth media division of a public access media center in New York, differentiate PAR from other research methods: Unlike conventional research in which the authority of the researcher/academic is often overemphasized—thus reproducing already existing expert systems, preserving the elite status of universities, and reinforcing power imbalances between universities and communities…[this study] challenges this model by shifting and expanding how research is conceptualized, conducted, and applied…the project assumes a youth-centered, participatory, feminist epistemology in which young people’s everyday experiences shape the research design and selected methodologies. (p. 161)
While potentially uncomfortable, this tension is precisely why PAR can be so helpful to intersectional research. It can work to destabilize the relationships of power that are at the heart of research and to provide a forum for marginalized communities to articulate their own needs and perspectives. Being directly confronted by community members’ articulations of their own needs and perspectives affords intersectional researchers the important opportunity to gauge how honestly their methods center participants in the knowledge creation process. In the context of a youth-focused PAR project, the creation of a youth advisory committee with decision-making input is another step that can facilitate that gauging process.
Understanding Experiences Through Concept Mapping and Q-Sorting
Concept mapping (Trochim, 1989) and q-sorting (Brown, 1980; Stephenson, 1953) can also be helpful to a social scientist interested in intersectionality for two reasons. First, each method is designed to deepen our understanding of lived experiences, thus responding to the call of many intersectionality scholars for the voices of those subordinated by multiple systems of domination to be heard. Second, research participants drive the outcome of the research process each step of the way, reducing the risk that the knowledge they cocreate with researchers will be interpreted through a dominant lens.
As mixed methods, concept mapping and q-sorting use traditional quantitative methods (cluster analysis in the case of concept mapping, modified factor analysis in the case of q-sorting) to explore the perspectives of individual research participants. These methods are effective for examining conceptual spaces that are complex and admit multiple characterizations or are understood from a variety of lived perspectives. Thus, for example, van der Meulen and colleagues (van der Meulen & Glasbeek, 2013; Wright, Glasbeek, & van der Meulen, 2015; Wright, Heynen, & van der Meulen, 2015) used concept mapping techniques to explore conceptualization of closed-circuit camera surveillance in order to compare how the various aspects of/reactions to this surveillance identified through the concept mapping exercise (e.g., “invasive and intrusive,” “assistive and supportive”) were rated across different groups of women (e.g., sex workers, racialized, and older women; club-goers and low-income women). Q-sorting has been used to examine privacy perspectives among younger (Morton & Sasse, 2014) and older (Morton, 2014) adults, identifying a variety of different positions on informational privacy issues. Although no research has explicitly contrasted the results of the two methods, the two methods have been used to gain insight into a common issue, specifically conceptions of feminism (Linton, 1989: concept mapping; Snelling, 1999: q-sorting).
Concept mapping and q-sorting offer different, and potentially complementary, approaches to the study of diverse perspectives. Both methods start with a set of statements that reflect the focus of inquiry—the concept or issue under study. These statements can come from a wide range of sources: They can be generated in consultation with participants, drawn from prior interaction with participants (e.g., open-ended survey questions, interviews, or focus group discussions on the issue), selected from press coverage or other published material relevant to the issue, extracted from prior research, and/or developed from theoretical considerations. Often, a number of different approaches are used to construct the statement set or concourse (a term used in q-sorting). The goal for both methods is to start with a set of statements that reflects the entirety of the concept of interest. A second consideration is the selection of subjects (or p-set in q-sorting terminology). Both methods typically use purposive sampling techniques to select participants. The goal in sample selection is not formal representativeness. Instead, a broad group of participants is selected on the basis of the insights or perspectives that they bring, and the aim of the sampling endeavor is to ensure that a full range of relevant perspectives is represented in the sample. Sample sizes in q-sorting and concept mapping studies are relatively small (usually under 40 participants) in contrast to the large sample sizes required by traditional quantitative methods.
The goal of concept mapping is to identify a common understanding or organization of a concept of interest—one that reflects and collectively organizes the perspectives of all participants, while exclusively privileging none. Assuming careful identification and selection of both statements and participants to be inclusive of the full range of perspectives and positions, with sensitivity to and inclusion of those occupying multiply marginalized positions, a “concept map” can provide the basis for a collaborative dialogue or can serve as a basis for the comparison or contrast of attitudes or experiences across different intersectional positions. In our research, we will use this method to understand how young people think about their “digital shadow” and the range of information that contributes to this profile. The results will organize the multiple pieces of information that go into the digital shadow into the categories that best reflect the thinking of all participants; the method will also allow us to contrast thinking about these categories of personal information across various categories (e.g., racialized vs. nonracialized youth). The method explores the experience of individuals affected by various intersectional positions, without being rooted in an assumption of difference or contrast between various groups or categories. The method allows and even encourages dialogue across intersectional positions that focuses on similarity as well as difference, thus resisting the reification of preexisting categories.
By contrast, the goal of q-sorting is to identify the different subjective positions expressed by a range of participants with respect to the issue of interest. In the context of our research, this might take the form of an exploration of the centrality of various aspects of the digital shadow to the person or self. The results of the q-sorting analysis identify the different subjective perspectives expressed across the group of participants. In our case, these different perspectives could include one that identifies preferences or tastes as most central to the self and another that identifies social contacts, family, and friends as most central. The analysis examines together the opinions of all participants (each expressed in an individual “sort,” described below in more detail) to identify systematically different perspectives. Those holding different perspectives may occupy different intersectional positions—but the analysis does not presume these differences and indeed proceeds without information about the individuals whose attitudes are reflected in each sort. Similar to concept mapping, therefore, q-sorting can reveal but does not presuppose categorical differences, and in this way assists us to understand various intersectional positions and to resist any temptation to set up an a priori contrast or focus on a presumed difference.
In concept mapping, participants are asked to sort the statements into categories or “piles”; at the same time, they are asked to rate each item according to specific characteristics. Little other instruction is provided, save that there must be more than one pile, and at least some piles with more than one item, and participants are encouraged to sort according to whatever criteria make sense to them. Thus, the piles or categories produced by each participant represent their individual organization of the topic under consideration. The data for each subject consist of a record of which items are “sorted” together, and data from all subjects are combined using multidimensional scaling and clustering techniques to identify a common “map” that best reflects the statement organization of all participants. Interpretation of the data involves naming of the clusters and analysis of their relative placement. In addition, rating scale averages can be computed for each cluster and compared across subgroups of participants (e.g., on the basis of gender, sexual orientation, racialization).
Concept mapping, therefore, extracts an understanding or description of a concept that all participants can (to some degree) relate to and reflect upon. In doing so, the method necessarily de-emphasizes conceptual disagreements or qualitatively different perspectives, creating the risk that ideas from participants from previously ignored communities will continue to be obscured. For this reason, it will be important to supplement this mixed method with the use of other critical qualitative methods that better ensure that the experiences of the most marginalized are more fully articulated. At the same time, the method allows for explicit comparison of cluster ratings across identified participant groups. As a tool, concept mapping can serve as a foundation for dialogue across or between individuals affected by different intersections in the matrix of domination; this is particularly true when item-rating data are used to compare cluster ratings across groups. This method, therefore, is valuable both to explore the conceptual understanding held by those affected by a specific intersection of oppressions (e.g., masculinity as experienced and conceptualized by Black men; Hunter & Davis, 1994) and to examine the nuanced differences between individuals who are differently located in the matrix of domination (e.g., differences in the importance of various community supports for sexual minority youth; Davis, Saltzburg, & Locke, 2009).
In q-sorting, participants carry out a structured sorting task, organizing the statements (concourse) in relation to a characteristic identified in a “focus question.” A typical q-sorting study might invite participants to sort statements about cyberbullying, for example, according to how much each statement reflects their own particular understanding of the concept. Participants are instructed to sort the items into a prespecified and roughly normal distribution, ranging from those statements that in their opinion least typify the focus question to those that most typify the question. The resulting sort is an expression, for each participant, of an individual perspective. A modification of factor analysis (Brown, 1980; Stephenson, 1953) is applied to the collection of q-sorts to identify characteristically different sorts that reflect the perspective of more than one participant. Thus, the analysis correlates participants with each other in order to articulate identifiably different positions within the context of the sample studied and the items sorted.
Q-sorting is oriented toward identification of a range of positions on the issue or concept under study and thus “is particularly useful in fields where questions of difference are important” (Senn, 1996, p. 203). These positions emerge as “factors” in the analysis. Each factor is “defined” by two or more participants, and the factor can be described in terms of the items that these participants rate particularly high or particularly low in the sorting task. The different positions are identified through the analytic process, and while it is possible to inquire post hoc whether identifiable social locations (e.g., the position of a person affected by the intersection of categories of race and gender) are related to particular ways of sorting the items, the method neither requires nor allows presupposition of different perspectives tied to specific social locations. As with concept mapping, this technique can be used to identify and explore different perspectives within an intersectional location (e.g., different experiences of sexual identity formation among gay men; Fassinger & Miller, 1997) or to identify a range of perspectives expressed by participants occupying different social locations (e.g., the study by Barata, 2007, of the perspective of domestic abuse survivors on the criminal justice system response, including both white and racialized participants).
Q-sorting and concept mapping have been used in conjunction with other participative methods (e.g., photovoice; Haque & Rosas, 2010; performance art, Maxwell, 1999) to explore and represent participant experiences and perspectives. Both methods lie outside the frames typically drawn around “qualitative” and “quantitative” methods, and this position represents both a weakness to be addressed and a strength that can be capitalized upon. The weakness lies in the fact that these methods can be difficult to describe since their assumptions and practices are unfamiliar to researchers whose work fits securely within existing qualitative or quantitative methods. The strength is that through innovative application of techniques “at the intersection” of qualitative and quantitative methods, these techniques can provide unique insight into the complex lived experience of those subordinated by multiple systems of oppression (e.g., sexism, racism, and homophobia).
These two methods are examples of what Stenner and Rogers (2004) have termed “qualiquantological” approaches to research. Feminist scholars have identified these methods as valuable research tools (Campbell & Wasco, 2000). Campbell and Salem (1999), for example, identify concept mapping as a “useful tool for capturing women’s voices” (p. 66), and Kitzinger (1986), in the context of an exploration of accounts of lesbianism, discussed q-sorting as an approach to “describe the different ways that people construct their realities” (p. 152).
The key is that both q-sorting and concept mapping place participants, and participant perspectives, at the center of the research process. Moreover, these methods are consistent with a PAR philosophy because they allow for participant involvement at all stages of the research process from the development of research materials through the stages of data collection, data analysis, and data interpretation—and necessarily place participants, and their perspectives, at the center of the data collection process, consistent with the aspiration of critical feminist theories of intersectionality. This can help to “equalize the power relations between scientists and those they study” (Riger, 1999, p. 92; see also Billard, 1999), thus enhancing the likelihood that results reflect the perspectives of the participants rather than those of the researchers. Q-sorting and concept mapping can be used as methods of data collection and interpretation in several research contexts, including PAR (Billard, 1999), to provide specific insight into various lived experiences. The results of q-sorting and/or concept mapping exercises can inform theoretical and policy dialogue.
Deliberative Dialogue
Like concept mapping and q-sorting, deliberative dialogue is a method that reverses the traditional direction of research, where the researchers define the categories of interest, and instead creates a way for participants to “better acknowledge and ground the differences among [them] and negotiate the means by which these differences will find expression in constructing group politics” (Crenshaw, 1991, p. 1299). 6 It is particularly useful in meeting Crenshaw’s goals because it brings people from marginalized communities together to engage in an overtly political process that defines issues and solutions from their perspectives. In addition to better ensuring that voices that are often neglected or ignored are actually heard, deliberative dialogue addresses a second caution from intersectionality: the need to move beyond “business as usual” mainstream research approaches that purport to describe categories of people (e.g., people with disabilities) toward methods that both critically explore the structural causes of inequality and enable the liberatory potential of intersectionally informed group politics (Alexander-Floyd, 2012, p. 4; Morris, 2006, pp. 2, 4).
Deliberative dialogue is accordingly a useful PAR method because it allows participants to explore through deep discussion their perspectives and needs, as well as solutions that respond to those needs. Participants can draw on their own rich, lived experiences with intersecting vectors in the matrix of domination without either reifying identity categories or comparing some identity experiences to an ideal or dominant norm. Moreover, deliberative dialogue is a policy-oriented method to unpack power structures as participants themselves describe the meaning of those structures based on their own experiences. Deliberative dialogue practices also involve creating space to acknowledge the differences among members and the power structures within the group. This method can therefore embrace the social justice–seeking objectives underlying intersectionality by bringing together equality-seeking groups to define the policy problem and articulate policy responses that meet their needs and aspirations, while at the same time recognizing that contradictions and conflicts within the group may not always lead to consensus on what those policy responses should be (Young, 2002).
Escobar (2014) conceptualizes deliberative dialogue as a community of inquiry, as a mechanism for reinvigorating public discourse by bringing together an “interpretive community” of experts and citizens to create “a persuasive understanding of the issues under investigation…developing arenas and forums in which knowledge can be debated and interpreted in relation to relevant policy issues” (Fischer, 2003, p. 13, cited in Escobar, 2014, p. 483). Deliberative dialogue takes place in a public setting designed to be a “safe space to engage in meaningful dialogue” about participants’ beliefs and underlying reasons for those beliefs. Those beliefs can be subject to challenge in a context crafted to encourage “relationship building and deep understanding,” rather than the “advocacy dynamics” that typically involve “prepackaged messages” and “well-rehearsed monologues” (Escobar, 2014, p. 486).
Deliberative dialogue builds on Habermas’ theory of communicative action. It incorporates insights about humans’ intrinsic ability to develop consensus/understanding through language when they are engaged in communicative processes unhindered by strategic goals and placed in settings structured to ensure equal opportunity to speak, challenge others, and to defend their own claims (Pfeil, 2013, p. 124). Critical to the exercise is ensuring that expert opinion is not permitted to dominate in order to create a “diverse and multi-disciplinary community of inquiry…that disrupts the stereotypical expert-led consensus and engagement dynamics of some [public engagement] processes” (Yankelovitch, 1991, p. 485). Through this dynamic, those traditionally thought of as experts learn from the expertise of those subordinated by systems of domination, to ensure that traditionally neglected voices are both fully included and heard within public discourse, and to maximize the opportunity for skills sharing and development among members of equality-seeking communities.
It cannot be assumed that those whose voices are intended to be centered will necessarily feel safe to speak in the process. Instead, steps must be taken to balance any power dynamics between groups of participants by insuring that participants from subordinated groups are provided with the support and resources they require in order to fully and meaningfully participate. 7 Although there is no single approach (Lavis, Boyko, & Gauvin, 2014), various researchers have sought to create open spaces for dialogue that share common characteristics. First, deliberative dialogue focuses on a problem that catalyzes the community to undertake an inquiry (Moses, Saenz, & Farley, 2015; Pfeil, 2013; Shields, 2003, p. 511, cited in Escobar, 2014, p. 483). Second, a skilled facilitator who is not a stakeholder (Lavis et al., 2014; Li, 2015) facilitates discussion in a way that ensures equal opportunities to express views and perspectives (Pfeil, 2013). Third, ground rules are set for deliberation that encourage equal participation by all by requiring respectful, compassionate, active listening that encourages deeper understanding (Lavis et al., 2014; Moses et al., 2015, p. 197; Pfeil, 2013, p. 126). Fourth, participants from a spectrum of social locations are recruited to ensure a “variety of perspectives [and] open dialogue” (Lavis et al., 2014, p. 1291; Li, 2015, p. 35; Moses et al., 2015; Pfeil, 2013). Fifth, informed discussion is facilitated by circulation of plain language background information to engender equal awareness of the issue among participants beforehand, with experts on hand to answer questions on request (Escobar, 2014, p. 482; Lavis et al., 2014; Li, 2015, p. 28). Sixth, discussion is facilitated to encourage sensemaking, “educative dialogue, reason giving and public information” (Moses et al., 2015, p. 197) rather than to create consensus or compel collective decision-making in order to challenge expert frames (Escobar, 2014; Lavis et al., 2014; Moses et al., 2015; Pfeil, 2013). Seventh, relevant community-based practices for creating safe spaces for discussion can be adopted (Powell, 2017).
Incorporation of deliberative dialogue should support our project in addressing methodological concerns raised by intersectionality in four ways. First, intersectionality theory posits that the existence of categories is not necessarily the problem. Instead, the key problem is “the particular values attached to [categories] and the way those values foster and create social hierarchies” (Crenshaw, 1991, p. 1296). Deliberative dialogue is oriented toward equalization of power and resources, to listen to people who are traditionally not listened to in policy-making processes, and to put “the experiences and perspectives of people with the least social, economic and political power front and centre throughout the research process” (Morris, 2006, p. 23). Further, deliberative dialogue is not premised on reifying socially constructed categories. Instead, the primary task of organizers is to identify relevant stakeholders by gathering those who are typically excluded, engaging recruitment practices that allow participants to define their own groups and/or social locations in the context of a process focusing on questions about the institutions and forces “that create fully to partly excluded groups,” what maintains them and what can be done to change them (Li, 2015, p. 28; Morris, 2006, p. 26).
Second, deliberative dialogue is designed to enable people to discuss difficult issues, issues of “public moral disagreement” (Moses et al., 2015, p. 196), in a purposeful process specifically designed not to silence subordinated voices. The method gives priority to those at the heart of the research or policy question, allowing them to define the problem from the bottom up and in diverse ways (Morris, 2006, pp. 27, 28). Deliberative dialogue’s recruitment methods also help to combat reification of social constructions of groups by recruiting people “whose activism addresses…sites of intersection” (Morris, 2006, p. 31). Moreover, researchers’ power is constrained by narrowing their role to that of dialogue facilitators (Morris, 2006, p. 32) and can be further constrained by training participants (where necessary) to facilitate their own dialogue.
Third, deliberative dialogue can serve to build community and develop group politics and policy reform (Crenshaw, 1991, p. 1299). Like concept mapping and q-sorting, deliberative dialogue stresses the importance of process over outcome and sensitivity to the dynamics of interactions between participants and between participants and researchers (Pfeil, 2013). Its overtly political nature also allows for participants to share skills, resources and information, and to build capacity in areas such as media relations, lobbying, and conducting research (Morris, 2006, pp. 32, 33). In this way, the deliberative dialogue process publicly acknowledges the strength and resiliency of participant groups, while also offering the prospect for sharing and developing new skills and communities (Morris, 2006, pp. 26, 33).
Fourth, deliberative dialogue “fills the gap between policy analysts and the public by (co)designing participation, creating an arena for expressing and communicating and promoting deliberation” (Li, 2015, p. 35). While it can be challenging to directly connect deliberative dialogue to the formal policy-making process in order to create change (Escobar, 2014), deliberative dialogue offers the prospect of developing alternative policies that are reflective of the needs and experiences of participants. As such, deliberative dialogue “can contribute to a transformatory praxis,” offering the potential to “generate a shift away from discourses of violence in the public sphere” (Pfeil, 2013, p. 121).
Conclusion
Intersectionality theory and practice offer valuable insights for “getting at equality” in a number of fields, two of which have been focused on in this article. First, unless matrices of domination and their relation to people’s lives are recognized, analyzed, and meaningfully addressed, policy and social justice movements risk further subordinating those whose social locations are negatively affected by intersecting systems of domination while further privileging beneficiaries of those systems. Second, in order for social science research to contribute to understandings of how these matrices operate (and how they have been and continue to be resisted), attention must be directed to “disremembered subjects” (May, 2015, p. 53, quoting Foreman, 2013, p. 316) by centering the knowledge, expertise, and techniques of resistance of those who have traditionally been marginalized. While recognizing that categories like race and gender have social and political salience and can both limit and enhance individual and group opportunities, intersectional analysis seeks to deconstruct traditional, single-axis understandings of their effect on lived experiences. Bridging the perceived gap between intersectional analysis as a theory and intersectional analysis as a mode of social science inquiry, however, has proven challenging.
Researchers claiming to pursue intersectional investigations have employed quantitative and qualitative methods, both of which have been subject to praise and to criticism in terms of achieving the equality- and justice-seeking goals of intersectionality. While quantitative research on intersectionality can explicitly address the interactions between systems of domination, thus addressing intersectionality theory’s insights on the interactive, nonadditive impact of socially constructed categories like race and gender, quantitative analysis when deployed on its own does not center the knowledge of those traditionally excluded. Instead, it relies on traditional externally imposed categories and factors to group people without knowing whether those factors are relevant to participants’ lived experiences or consistent with their own understandings of their social location.
While qualitative methods create rich opportunities for learning about social location from those whose knowledge and expertise have traditionally been subordinated, and allow for self-definition in relation to social location, they are limited in their capacity to demonstrate how socially constructed categories and related oppressions work to produce both subordination and power. Like quantitative methods, qualitative work can also risk reification of contested categories depending on the assumptions made by researchers in the process of identifying participants.
In light of these benefits and drawbacks, the challenge to researchers involved with projects such as ours is to determine how to conduct social science research while also advancing intersectional analysis as a social justice practice. Our proposed PAR with mixed methods approach aims to address the diverse social justice objectives of our partners, while also honoring key insights of intersectionality theory. Adopting a PAR approach informed by critical analysis offers the opportunity to promote robust community/research collaboration, affording greater opportunities for learning through active direction from those traditionally excluded from research and policy fora. Concept mapping and q-sorting are consistent with this approach because they offer participants a greater degree of autonomy and minimize (albeit without eliminating) the risk of re-entrenching problematic categories by initially sorting information according to the responses of participants, while still providing an opportunity to examine interactive effects. Deliberative dialogue also extends coresearcher roles to participants, thereby enabling those traditionally excluded from, or silenced in, dialogue to drive inquiry and responses framed to address their needs and experiences, while building political community and movement for change.
These methods individually and together fall short of fully addressing the primary objectives of intersectionality analysis, both as a theory and as a social justice practice. As Hill Collins and Bilge (2016) have warned, “a tidy methodology for intersectional analysis” remains elusive, and our proposed approach is no exception to that observation. However, each of our proposed methods offers something of relevance to the social justice objectives of our partner organizations, while also at least partially getting at some of the more salient goals of intersectionality. As we move forward, we expect to encounter challenges in implementing these methods in accordance with a principled commitment to improving the lived experiences of those who are socially, politically, and economically marginalized by matrices of domination. We hope that the lessons learned from intersectionality and from the voices of those traditionally excluded will guide us in navigating those challenges.
Footnotes
Acknowledgments
The authors thank the Social Sciences and Humanities Research Council of Canada for funding for The eQuality Project, of which this article forms a part, our amazing student researchers Suzanne Dunn, Virginia Lomax, and Michel Mersereau, and the anonymous peer reviewers who contributed invaluable feedback and insight on prior versions of this article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
