Abstract
This contribution proposes nimble intersectionality in response to McBride et al.’s article about intersectional research in the field of employment and industrial relations. Although the authors’ call for all researchers to be ‘intersectionally sensitive’ is positive, regrettably, by highlighting the problems with intersectional methods, they reinforce the widespread perception that they are too difficult to implement. While intersectionality is undeniably complex, this article argues that a nimble approach can help resolve methodological dilemmas. By resolving four basic methodological questions at the onset of a study, researchers can successfully use an intersectional approach to explore age, gender, ethnicity, race and class in employment.
Introduction
This contribution responds to McBride et al.’s (2015) recent article about the importance of intersectionality in employment research. They join a long list of perplexed scholars who wish for a more definite intersectional methodology, without reducing it to a ‘painted by numbers analytical enterprise’ (Crenshaw, 2011: 230). Commendably, McBride et al. (2015) suggest that researchers should be ‘intersectionally sensitive’, not solely to differences between groups, but also to those between individual group members. However, this article contests their position that intersectionality’s complex methodological challenges principally concern a ‘smaller number’ of intersectional specialists (McBride et al., 2015: 334). Rather than distinguishing between types of intersectional researcher, this article aims to fill the gap for a practical intersectional methodology by proposing a nimble approach. Here, nimble denotes the smart and agile application of intersectional theory by researchers, while remaining true to its core principles. The discussion is organized as follows: first, the importance of contextual theoretical and methodological framing is highlighted. Attention then focuses on the difficulties concerning how individual differences are categorized and tracked. The contribution concludes that taking a ‘nimble’ approach can provide solutions to intersectional methodological dilemmas.
A brief reminder of intersectionality’s roots will help ground the discussion. Crenshaw (1991) introduced the term ‘intersectionality’ in 1989 when she explored the discrimination and oppression experienced by black women in the United States (Nielsen, 2013; Walby et al., 2012). Feminist academics, for example, Collins (1998) and Nkomo (1991), continued the debate she ignited, arguing that how women (and men) experience inequalities is shaped not only by gender but also by the interconnection of gender with other social categories of difference. Gender or race cannot be studied in isolation, as a selective lens only reveals fragments of how an organization works (Britton and Logan, 2008) and interpretations are frequently associated with dominant or privileged norms. Therefore, a black feminist research perspective exerts a powerful effect on current meanings about how intersectional studies ‘should be’ conducted.
However, aspiring intersectional researchers query if it really matters whether ‘intersectionality’ is viewed as a means to analyse categories of difference (McCall, 2005), process (Acker, 2012; Yuval-Davis, 2006), paradigm (Hancock, 2007), simultaneous process (Holvino, 2010), matrix (Dhamoon, 2011), analytical process (Winker and Degele, 2011) or tool (Anthias, 2013). Yes, I believe it matters profoundly, as the choice of lens influences every stage of the research design, theoretically and practically. Studies would be more methodologically sound and useful if the ‘specific assumptions that the researcher makes about intersectionality were made more explicit’ (Choo and Ferree, 2010: 146).
Four basic methodological questions
The next section will explain how understanding the practical implications of four methodological questions will enable nimble intersectional researchers (expert or not) to make their assumptions transparent and sound. The four questions are:
Is it an intersectional study?
What is the intersectional framing that fits the context of the research?
Should the study be based on individual identity or organizational and societal processes?
What are the meanings attached to categories of difference?
Is it an intersectional study?
The answer to the first question may appear straightforward – it is an intersectional study if more than one category of difference is studied, and the aim is to reveal difference (Hancock, 2007). However, all too frequently, the question of what exactly constitutes an intersectional study is polarized between critical race and feminist understandings. Bilge (2013) argues that a methodology designed to explore the oppression of black women has been commandeered by European feminists (e.g. Lutz et al., 2011), as an intellectual exercise to explore other dimensions of difference, such as disability. Nash (2008) contests this perspective; to associate the method only with black marginalized women is as blinkered as the former privileged white middle-class feminist lens. Crenshaw (2011) considers the differences between theoretical positions are less important than the diverse aims and accomplishments of intersectional studies and projects across various disciples.
Currently, refined by these periodically opposing traditions, intersectionality has developed into a theoretical framing that can be used to explore the complex intersecting relationships between diverse groups and social structures, beyond the confines of women of colour, in different contexts. However, notions of privilege and penalty are subjective and viewing different categories of people as only attracting privilege or suffering penalty cannot capture the complexity of multiple intersections of points of difference (Dhamoon, 2011). Methodological decisions may exclude people who become, in essence, ‘missing voices and experiences’ (McBride et al., 2015: 338).
Therefore, intersectional researchers are charged with investigating the ‘silences’ of individuals who deviate from the more obvious and dominant norms. In contemporary iterations of intersectionality – for example, simultaneous process (Holvino, 2010) or translocation analysis (Anthias, 2013) – dominant/privileged (frequently white in western contexts) individuals are not positioned at the centre of research. Rather, researchers track how the interconnections between sex, race, ethnicity and sexuality produce patterns of equality and inequality, dominance and subservience, in particular social contexts (Calas et al., 2014). To summarize, a study is intersectional when the aim is to reveal differences (more than one category), between different groups, and/or in different groups, and where dominant/privileged individuals are not positioned at the centre of the research.
What is the theoretical framing that fits the context of the research?
The second step in a nimble intersectional study is to choose a theoretical foundation that reflects the sectoral and locational context. It is necessary to review literature, empirical and critical, detailing how work is organized across the specific sector. In organizations, advantage and disadvantage is ‘empirically variable by context’ (Walby et al., 2012: 234), revealing nuances of hidden power differentials in workplace relationships (Acker, 2012; Dhamoon, 2011; Hancock, 2007; Holvino, 2010; Winker and Degele, 2011). Hence, a suitable framing, such as career or labour market theory, is required to interpret patterns for both the privileged and non-privileged in a particular context. To illustrate, McDowell et al. (2014) use labour market theory to highlight the social and economic exclusion experienced by young South Asian men in Greater London. In an intersectional sense, not all theories are equal; for example, many ‘boundaryless’ career studies focus on the professional elite, neglecting the disadvantages of precarious work arrangements for unskilled workers (Roper et al., 2010). Contemporary rather than traditional approaches may be more successful in analysing the practices that regulate employment at different hierarchical levels – for example, the evolving focus on ‘boundaries’ in ‘boundaryless careers’ exposes critical career junctures that exist in all careers (Inkson et al., 2012).
Defining appropriate and specific research questions is also of paramount importance. An intersectional approach must pay attention to both the practical and theoretical aspects to be effective as both a ‘normative and empirical’ framework (Hancock, 2007: 251). Following the initial contextual review, research questions must be broken down into separate and specific foci, as it is immensely difficult to link wide-ranging data sets at a later stage. One way to organize intersections between aspects of difference is to ‘split the data entanglements’ (Fletcher et al., 2012). For example, the first research question might explore the expected employment outcomes in the sector. Subsequent research questions can then centre on the experiences of diverse individuals against the contextual background revealed by the first question. This separation positions the experiences of both marginalized and privileged against what is perceived as ‘normal’. As ‘normal’ generally reflects the norms of the dominant group – for example, managerial hetero-normal norms (Pringle, 2008) or whiteness (Bilge, 2013; Holvino, 2010) – it is crucial that analysis uncovers advantage and disadvantage by challenging habitual organizational practices. Only by understanding ‘normal’ can the intersections of difference be tracked during analysis.
Sectoral context gained from a broad initial review (Winker and Degele, 2011) may indicate what categories of difference to research, as particular social groups will dominate some employment sectors. In diversity studies, for example, Tatli and Özbilgin (2012: 188) expose some weaknesses of the traditional ‘etic’ intersectional approach, where specific categories of difference are predetermined and static. In contrast, the emic approach is based on an organization’s historical and geographical background, revealing the social categories ‘creating and sustaining privilege and disadvantage in the specific context’ (Tatli and Özbilgin, 2012: 191). Therefore, the comprehensive examination of literature previously undertaken by the nimble researcher may indicate salient categories of difference, while recognizing that some categories will remain hidden. For example, in hospitality career studies, obvious categories, such as ethnicity, emerge at an early stage; however, the way sexual orientation influences progression is less obvious.
Should the study be based on individual identity or organizational and societal processes?
McBride et al. (2015: 338) query whether the underlying motivation of a study is ‘to understand the lived experience of workers at the intersection or to better understand the dynamics of power at work?’ Accordingly, nimble researchers must decide whether analysis will commence with examining individual identity or with organizational processes. Due to intersectionality’s black feminist origins, many earlier intersectional studies were based on identity (Proudford and Nkomo, 2006) and addressed the question of ‘who’ an individual was (Yuval-Davis, 2006). Intersectional ‘process models’ focus on the ‘context and comparison at the intersections as revealing structural processes organizing power’ (Choo and Ferree, 2010: 134), rather than centring on individual identities.
The advantage of an identity-based approach may be that it enables researchers to investigate two dimensions of socially constructed identity simultaneously, without privileging one over the other; for example, women’s ‘double-bind’ of gender and leadership (Fletcher et al., 2012). However, the disadvantage of an identity-centred approach that explores more than two dimensions of difference is the difficulty of balancing the complex intersections of multiple identities. Identity remains a central tenet in McCall’s (2005) three innovative approaches to analysing categories of difference; although McCall suggests analysis should include structural levels, she does not indicate specifically how such analysis should proceed. This remains a key problem, because ‘at the heart of the [intersectional] debate is conflation or separation of the different analytical levels’ rather than the relationship between categories (Yuval-Davis, 2006: 195).
Practically, identity-based studies are problematic and Acker (2006b, 2012) warns that relationships between competing aspects of individual identity are very difficult to track. Dhamoon (2011) reiterates Acker’s advice to focus on processes and systems. Acknowledging the ‘boundedness of identity’, especially when multiple aspects of identity are involved, reiterates ‘the norms it aims to challenge’ (Dhamoon, 2011: 233). However, deciding to take an organizational process approach based on fixed categories will also present pitfalls if the process model does not reflect the extent of an individual’s self-determination within organizational and societal systems. Nimble researchers should choose a model that considers social identity to be ‘constructed through … or co-constructed with, macro and meso categories and relations …’ (Choo and Ferree, 2010: 134), thus revealing the dynamic nuances of identity and societal location. Analysis then reflects the power differentials and individual agency of workers within the context of societal structures and managerial practices, rather than viewing them as passive recipients of organizational processes.
It can consequently be difficult to arrive at a decision about the most suitable approach. Without prior consideration, practical issues associated with either approach may only emerge when the study is well advanced. Winker and Degele (2011) appear to offer the only detailed method to link individual identity with organizational processes and institutional structures. Their multi-level model of intersectional analysis is designed to overcome the vexing issues highlighted by Yuval-Davis (2006): how to track dynamic categories of difference; how to prioritize intersecting categories; and how to connect different levels of analysis (e.g. from individual level to policy/societal/global level).
The principal challenge associated with Winker and Degele’s (2011) model is that their first step of analysis focuses on identity, creating a problem in a retrospective or longitudinal study. In this researcher’s experience, the fluidity of categories such as age, intersecting with aspects of gender (e.g. motherhood) that change over the course of participants’ employment histories, renders the lists of competing and intersecting identities almost limitless. It is impractical to manage the relationships in the data sets while linking individuals’ myriad identities with workplace practices and structural frameworks. However, the nimble researcher may adapt their method so that the first step of analysis focuses on specific industry norms, before connecting them with organizational processes associated with workplace advantage and disadvantage, rather than individual identity (see Mooney et al., 2014). At the final stage of multi-level analysis, intersections of organizational processes and institutional structures are linked to individual identity categories.
What are the meanings attached to categories of difference?
There will be further issues with defining ‘difference’ irrespective of the decision to take an identity or process-based approach. Such debate is not specious posturing as each theoretical position has practical ramifications; this section aims to clarify the impact of these decisions.
Establishing the categories of difference to study
Challenges surround deciding whether to establish categories of difference at the onset of the study (Holvino, 2010) or whether to allow categories of difference to emerge during data analysis (Tatli and Özbilgin, 2012; Winker and Degele, 2011). Acker (2012: 221) argues, for instance, that questions about the category under investigation, such as gender or race, need an a priori intersectional determination. Practical considerations support an early decision by the nimble researcher to resolve which categories to investigate. Even sophisticated data organizing software, such as NVivo, will find it difficult at later stages to connect individual categories of difference with organizational processes and institutional structures, if not identified and coded at the beginning of a study. Scrutinizing sector-specific environments in advance, and being alert to contextual factors, reduces the possibility that researchers will overlook less obvious categories. However, the risk will remain; ‘perfectly’ covering all dimensions of difference – visible or invisible – is an impossible objective.
A further contested dilemma with practical implications rests with the decision of whether or not to privilege one particular category, for example, race/ethnicity, over another, such as gender. Intersectionality has complicated approaches to gender, in that gender tends not to be studied insularly but in combination with race and class (Acker, 2012). As previously observed, researchers following a black feminist tradition, such as Bilge (2013), believe race should be the primary category. Other feminist scholars, such as Broadbridge and Simpson (2011), argue intersectionality has possibly undermined gender research in certain fields, such as management. They suggest that a potential solution might be to give gender the primary position in such studies, while acknowledging ‘plurality of differences’; for example, gender could be studied ‘with’ race and class, as opposed to gender ‘and’ race and class (Broadbridge and Simpson, 2011: 476). Research aims will drive the final choice, and nimble researchers should take comfort from Hancock’s (2007) conclusion that paying equal attention to categories of difference does not imply that the relationship between them is known or equal (this author’s italics). Ultimately, although research objectives, context and researcher positioning all influence decisions, the data must be allowed to speak and reveal the weighting of each category of difference in that particular context. Resolving this point leads to the next significant decision about who to include in the study.
Defining the context and parameters of the study at an early stage, as indicated earlier, enables nimble researchers to distinguish significant aspects of difference. Contemporary iterations of intersectional theory contest more established notions about appropriate sample sizes and populations. For example, Özbilgin et al. (2010) argue that the focus on an idealized worker – primarily female, heterosexual, white and middle class – in many diversity and work–life balance studies, neglects the experience of those from other groups, for example, gay parents. Therefore, intersectional studies that include men and women and participants from different ethnic backgrounds are more likely to reveal diversity of experience; supporting McBride et al.’s (2015: 338) view that including participants ‘with power’ helps to illuminate ‘systems of domination’. Otherwise, diversity may be only linked with those considered to be marginalized in a particular context, for example, women in management studies.
Unfortunately, it remains a considerable challenge to give voice to those who appear to be privileged without further entrenching their privilege (Broadbridge and Simpson, 2011). A core principle of intersectionality is that researchers recognize that diversity also exists within each grouping of difference; therefore, in this sense, McBride et al.’s (2015) argument to adopt an ‘intersectionally sensitive’ approach as a minimum requirement is compelling. Dhamoon’s (2011) acknowledgement that privilege and penalty are not mutually exclusive underscores the obligation; depending on the contextual circumstances (this author’s italics), an individual from an ethnic minority may experience simultaneous privilege and oppression and some women may be more privileged than some men.
Interpreting categories of difference
Socially ascribed categories of difference are imbued with different connotations depending on researchers’ perspectives, and the nimble researcher will ensure that the chosen lens matches the epistemological and ontological base of the study. Holvino’s (2010) ‘simultaneous process’ model highlights intersectionality’s perennial problem: the effects of gender, class and colour are inseparable and it is difficult to see exactly when one effect supersedes the other(s). Therefore, to avoid repeating McBride et al.’s (2015) excellent overview, this section will highlight how the specific ways that primary categories such as age, gender, race/ethnicity and class are conceptualized and interpreted affect study design.
In comparison to the more usual fixed categorizations, Hancock’s (2007) visualization of difference aligns with a process-based approach; she suggests ‘difference’ is dynamically expressed at both structural and individual levels and that meanings vary according to the particular context. Although Walby et al. (2012) consider Hancock’s (2007) understanding of ‘fluid’ categories difficult to implement, it becomes easier to understand ‘fluid’ in the organizational sense, expressed as social role performance. The significant advantage of viewing categories as ‘performance’, instead of rigid classifications, is the superior way in which nimble researchers can explore how difference is enacted through performance norms, at individual, organizational and societal level. Data are produced that capture ‘both the agency of individuals in making the world they inhabit and the enabling and constraining forces of the world as it has been produced’ (Choo and Ferree, 2010: 134). Age, gender, ethnicity and class may all be constructed and analysed as ‘performance’, as the following brief outline explains.
Age is particularly salient for workers’ employment opportunities or lifestyle choices and is frequently investigated as a fixed chronological dimension, for example, 20–30 year olds. Regrettably, this orientation neglects the rich meanings associated with being older or younger than prevailing norms. Therefore, Winker and Degele’s (2011) notion of ‘body’ is a highly significant contribution to intersectional theory, encompassing an individual’s chronological age and associated suitability (the fusion of desired appearance, performance and life stage) for a particular job. The performance perspective of gender, likewise, is associated with the differing social roles of men and women – gendered performance is anticipated and defined (by individuals themselves as well as others) through workplace practices and processes (Acker, 2012; Lewis and Simpson, 2010). In intersectional studies, a performance lens further enhances researchers’ ability to explore how masculinity and femininity are ‘performed’ in workplaces.
Whether to choose a ‘race’ or an ‘ethnicity’ lens remains deeply contested terrain with ontological and practical implications for researchers. Acker (2006b: 444) regrets that feminist scholars have never really defined race; her perspective views race as ‘socially defined differences based on physical characteristics, culture, and historical domination and oppression, justified by entrenched beliefs’, which may or may not accompany ethnicity as grounds for oppression. The notion of ethnicity, on the other hand, acknowledges variances in ‘cultural markers such as language, values, traditions and national origin’ (Atewologun and Singh, 2010: 334), which are socially constructed. Anthias (2013) responds to the way globalization has changed meanings about social positioning in intersectional studies. Her translocational perspective considers the categories of class, gender and race ontologically ‘valueless’, suggesting social class identities are formed from the experiences of ‘transnational migrant “others” in particular localities’ (Anthias, 2013: 124).
A practical challenge exists about interpreting race or ethnicity as ‘performance’ according to these definitions. Again, harking back to the importance of specific context, the way statistics are reported may sway nimble researchers’ decisions to explore either race or ethnicity. Not all countries use the term ‘race’ as a demographic descriptor, for example, New Zealand. Furthermore, it is not always realistic or practical to represent the individual racial or ethnic identity of participants if they are few in number. In such cases, the term ‘ethnic minority’ may be adopted to describe demographic minorities, resolving the challenge of reductive analysis (Dhamoon, 2011) associated with the positioning of small minority groups (Walby et al., 2012). As suggested earlier, there are advantages associated with categorizing participants as either belonging to the dominant or non-dominant group. By deliberately not positioning dominant white individuals at the centre of research (Calas et al., 2014), studies will not infer that all people of minority ethnicity experience the same complex mix of privilege/penalty or advantage/disadvantage.
Finally, although a ‘class’ lens occupies the last position in the current discussion, this contribution argues that the under-researched category of class should be ‘rediscovered’ in employment studies, as class reflects privilege in ‘organizational hierarchies of power and reward’ (Acker, 2006a: 141). For intersectional researchers, ‘class’ is perceived as abstract and difficult to track (Walby et al., 2012) and regrettably it is beyond the scope of this article to convey its multi-faceted social meanings. However, in employment studies, a nimble way to handle class is to view it as ‘occupational class’, conveying the tangible and intangible expressions of rewards. Rewards link with performance; for instance, access to training. Viewing class as embedded in organizational processes rather than as a descriptor (Choo and Ferree, 2010) gives researchers a ‘window’ (Scully and Blake-Beard, 2006: 446) to explore the effects of individual differences in specific employment settings. Although here, only four categories of difference were construed as performance, other dimensions of difference such as disability could, and should, be added to this group.
Presenting intersectional findings
It is mystifying how many articles with an intersectional focus fail to recommend technological tools that improve the analysis and presentation of findings. There are exceptions – Dhamoon (2011: 237) suggests using a ‘matrix’ design to express multi-dimensional intersections and Choo and Ferree (2010: 136) indicate that intersectional models, which track interactions between class and gender rather than the ‘main effects’, would benefit from computerized feedback programs. Indisputably, software programs such as NVivo and ATLAS.ti offer substantial advantages to nimble researchers by organizing and tracking elaborate intersections in identity and process-based studies.
Intersectional researchers must eventually realize how the labyrinth-like nature of intersectional findings gives intersectionality its critical edge. Software applications, for example, ‘Inspiration’, can assist in presenting complex findings more clearly, translating the nuanced relationships between aspects of difference at individual and structural level into graphic ‘maps’. When the overarching research question is further separated into smaller, more manageable ‘chunks’ as advised earlier, such a ‘map’ can respond to the scope of the study, charting the norms, sectoral structures and institutional regulators in a particular employment context. At later stages, further maps may then be superimposed over the first, to show how individual differences are reflected in variations of the anticipated employment constructs. In this way, visual representations clarify and enhance the presentation of empirical and normative data provided by well-designed intersectional studies (Hancock, 2007; Walby et al., 2012).
A nimble intersectional approach: the way forward
Intersectionality can help researchers convey the separate and cumulative effects of being ‘different’ in more than one dimension across diverse employment settings. The nimble intersectional approach means researchers should make early decisions about theoretical framing – whether to base the study on individual identity or processes, how they will research ‘differences’ and how they will present their findings. Each decision will have practical implications. The methodological considerations apply equally to quantitative researchers. Bowleg (2012) decries the lack of intersectional teams, including statisticians and qualitative analysts, in public health research, which would clearly benefit from an intersectional approach.
In conclusion, ‘intersectionality applies to everyone – no one exists outside the matrix of power’ (Crenshaw, 2011: 230). Intersectionality’s real strength lies in its ability to reveal the experiences of the ‘sometimes’ marginalized and the ‘sometimes’ privileged, providing empirical and normative contributions to knowledge about work and employment relations in specific sectors. Unfortunately, intersectionality as method has remained marginalized, arguably because a definitive methodology has been lacking. The introduction questioned whether it ‘really matters’ how intersectionality is conceptualized. Yes, ‘it matters profoundly’ because when ontological decisions are explicit and coherent, the research design becomes easier to implement and justify, even by non-intersectionality experts. Therefore, nimble researchers who address four basic methodological questions can successfully unleash intersectionality’s emancipatory possibilities in employment studies.
Footnotes
Acknowledgements
The author would like to thank Paul Brook and the three anonymous reviewers for their detailed feedback, which significantly improved the coherence of the article’s argument. The author would also like to acknowledge how the lively debates over the years with Irene Ryan, Candice Harris and Judith Pringle have helped to deepen the author’s understanding of intersectionality as concept and method.
Funding
The author received no financial support for the research, authorship and/or publication of this article.