Identity Pie and Descriptions

ORIGINAL ARTICLE

When Black + Lesbian + Woman ≠ Black Lesbian Woman:
The Methodological Challenges of Qualitative
and Quantitative Intersectionality Research

Lisa Bowleg

Published online: 21 March 2008
# Springer Science + Media, LLC 2008

Abstract The notion that social identities and social
inequality based on ethnicity, sexual orientation, and sex/
gender are intersectional rather than additive poses a variety
of thorny methodological challenges. Using research with
Black lesbians (Bowleg, manuscripts in preparation;
Bowleg et al., Journal of Lesbian Studies, 2008; Bowleg
et al., Cultural Diversity & Ethnic Minority
10:229–240, 2004; Bowleg et al., Journal of Lesbian
Studies, 7:87–108, 2003) as a foundation, I examine how
these challenges shape measurement, analysis, and inter-
pretation. I argue that a key dilemma for intersectionality
researchers is that the additive (e.g., Black + Lesbian +
Woman) versus intersectional (e.g., Black Lesbian Woman)
assumption inherent in measurement and qualitative and
quantitative data analyses contradicts the central tenet of
intersectionality: social identities and inequality are inter-
dependent for groups such as Black lesbians, not mutually
exclusive. In light of this, interpretation becomes one of the
most substantial tools in the intersectionality researcher’s
methodological toolbox.

Keywords Intersectionality research methods .

Black lesbians

Introduction

Black lesbian poet Audre Lorde’s (1984) description of “…
constantly being encouraged to pluck out some aspect of
myself and present this as the meaningful whole, eclipsing
and denying the other parts of the self” (p. 120) highlights
eloquently the complexity of intersectionality. For Lorde
and other Black lesbians, one’s identity as a Black lesbian
is the meaningful whole; it is not a mere addition of
ethnicity, sexual orientation, and sex/gender. For research-
ers interested in designing and conducting intersectionality
research, the notion that social identities and social
inequality based on ethnicity, sexual orientation, sex/gender
(and one could add a host of other identities such as class,
disability status, etc.) are interdependent and mutually
constitutive (i.e., intersectional; Collins 1995, 1998; Crenshaw
1989, 1991; Weber and Parra-Medina 2003), rather than
independent and uni-dimensional poses a variety of thorny
methodological challenges. These challenges shape key
aspects of the research process such as measurement, data
analysis, and interpretation. Using research with Black
lesbians as a foundation (Bowleg, manuscripts in prepara-
tion; Bowleg et al. 2008, 2004, 2003), I focus in this article
on some of the methodological challenges of conducting
intersectionality research using qualitative and quantitative
methods.

The solipsism that equates women with Whiteness, and
Blackness with men rendering the experiences of people who
are women and Black invisible (Hull et al. 1982; Spelman
1998) pervades contemporary scholarship, policy, and
thought. Copious examples exist, but one will suffice. In

Sex Roles (2008) 59:312–325
DOI 10.1007/s11199-008-9400-z

The Trials and Tribulations Study was supported by a 2000 University
of Rhode Island Council of Higher Education Grant.

I presented portions of this article as Black + Woman + Lesbian?
Black x Woman x Lesbian?: Some Conceptual and Methodological
Challenges of Intersectionality Research at the “Practical Issues in
LGBT Research” symposium at the August 2006 Annual Convention
of the American Psychological Association Convention, New Orleans,
LA, USA.

L. Bowleg (*)
Department of Community Health and Prevention Program,
School of Public Health, Drexel University,
1505 Race Street, Mailstop 1032,
Philadelphia, PA 19102-1192, USA
e-mail: [email protected]

1994, the National Institutes of Health released guidelines
requiring researchers to describe their plans for the
“Inclusion of Women and Minorities.” The rationale for
the guidelines were sound but the continued declaration of
“women and minorities” as if these were two mutually
exclusive groups signaled an entrenched misunderstanding
of how women’s experiences as women also intersect with
their experiences as members of ethnic minority groups, as
well as other historically oppressed social groups.

The discipline of psychology has not fared well in terms
of promoting the understanding of intersectionality. Despite
an abundance of theories on social identity within psychol-
ogy, the prevailing view of social identities is one of uni-
dimensionality and independence, rather than intersection.
A notable exception is Ransford’s (1980) multiple jeopar-
dy-advantage (MJA) hypothesis which posits that people
occupy various social status positions that intersect to create
a “unique social space” (p. 277). This unique space manifests
as outcomes that one’s social status location (e.g., race) alone
cannot explain. Instead, this space can be explained only by
the intersection of one of more social status positions (e.g.,
race, sex, class, sexual orientation) to yield multiple jeopardy
(i.e., the intersection of two or more low social status
positions) or multiple advantage (i.e., the intersection of two
or more high social status positions). Deaux’s (1993) work
reconstructing social identity to recognize multiple dimen-
sions of social identity is another exception to the rule, as
are numerous examples within feminist psychology.
Though explicit mention of the term intersectionality is
rare, feminist psychology has been far more progressive
than mainstream psychology in recognizing the intersec-
tions between women’s experiences of structural inequality
based on race, gender, class, and sexual orientation (e.g.,
Greene 1997; Reid and Comas-Diaz 1990; Weber 1998).

Not surprisingly, the Black feminist literature is replete
with narratives and analysis of Black women’s experiences
of the intersections of race, gender, class, and/or sexual
orientation (see Collins 1991; Davis 1983; Hooks 1981)
and notions of double (Beale 1970) and triple jeopardy
(Greene 1995) based on these identities. But while a
plethora of scholarship on the intersection of race, class,
gender, and sexual orientation exists in the multidisciplin-
ary literature, there is a paucity of literature on intersection-
ality from a methodological perspective (Cuadraz and Uttal
1999; McCall 2005). Thus, researchers interested in
conducting intersectionality research often have to self-
teach and learn through trial and error. My goal in writing
this article is to help intersectionality researchers sidestep
some mistakes, as well as learn from the insights gained
from the qualitative (Bowleg, manuscripts in preparation;
Bowleg et al. 2008, 2003) and quantitative (Bowleg et al.
2004) intersectionality research with Black lesbians of
Bowleg et al. Far from being a comprehensive treatise of

the methodological issues that arise in intersectionality
research, I address here three issues with which intersec-
tionality researchers must grapple: developing questions to
measure intersectionality, analyzing intersectionality data,
and interpreting them.

Black Lesbians in Micro and Macro Perspective: A Brief
Overview

Because the lives of Black lesbians are rooted in structural
inequalities based on the intersections of sexual orientation,
sex, gender, and race (see Greene 1995), Black lesbians are
an ideal population in which to study intersectionality.
Intersectionality examines how distinctive social power
relations mutually construct each other, not just that social
hierarchies exist (Collins 1998). At the micro level, a small
empirical literature base has examined the intimate relation-
ships (Hall and Greene 2002; Mays and Cochran 1988;
Peplau et al. 1997), health care (Cochran and Mays 1988),
mental health (Cochran and Mays 1994), workplace
(Bowleg et al. 2008), active coping (Bowleg et al. 2004),
and multiple minority stress and resilience (Bowleg et al.
2003) experiences of Black lesbians. Other relevant
scholarship, most of it focused on predominantly White
middle-class lesbian, gay, and bisexual (LGB) populations
has addressed the minority stress experiences of LGBs
(Brooks 1981; DiPlacido 1998; Meyer 2003) and the dual-
identity experiences of lesbians (Fingerhut et al. 2005).

A macro level analysis of economic inequality from an
intersectional perspective demonstrates aptly how the social
hierarchies of race, sex, and sexual orientation are mutually
constructed in the lives of Black lesbians. According to the
National Gay and Lesbian Task Force’s (NGLTF; Dang and
Frazer 2004) analysis of Black same-sex household data
from the 2000 U.S. Census Bureau, Black same-sex couples
reported an annual median household income ($49,000) of
$2,000, lower than that of their Black married heterosexual
counterparts ($51,000). Black female same-sex couples
however reported a median income of $9,000 less than
Black married heterosexual couples, $7,000 less than Black
male same-sex couples, and a stunning $21,000 and $29,000
less than White female and male same sex couples
respectively, illustrating clearly how structural inequalities
grounded in intersections of race, sex, and sexual orientation
affect Black female same-sex couples adversely.

Pitfalls and Insights: Lessons from Intersectionality
Research with Black Lesbians

Two studies, the Black Lesbians Stress and Resilience Study
(BLSR), a mixed methods study with Black lesbians in
southern California (Bowleg, manuscript in preparation;

Sex Roles (2008) 59:312–325 313313

Bowleg et al. 2008, 2004, 2003), and a qualitative study
with a subsample of Black lesbians in Washington, DC who
were part of the Trials and Tribulations Study (TT), a larger
study of Black lesbian, gay, bisexual and transgender
(LGBT) people (Bowleg, manuscript in preparation) pro-
vide the foundation for the methodological challenges that I
highlight in this article.

The goal of both studies was to explore and examine
experiences of multiple minority stress and resilience relevant
to the intersections of race, sex/gender and sexual orientation
for Black lesbians. Despite the researchers’ interest in the
intersection of and social inequality based on these identities,
much of it prompted by the primary author’s own experience
as a Black lesbian, the research team knew virtually nothing
about intersectionality theory or research. The proof: none of
the literature review sections of these articles reference a
single intersectionality theorist, or even mention the word
intersectionality. Instead, the prevailing wisdom of the triple
jeopardy approach to Black lesbians’ experiences (e.g.,
Greene 1995) informed much of the empirical exploration
of what Bowleg et al. (2003) called at the time “multiple
marginalized identities” (p. 89). The researchers’ realization
that virtually every methodological choice made in these
studies reflected an additive approach (Black + Lesbian +
Woman), antithetical to the theoretical fidelity of intersec-
tionality would come later, most of it revealed through the
research participants’ poignant and complex narratives about
the intersections of ethnicity, sex/gender, and sexual orienta-
tion in their lives. Thus, trial and error, those two marvelous
teachers, inform the methodological issues that I discuss rel-
evant to measurement, data analysis, and interpretation. For
each of these three domains, I present a key methodological
challenge, use examples from the research of Bowleg et al.
(Bowleg, manuscript in preparation; Bowleg et al. 2008,
2004, 2003) to highlight relevant issues, and conclude with
insights and recommendations for addressing the challenge.

Measuring Intersectionality

It is so obvious as to not even warrant mention: the wording
of questions shapes how participants respond to them.
Accordingly, a bounty of excellent resources exist for
researchers interested in designing good qualitative (e.g.,
Patton 2002) and quantitative (e.g., Bradburn et al. 2004)
questions. Obviously, asking good questions is vital to
intersectionality research too, but doing so well can be quite
challenging. At issue is how to ask questions about
experiences that are intersecting, interdependent, and
mutually constitutive, without resorting, even inadvertently,
to an additive approach.

The additive approach posits that social inequality
increases with each additional stigmatized identity. Thus,

a Black lesbian would be multiply oppressed because of the
combination of her ethnicity, sexual orientation, and sex/
gender (i.e., triple jeopardy). Critics reject the additive
approach because it conceptualizes people’s experiences as
separate, independent, and summative (Collins 1995;
Cuadraz and Uttal 1999; Weber and Parra-Medina 2003).
Furthermore, they disavow the additive approach’s impli-
cation that one’s identities and/or discrimination based on
these identities can be ranked (Collins 1991; Cuadraz and
Uttal 1999; Weber and Parra-Medina 2003). Weber and
Parra-Medina (2003) have asked rhetorically: “How can a
poor Latina be expected to identify the sole—or even
primary—source of her oppression? How can scholars with
no real connection to her life do so?” (p. 204). They
contend further that people can be members of dominant
and subordinate groups (e.g., a White man with a physical
disability) simultaneously thereby rendering the ranking
exercise futile (Weber and Parra-Medina 2003). Alas, what
holds in theory does not always translate easily to practice.
Indeed, I would argue that it is virtually impossible,
particularly in quantitative research, to ask questions about
intersectionality that are not inherently additive.

Lessons from the BLSR and TT Studies of Bowleg et al.

The conceptual framework of triple jeopardy (Greene 1995)
shaped the design of both the BLSR (Bowleg et al. 2008,
2003) and TT (Bowleg, manuscript in preparation) studies.
Applied to Black lesbians, this framework is implicitly
additive: Black lesbians are subject to prejudice and
discrimination based on their ethnicity, sex, and sexual
orientation. Three lessons from these studies are key: (1)
ask an additive question, get an additive answer; (2) the
problem of attempting to measure intersectionality through
addition; and (3) ask precisely what you want to know.

Ask an Additive Question, Get an Additive Answer

Consistent with the additive approach, Bowleg (manuscript
in preparation) posed questions in the qualitative phase of
the TT study that implied that participants’ identities could
be isolated and ranked:

Some of the people we’ve spoken to have told us that
when it comes to their identities, they are Black first,
and gay, lesbian or bisexual second. Other people said
that they are gay or lesbian first and then Black or
female, second. Still others have said that they don’t
feel as if they can rank these identities. In terms of
your life, do you rank these identities, that is by race,
sexual orientation, gender or anything else?

Not surprisingly, many interviewees responded in kind.
That is, they ranked their identities. For example, Loretta, a

314 Sex Roles (2008) 59:312–325

33 year old lesbian noted that she did rank her identities: “I
think I do. I’m African American first but for a while I was
lesbian first and before that I was just [Loretta] and couldn’t
understand what all the fuss was about” (Bowleg, manu-
script in preparation). Although Maggie, a 27 year old
lesbian initially challenged the request to rank her identi-
ties, noting, “[No]. I’ve thought about that and I don’t think
I can,” (Bowleg, manuscript in preparation) she nonetheless
proceeded to do just that:

No, I would say that I’m gay first because being a
lesbian has had such an impact in my life that it has
put me into a different category than just being an
African American. It seems like if I were going to be
discriminated against about something that would be
the first thing. If someone had a choice to hate me or
discriminate against me for something that I was that
would probably be the first thing picked. And that is
the thing I feel I am discriminated against the most. So
then that seems to have the biggest impact so I guess
that’s why it gets first place. And then second place is
being Black. Regardless of where I go being Black in
any part of the world being Black is an issue. Even in
Black countries it’s an issue. Black women and White
woman get treated differently in every country
(Bowleg, manuscript in preparation).

By contrast, others such as Karen, a 36 year old lesbian,
reflected the intersectionality perspective with their rejec-
tion of the notion that they could rank their identities. Karen
observed, “No, I always resort to ‘there is no higher
political repression.’ So I personally don’t ascribe to that
I’m Black first, lesbian second, woman third. I’m all those”
(Bowleg, manuscript in preparation).

Attempting to Measure Intersectionality Through Addition

Another question from the Bowleg (manuscript in prepara-
tion) TT study asked: “… If someone dropped in from
another planet and asked you to tell them about your life as
a Black lesbian woman. First, what would say about your
life as a Black person?; Woman? Lesbian?; and Black
lesbian woman?” It is obvious now in retrospect that a truly
intersectional question would simply ask the respondent to
tell about her experience without separating each identity.
This is precisely what Karen implied in her response to the
question about her life as a Black or African American
woman?: “Well, you probably could combine all those
statements” (Bowleg, manuscript in preparation). The
research suggests that even if the interviewer omitted the
question singling out each identity, respondents might still
seek to do so. For example, at a later point in the interview
when asked, “In terms of your own life, what are some of
the things you like most or the advantages about being

Black and lesbian?” Karen countered, “Not Black, lesbian,
and woman? Just Black and lesbian?” (Bowleg, manuscript
in preparation). Her questions seeking clarification high-
light the importance of articulating intersectionality explic-
itly in interview questions. Even if a respondent asks an
interviewee to disaggregate identities, it seems advisable for
the interviewer not to do so, but to instead invite the
interviewee to discuss her identities and experiences
however they best resonate with her. Karen’s counter
questions are also a fitting example of the problem of
assuming that the experience of being a woman is
subsumed within that of being lesbian.

Ask Precisely What You Want to Know

The aforementioned measurement mistakes notwithstand-
ing, an interview question in Bowleg’s (manuscript in
preparation) BLSR study elicited narratives that captured
the experience of intersectionality. For example, in response
to the question, “What are some of the day-to-day
challenges that you face in terms of your race, gender
and/or sexual orientation?” Nancy, a 44 year old lesbian
with a physical disability stated: “Getting listened to. I
think that a lot of time people discredit me because I am a
Black lesbian, who walks with a cane most of the time”
(Bowleg, manuscript in preparation). Ethnicity, sex, sexual
orientation and disability intersect in Nancy’s narrative,
these are not discrete identities.

In the TT study, Johanna, a 36 year old woman who said
that she sometimes identified as lesbian and other times as a
lesbian-identified bisexual depending on her audience,
described the intersection of her identities this way:

I clearly … see myself as Black first. Although … I feel
that … I am not just Black, but I’m also a woman, I’m
lesbian identified bisexual, I also come from a
working class background. So I see those other parts
of myself (Bowleg, manuscript in preparation).

Johanna’s presentation of her identities was seamless;
the absence of the conjunction and in her description
underscores her perception of her identities as intersectional
rather than additive. Noteworthy in Johanna’s mention of
socioeconomic class, an identity that the interviewer did not
ask about explicitly, is the reality that interviewers are
limited in the number of different identities about which
they can ask questions. It is simply not practical for an
interviewer to ask an exhaustive list of questions about
intersecting identities (e.g., class, disability status, etc.). If
the researcher asks the question well, however (i.e., by
inviting participants to discuss any other dimensions that
are important to them), then the interviewee can add, as
Johanna did class and other dimensions that the researcher
might otherwise have overlooked.

Sex Roles (2008) 59:312–325 315315

Asked how she typically described herself, Kim, a
33 year old lesbian interviewee from the TT study
explained, “I think I usually describe myself as a Black
lesbian or African American lesbian cause it feels like I
have to carve myself some space for myself there because
there isn’t already” (Bowleg, manuscript in preparation).
Thus, for Kim the intersection of these identities formed an
interdependent identity that she presented to the world
rather than a summation of additive identities.

By contrast, in response to the same question others
such as Leslie, a 48 year old lesbian in the BLSR study
(Bowleg et al. 2003) separated each identity, illustrating a
disconnect between how the researchers intended to have
the question answered (i.e., with a focus on all of the
intersecting identities rather than single identities) and
how Leslie perceived and interpreted the question (i.e.,
additively):

Well, the primary challenge would be around race…
Because it’s like every day you get up and you don’t
know if you will get to work without one of these mad
dog police pulling you over and getting into a beef and
you get arrested; then you lose your job. You don’t
know if you’ll get home at night. You don’t know if
when you go shopping they’ll put security on you and
be following you around the store. The queer part is
probably something … I personally encounter in up
close relationships so it would probably be in a work
environment or just out in the street where maybe a
guy is hitting on me or something. And the woman
part is kind of like the same [as the queer part] where
you interact with men on the street and at work with
your coworkers or bosses (p. 14).

As for how one might measure intersectionality quanti-
tatively, none of the options are ideal. For example, in the
quantitative phase of the BLSR study, Bowleg et al. (2004)
gave participants the option of using a five-point Likert-
type scale (1 = strongly agree, 5 = strongly disagree) to
indicate the extent to which they agreed or disagreed with
statements such as: “Racism is a much more serious issue
in my life than homophobia” and “Racism is a much more
serious issue in my life than sexism” (p. 234). In retrospect,
this approach seems farcical for all of the obvious and
previously stated reasons. Nonetheless, it has prompted me
to think how I would ask the question with another similar
study. I remain stumped. The simplest, albeit inadequate
approach appears to be the inherently additive check all that
apply option:

In the past year, would you say that you have
experienced stress as a result of discrimination due to
your race, sex, and/or sexual orientation? If so, please
indicate by checking all that apply below, the response

that best describes the basis for the discrimination you
experienced. Was it primarily because of your:

Race Sex Sexual orientation

Insights about Measuring Intersectionality

The BLSR (Bowleg et al. 2008, 2003) and TT (Bowleg,
manuscript in preparation) research experiences of Bowleg
et al. have yielded some clear insights about asking
questions intersectionality. The most obvious is that no
part of the question should even hint at addition. For
example, if I were to ask a question about day to day
challenges today, I would ask something like this: “Now,
I’d like you to tell me about some of the day-to-day
challenges that you face as a Black lesbian woman.” That
is, I would not use a phrase such as “race, gender and/or
sexual orientation” in which the presence of the conjunc-
tions and/or could imply that I wanted the experience
recounted serially (race, then gender, then sexual orienta-
tion) or that these identities could or should be separated.

Going forward, there are two key points to which
researchers should attend in constructing questions about
intersectionality. First, questions about intersectionality
should focus on meaningful constructs such as stress,
prejudice, discrimination rather than relying on demograph-
ic questions alone (Betancourt and Lopez 1993; Helms et
al. 2005; Weber and Parra-Medina 2003). Widespread
advocacy for the infusion of multicultural perspectives
within psychology (e.g., Sue et al. 1999) notwithstanding,
psychologists’ tendency for considering ethnic and racial
categories to be conceptually meaningful persists in much
research. The reality though is that concepts such as race
and class are socially constructed, and as such, explain
virtually nothing in and of themselves (Helms et al. 2005).
Thus, a study with an ethnic minority or ethnically diverse
sample that includes demographic measures of racial or
ethnic identification, socioeconomic status (SES), and
sexual orientation, for example, is not intersectionality
research de facto. By contrast, a similar study that focused
on the dimensions of experience (e.g., annual earnings,
access to health care, stress experiences, etc.) shaped by the
participants’ experiences of intersecting identities of racial or
ethnic identification, SES, and sexual orientation would
exemplify intersectional research. Second, questions should
be intersectional in design; that is they ought to tap the
interdependence and mutuality of identities rather than imply
as the BLSR (Bowleg et al. 2008, 2003) and TT (Bowleg,
manuscript in preparation) studies of Bowleg et al. did, that
identities are independent, separate, and able to be ranked.

My clarity on the aforementioned points notwithstand-
ing, there are other measurement issues with which I

316 Sex Roles (2008) 59:312–325

continue to grapple, however. For example, I am increas-
ingly agnostic about how much energy ought to be
expended on asking the right question to measure inter-
sectionality. Overzealous focus on designing the perfect
qualitative or quantitative question harkens back to positi-
vism’s ontological tenet that there is some single fixed
reality (see Tashakkori and Teddlie 1998) about intersec-
tionality that can be measured if only the researcher had just
the right question. Yet, as Nancy (Bowleg, manuscript in
preparation) and Leslie’s (Bowleg et al. 2003) different
answers to the same question demonstrate, there is no
single reality about the experience of one’s intersecting
identities, only multiple constructed realities about one’s
own experience of intersectionality. As for asking questions
about intersectionality in quantitative research, I question
whether the positivistic assumptions implicit in quantifica-
tion are compatible with intersectionality research. McGrath
and Johnson (2003) have captured the dilemma aptly:

Quantification imposes a very strong meaning system
on the information thus gathered—the meanings that
are implicit in various arithmetics and mathematics.
This, in turn, imposes many assumptions about sub-
stantive elements and relations (e.g., linearity, unidi-
mensionality) that go with the meaning system (p. 36).

Interdependence, multi-dimensionality and mutually
constitutive relationships form the core of intersectionality,
attributes that contradict the positivist assumptions inherent
in most quantitative approaches. Since I use both quantita-
tive and qualitative methods in my research, it should be
obvious that I have no interest in resurrecting that tired and
ultimately futile debate about the superiority of quantitative
versus qualitative methods. Rather, my argument here is
that the positivist paradigm that undergirds much (but not
all) quantitative research appears to be orthogonal to the
complexities of intersectionality. A researcher’s philosoph-
ical or “qualitative stance” (Marecek 2003, p. 49) exempli-
fied by an epistemological commitment to “situating …
investigations in specific historical, social, and cultural
contexts” (Marecek 2003, p. 56) is paramount; not whether
the questions they ask to measure intersectionality are
qualitative or quantitative.

Analyzing Intersectionality Data

The next step after asking questions about intersectionality
is to analyze the amassed data. Quantitative and qualitative
analysts face overlapping concerns. First, is the imposition
of the researchers’ philosophical paradigm and the relevant
assumptions with which the analyst approaches the data
(Baptiste 2001; McGrath and Johnson 2003); these will
“both shape and constrain the meaning(s) of the evidence”

(McGrath and Johnson 2003, p. 42). Second, is the
transformation of observations into data for …