Intersectionalized identities: sexual and racial minorities and their related mental health risks


Jeramy Townsley

Dec 2009


Minority status brings with it several layers of social inequality. At the macro level, we see economic, political, legal, and status stratification, along with various forms of informal discrimination. At the micro level we see stigmatized identities and stigmatized families, especially when those families fail to conform to the expectations of the majority. Whether the minority statuses are religious minorities (such as the recent media attention to polygamous Mormon families), racial minorities, or sexual minorities, hegemonic structures marginalize all who do not meet idealized, externally imposed standards.

Recent work has highlighted the influence of overlapping stigmatized identities that compound the inequalities suffered by these individuals. Collins (2005) describes the matrix of domination which traps individuals and groups into interlocking systems of oppression. These forms of oppression are experienced to a greater degree when individuals lay at the crossroads of multiple subordinated identities, or intersectionality. Not only do such individuals face stigma from the majority, but from other minorities as well, as those minorities accept hegemonic schema and reproduce them onto other minority groups (Battle 2002). In exploring intersectionality and marginalization, Collins describes this often overlooked cross-stigmatization between minority groups, resulting in few sites for cross-minority organization and resistance (2005).

Evidence of stigmatization can be found in several areas of social and personal life. Data on hate crimes, income and hiring inequality, and housing discrimination are classic and robust measures of discrimination based on race, class and gender (Massey 2007). Other measures look at individualized effects, like evidence of stress and personal adjustment difficulties. Sexual minorities face significantly increased risks of drug abuse, mental illness and suicide attempts (Udry 2002; Fergusson 1999) that typically are not seen in racial minorities, problematizing simplistic models of emotional vulnerability caused solely by placement into “minority” status. This risk for sexual minorities is even greater for bisexuals than for those who identify as gay/lesbian (Paul 2002; Austin 2008), presumably since there are few norms for the bisexual identity, and they are stigmatized by both the straight and the gay/lesbian communities. Some sexual minority communities have begun to be more inclusive, with growing acronyms which started as GL (gay/lesbian) and have grown to GLBTQI (gay/lesbian/bisexual/transgender/ queer/intersex). However, much of this inclusiveness has thus far been merely symbolic, as evidenced by the gay/lesbian communities’ lack of political support for “queer” legalities (polyamorous families, trans-inclusive language in rights bills, discussions about intersex, etc) as well as stereotypes about the bisexual identity: “they are more promiscuous,” “they are just closeted gays,” “they can’t make up their minds” (Henderson 2009). These patterns make the bisexual more vulnerable among the sexual minorities than the historically more empowered gay and lesbian identities.

The existence of the bisexual identity problematizes simplistic binary categories of sex and gender. For the essentialists, who propose the biological basis for a fixed male/female dichotomy, the bisexual (or traditionally, any non-heterosexual identity) represents a medical or psychiatric pathology. For the constructivists, who propose more dynamic, socialized sexual identities, the bisexual represents a variation on a multitude of divergent sexualities (Seidman 2003; Weeks 2003; Weinberg 1994). Queer theory draws from constructivist and feminist models of sexuality, challenging binary hegemonic norms of both sex and gender. The constructivist model helps us understand the ways our identities are continuously shaped and reshaped throughout the life course. The creative construction by the individual of our sex and gender identities (agential), framed by structural constraints provides a fertile source of sociological exploration of the interplay between these two forces, given the profound changes in the past century to these identities. It also provides us a model for understanding why identities can become problematic for individuals when faced with erotic and affectionate feelings towards same and opposite sex individuals that are not socially sanctioned, and how the gay, straight or bisexual (or other) identities fluidly and chaotically become social and personal constructions, not fixed identities in nature. Further, critical constructivism adds the dimension of conflicts between groups in power (majority) and the rest of society (minorities), integrating that analysis with the socially constructed nature of these conflicts, intertwining classic conflict theory with symbolic interactionism (Heiner 2006).

Sexual orientations can have multiple components, including factors such as attraction, behavior, fantasy and identity, none of which necessarily imply the other (Sell 1997). For example, an individual may self-identity as “heterosexual” yet have same-sex attractions or even behavior. Conversely, an individual may self-identify as gay yet may be in an heterosexual marriage, staying in the marriage for cultural reasons (status, children, etc). Both of these patterns may be termed being “closeted” or on the “down-low” (Collins 2005). Others may experience same-sex attractions and fantasies, but do not desire them, and may either self-identify as heterosexual, or even “ex-gay,” similar to the 12-step model of identifying as a member of your “addiction” identity, with the assumption that persistent, unwanted sexual feelings represent pathology similar to cravings for alcohol (Erzen 2006). In any of these situations, adaptational difficulties may occur, both social, and personal, each of which may be dependent on the cultural meanings of these factors. One community may value individual authenticity over conforming to socially-approved sexual identities. In this case, an individual having same-sex behavior but “passing” as straight may induce a crisis of identity that surpasses the stigmatizing effect of identifying as gay (Kroeger 2003). Another community may value traditional sexual roles, to the extent that having same-sex behaviors and attractions must be repressed, and such repression may not necessarily lead to the same level of emotional challenge as in the former community.

Based on Collins’ studies on intersectionality, it is reasonable to expect that individuals who fit into overlapping stigmatized identities will suffer greater risk than those who just fit into one minority identity group. The purpose of this study is first to seek further evidence that the intersection of stigmatized identities (female, non-white, poverty, no college, sexuality) puts those individuals at greater risk for emotional vulnerability. Second, it will be shown that not only is sexual identity a risk factor for emotional vulnerability, but that individual sub-components of sexual orientations (attraction, behavior, identity) are themselves risk factors. Finally, this paper will explore the proposal that the different sub-components of sexualities put individuals at differential risk, depending on cultural factors.



This study utilizes data from the National Longitudinal Study of Adolescent Health (Add Health). Add health is a stratified sampling of high school students from 1992-2002, using a random sample from all high schools in the U.S. Wave 1 was collected from students in grades 7-12 from 1994-1995. Wave 3, which is used here, samples that same cohort from 2001-2002. While longitudinal analysis is possible with this data, the measures important for this study require only data collected from the most recent time-frame in order to measure current feelings about their social identity regarding sexual orientation, and to contrast that with mental health measures. Wave 3 Add Health data surveys young adults age 18-26, n=15,170 (Rostovsky et al. 2008). Several methods of data collection were used, including face-to-face interview, paper and pencil interview, and audio computer-assisted self interview (ICPSR Web site 2009). The particular dataset used for this study (Dataset 12, Wave 3, Public In Home Questionnaire) has a population of 4882, and most of the questions incorporated here have response rates above 90% (questions about income were limited to 3835 participants).


Sexual orientation

Wave III Add Health asks three questions related to sexual orientation (Rostosky, Danner, Riggle 2008; Udry, Chantala 2002; Pearson, Muller, Wilkinson 2007; Marshall, Friedman, Stall, Thompson 2009; See also Sell 1997 for a description of construct validity of the various measures for sexual orientation. Separating fantasy, attraction, identity and behavior is similar to the Klein scale, which expands on the simplistic binary of the older Kinsey scale). First, the sex to whom the individual is attracted. Both males and females were asked the following questions: “Have you ever had a romantic attraction to a female?” and “Have you ever had a romantic attraction to a male?” The response options were “No,” “Yes,” Refused,” “Don’t Know,” and “Not Applicable.” The second measure was specifically about identity: “Please choose the description that best fits how you think about yourself,” with the response options ranging from “1 100% heterosexual (straight)” (89%), to “3 Bisexual-attracted to men and women equally” (1.5%), to “5 “100% homosexual (gay)” (0.6%), and “6 Not sexually attracted to either males or females” (0.5%). Other options were “Refused” (0.4%), “Don’t Know” (0.2%), and “Not Applicable” (0.3%). For this study, both for theoretical considerations (any “non-bipolar” sexual identity was tested for significance, not the degree of bipolarity) and for practicality (the sample of non-heterosexuals was small), only the extremes on either side are used to identify “heterosexual” and “homosexual,” and all intermediary responses are considered “bisexual.” The third question asked about behavior, with the individual being able to list the sex of each of his/her past sexual partners. For this study, having one or more same-sex partner is used as a measure of bisexuality, unless no partners were opposite-sex, which is used as a measure of strict homosexuality.

Mental Health

Items measuring mental health were constructed for the Add Health study and were adapted from widely used measures (Guterman, Hahm, Cameron 2002). Six specific Add Health measures are used here to assess mental health: thoughts about suicide in the previous 12 months, levels of satisfaction with life, frequency of crying in the previous 12 months, participation in a drug/alcohol program in the previous 12 months, participating in counseling or therapy in the previous 12 months, or having been in a mental hospital in the previous 12 months. First, Add Health asks “During the past 12 months, have you ever seriously thought about committing suicide?” with the response options of “No” (91%), and “Yes” (6.2%). A second measure asks, “In the past five years, have you spent a day or more in a facility where you were treated for a mental illness,” with response options of “No” (97.9%), “Yes” (2.1%). Two other measures ask about alcohol and other substance use. The first asks about alcohol use frequency in the past 12 months, with response options ranging from “0 None” to “6 Every day/almost every day.” A second if the responded has used illegal drugs not prescribed to the respondent, giving examples such as “LSD, PCP, ecstasy, mushrooms, inhalants, ice, heroin…” with response options of “No” (82.5%) and “Yes” (1.6%). The final measure for mental health status asks “You were sad, during the past seven days,” with response ranges from “0 Never/rarely” (59.1%) to “3 Most of the time/all the time” (1.6%).

Factor analysis (in SPSS) using principal component analysis against sexual orientation (attraction + behavior + identity) indicated a relationship between several components that were used to create two factors, as indicated in Table 1. Factor one is comprised of thoughts of suicide, life satisfaction, and frequency of crying, and seems to be measuring general mental health challenge. Factor two is comprised of three questions about recent therapy, whether for general counseling, therapy for drug/alcohol problems, or residence in a mental health hospital, and seems to be measuring interaction with mental health institutions. Together these factors explain 45% of the variance, representing the only factors with an eigenvalue greater than 1.0.

Table 1: Factor Loadings for Mental Health Measures Using Varimax Rotation




Thought about suicide (past 12 months)



Counseling or therapy (past 12 months)



Drug-Alcohol Program (past 12 months)



Mental Hospital (past 5 years)



Frequency of Crying (past 12 months)



Life Satisfaction vs. Dissatisfaction




Descriptive Results

2:01 PM 12/16/2009 Table 2 describes the major variables used in this study. The data represents results from 18-28 yr old participants, with the largest cluster of subjects in their early 20s (21-23 year olds comprise 51% of the sample). This limits generalizability of the outcomes to individuals under 30.

Men and women are fairly equally distributed in the sample, with sex presented as a binary forced-choice (“gender”). Education levels are quite high (over 55% in college or graduate school) in a full national sample, but similar to census data from this age group. In contrast, income (median=$10,000) is quite low compared to what might be expected from similar education levels, most likely indicative of the young age of the sample. Race is fairly well distributed, representing general population characteristics. For this study, the category “multiracial” does not include individuals who only identified in any single-race category. Sexual orientation in Table 2 is defined by a composite measure of homosexual or bisexual measures of attraction, identity or behavior, and is fairly equally distributed between most demographic groups.

Table 2 gives a summarized view of these factors for all demographic groups, while Table 3 gives specific correlations for each measure and factor for specific demographic risk groups. Table 4 summarizes the factors as they relate to sexual orientations.

Analytic Results

Table 3 indicates the relationships between risk demographic data and mental health indicators. Each of the five risk demographics are listed individually, as well as an intersection composite measure, which is simply an additive factor depending on how many risk factors a subject has: being non-white (Hispanic, Black, Asian or Native American), female, education level, income level, or GLB orientation. Income and education levels are inverted from each of the other measures—whereas being female, non-white and GLB identity represent high stigma, high income and education represent low stigma. Inverse correlation for these measures, therefore, contributes to higher risk intersection correlation.

Table 2: Descriptive Statistics


Bisexual or homosexual Orientation

N (% of row)

Self-Identified Sex



249 (11%)



493 (19%)

Age Group



215 (16%)



385 (15%)



142 (14%)


(by grade completed)

(some high school)


120 (19%)

(some college)


458 (15%)

(advanced college)


150 (14%)



12 (12%)

Income bracket



375 (16%)



176 (16%)



32 (9%)



3 (11%)



0 (0%)

Self-identified race



6 (16%)



420 (15%)



168 (15%)

Native American


8 (20%)



24 (14%)



114 (18%)

Table 3: Intersection of Various Risk Factors

Factor 1 Total

Factor 2 Total













GLB Orientation



Intersection of each of the above



** Correlation is significant at the 0.01 level (2-tailed); *. Correlation is significant at the 0.05 level (2-tailed).

Each of the risk-demographic measures is correlated to at least some of the mental health indicators. The strongest individual relationships are GLB orientation, low education, and poverty, while non-white and female statuses have the least number of positive relationships. Factor one, measures of personal unhappiness, have the strongest association with the risk demographic measures. Factor two, measures of utilization of institutional services, evidences relationships with GLB orientation and low education, but strong negative relationship with non-whites. As expected, the intersection of multiple stigmatized identities is generally related to greater mental health risks, especially when one considers just personal experience and not utilization of professional mental health services.

Table 4 describes the mental health risks of race and income compared to specific GLB sexualities. GLB identity is separated from both attractions and behavior, with the assumption that self-labeling with a stigmatized identity has social and personal consequences distinct from emotions and/or specific (potentially isolated or discreet) sexual relationships. Moreover, these consequences may be distinct for different cultures, even within the same nation-state. This is evidenced by the differences in mental health effects between blacks and white, wealthy and poor, depending on whether the individual self-labeled as GLB versus simply engages in GLB behaviors or has GLB attractions. Initial analyses (prior to an income analysis) indicated that both races had high risk of emotional vulnerability and utilization of mental health services. However, given the interactions between race and class can have differential effects, an analysis that separated class gives a more nuanced pattern. While all groups have high utilization of mental health services (except poor blacks and the groups without sufficient data), race and class played significant effects in emotional vulnerability, with the only heterosexual identifying group having emotional difficulty are wealthy whites—the others do not claim such difficulties. Of those who claim a queer identity, class plays an even more important role, reversing racial trends: poor blacks and wealthy whites do not claim emotional challenges, while wealthy blacks and poor whites claim significant distress.

Table 4: Mental Health versus the Intersection of Race, Class and Sexual Orientations

GLB Identity

GLB Attractions or Behavior

(Heterosexual Identity)













Cried a lot, life satisfaction or suicidal thoughts









Treated for mental illness, drug abuse or other counseling









** Correlation is significant at the 0.01 level (2-tailed); N/A-not enough data to calculate significance


The data provides further evidence of Collins’ intersectionality theory. Not only do individuals in stigmatized groups suffer greater emotional vulnerability, but the effects are additive—the more stigmatized groups you are in, the greater the challenges. Similarly, non-heteronormative factors, such as attraction, behavior, or identity, put those individuals at greater risk for emotional vulnerability. However, the effects differ by race and class, presumably because of the cultural meanings of queer identities. While both blacks and whites experienced emotional distress from self-labeling with the stigmatizing identity of “queer” (bisexual or gay), class culture seems to modify racial experiences.

Several social theorists discuss the importance of identity in modern/postmodern societies: Habermas (1974), in discussions about communicative rationality and personal meaning; Elias (1939), when discussing “the we-less I” of modern institutions (1939); Giddens (1991), in the reflexive projects of self; Castells (1997) and Touraine (1997), in discussing “New Social Movements” based around a search for identity; Bauman (1992), who describes how a loss of social anchoring points causes identity crises that force us to focus identity construction around our own bodies; and Gergen (1991), when discussing the saturation and fracturing of the self/selves. Within this process of constructing/negotiating social and personal identities, it often becomes necessary to deal with the tension between conflicting cultural values. Two possible values at work here are the tension between commitment to traditional values and community on the one hand, and authenticity and individualism on the other.

In the latter, society has constructed various, optional “selves” from which we can choose to identify. If we “feel” like we “are” a particular identity, yet flee from it, we may find ourselves in a situation similar to the Judeo-Christian Jonah, fleeing from a destiny only to be spat back out onto that destiny by a big fish (in this case, the complex interaction between social identities and our perceived emotional selves). Cultures that value authenticity and individualism over community may devalue individuals who simply perform social roles at the expense of their “true selves.” Even if that culture stigmatizes a particular identity, such as gay/bisexual identity, contemporary middle/upper class culture seems to value the individualistic quest for authenticity over conformity to traditional sexual roles. On the other hand, commitment to duty and tradition may lead a person to resist stigmatizing labels in order to appear to be committed to some other identity, like a religious, or racial identity, much as the “ex-gays” mentioned above may deny being gay, while still have same-sex attractions (Erzen 2006). While this “passing” itself may cause stress, a strong sense of identification to a different cultural identity may provide protective effects greater than dissonance related to identity-feeling conflicts (e.g., identifying as straight or ex-gay, but having same-sex attractions or engaging in same-sex behaviors).

In Distinction, Bourdieu analyzed Parisian culture, considering not only class and cultural status of individual groups, but also directionality—whether that group was increasing or decreasing both economically and in status (Bourdieu 1984). Understanding the differential effects of class on racial emotional challenge may be facilitated by utilizing insights from Distinction. Bourdieu found that while political class has some relationship to political choices and moral reasoning, the much more robust predictor is the history and trajectory of that group. Groups with threatened status, or with a tenuous history of growth, lean towards conservatism and traditional values. On the other hand, established groups, especially of the better educated, tend towards being open to social change. Applying those insights to the data above, the black middle class in the U.S. is a fairly new group, and the difference between white and black wealth is still tremendously high: “roughly three out of four white households own their own homes, while the same is true for less than half of all blacks households, … and the racial difference in homeownership has widened notably … between 1985 and 2005” (Hilber and Liu 2008).

Assuming Bourdieu is applicable to U.S. racial culture, it might be expected that members of the black middle/upper class might stigmatize queer identities at greater rates than the white middle/upper class, leading to greater mental health challenges of black middle/upper class queer individuals, while, in contrast white middle/upper class individuals may be encouraged to identify with non-traditional sexual identities, and thus have better mental health outcomes for identifying as queer. In contrast, those same middle/upper class whites who engage in same-sex behavior or have same-sex attractions yet deny a queer identity may suffer cognitive dissonance greater than the stabilizing effect of conforming to group identity. Black middle/upper class individuals who engage in same-sex behaviors or have same-sex attractions, but identify as heterosexual, would receive the full psycho-social benefit of conforming to group identity and expectations, despite hidden attractions or behaviors.

Poor whites may be influenced by the same class and status factors that affect middle/upper class blacks—tenuous and threatened position. Therefore, leaning more towards conservative political and moral positions, they would tend to reject non-traditional sexual identities, affirming heterosexual identities and stigmatizing queer identities. As expected, they show similar mental health challenge patterns regarding queer or heterosexual identification. The phenomenon that remains unexplained is the lack of emotional difficulties of queer poor blacks. While this group seeks professional mental health services (contrasted with “heterosexual” poor blacks who have same-sex sex or attractions, who do not tend to utilize such services), they do not seem to suffer the same ill effects as their white counterparts. Neither Collins’ intersectionality, nor Bourdieu’s class trajectory analysis seems to explain this final category. While one could propose that the intersectionality of poverty and race creates such high levels of stress that it overshadows any increases in stress caused by queer identity. However, as indicated in Table 3, race itself is not highly correlated with negative mental health outcomes within this sample. It is possible that whatever protective factors are found within this larger group may extend to those members with a queer identity. Further study may provide insight into these racially distinctive protective factors.


Austin, Bryn, Andrea Roberts, Heather Corliss, Beth Molnar. 2008. “Sexual Violence Victimization History and Sexual Risk Indicators in a Community-Based Urban Cohort of ‘Mostly Heterosexual’ and Heterosexual Young Women.” American Journal of Public Health 95: 1015-1020.

Battle, Juan, Cathy Cohen, Dorian Warren, Gerard Fergerson, Suzette Audem. 2002. Say it Loud: I’m Black and I’m Proud; Black Pride Survey 2000. New York: The Policy Institute of the National Gay and Lesbian Task Force.

Bauman, Zygmunt. 1992. Intimations of Postmodernity. New York: Routledge.

Bourdieu, Pierre. 1984. Distinction . Cambridge: Harvard University Press.

Castells, Manuel. 1997. The Power of Identity. Oxford: Blackwell Publ.

Collins, Patricia Hill. 2005. Black Sexual Politics: African Americans, Gender and the New Racism. New York: Routledge.

Elias, Norbert. 1939. The Civilizing Process. Oxford: Blackwell

Erzen, Tanya. 2006. Straight to Jesus: Sexual and Christian conversions in the ex-gay movement. Berkeley: University of California Press.

Fergusson, David. 1999. “Is Sexual Orientation Related to Mental Health Problems and Suicidality in Young People”? Archives of General Psychiatry 56: 876-880.

Gergen, Kenneth. 1991. The Saturated Self. New York: BasicBooks.

Giddens, Anthony. 1991. Modernity and Self-Identity. Stanford: Stanford Univ Press.

Guterman, Neil, Hyeouk Hahm, Mark Cameron. 2002. “Adolescent Victimization and Subsequent Use of Mental Health Counseling Services.” Journal of Adolescent Health 30: 336-345.

Habermas, Jurgen. 1974. "On Social Identity." Telos 19: 90-103.

Heiner, Robert. 2006. Social Problems: Introduction to Critical Constructionism. New York: Oxford University Press.

Henderson, Loren. 2009. “Between the Two” Bisexual Identity Among African Americans.” Journal of African American Studies 13: 263-282.

Hilber, Christian, and Yingchun Liu. 2008. “Explaining the Black-White Homeownerhsip Gap: the Role of Wealth, Parental Externalities and Locational Preferences.” Journal of Housing Economics 17:152-74.

ICPSR Web site ( accessed Oct 26, 2009.

Kroeger, Brooke. 2003. Passing: When People Can’t Be Who they Are. Cambridge: PublicAffairs.

Marshall, Michael, Mark Friedman, Ron Stall, Amanda Thompson. 2009. “Individual Trajectories of Substance use in Lesbian, Gay and Bisexual Youth and Heterosexual Youth.” Addiction 104: 974-981.

Massey, Douglas. 2007. Categorically Unequal. New York: Russell Sage Foundation.

Paul, Jay, Joseph Catania, lance Pollack, Judith Moskowitz, Jesse Canchola, Thomas Mills, Diane Binson, Ron Stall. 2002. “Suicide Attempts Among Gay and Bisexual Men: Lifetime Prevalence and Antecedents.” American Journal of Public Health 92: 1338-1345.

Pearson, Jennifer, Chandra Muller, Lindsey Wilkinson. 2007. Adolescent Same-Sex Attraction and Academic Outcomes. Social Problems 54: 523-42.

Rostovsky, Sharon, Fred Danner, Ellen Riggle. 2008. “Religiosity and Alcohol Use in Sexual Minority and Heterosexual Youth and Young Adults.” Journal of Youth and Adolescence 37: 552-563.

Seidman, Steven. 2003. The Social Construction of Sexuality. New York City: Norton.

Sell, Randall. 1997. “Defining and Measuring Sexual Orientation: A Review.” Archives of Sexual Behavior 26: 643-658.

Touraine, Allain. 1997. What is democracy? Trans. David Macey. Boulder: Westview Press.

Udry, Richard, Kim Chantala. 2002. “Risk Assessment of Adolescents with Same-Sex Relationships.” Journal of Adolescent Health 31: 84-92.

Walls, Eugene, Stacey Freedenthal, Hope Wisneski. 2008. “Suicidal Ideation and Attempts Among Sexual Minority Youths Receiving Social Services.” Social Work 53: 21-29.

Weeks, Jeffrey. 2003. Sexuality, 2nd ed. New York: Routledge.

Weinberg, Martin, Colin Williams, Douglas Pryor. 1994. Dual Attraction: Understanding Bisexuality. New York City: Oxford University Press.

Wheeler, Darrell. 2003. “Methodological Issues in Conducting Community-Based Health and Social Services Research Among Urban Black and African American LGBT Populations.” Journal of Gay and Lesbian Social Services 15: 65-78.

Return to Jeramy's Homepage

Feel free to e-mail me at

"Open your eyes. Don't let your mind tell the story here." Tonic, 1996

"Our lies have made us angry with the truth." Five O'Clock People, 1997 Intersectionalized Identities 1