This paper investigates the concept of internalized homophobia in both theory and research relating to lesbian and gay health. It offers a contemporary and critical review of research in this area, and discusses a range of recent findings relating to a range of health issues including HIV and AIDS. Whilst the concept has a resonance for gay men and lesbians, and is widely used in `lesbian and gay-affirmative' interventions, the paper demonstrates that research findings have been equivocal and the term is often used without full consideration of its sociopolitical consequences. The paper concludes that the concept does have a valuable role to play in health promotion work with lesbians and gay men but invites further discussion and examination of the construct.
This paper aims to present both an overview and critical evaluation of the usefulness of the concept of internalized homophobia1
The term `internalized homophobia' had been used throughout this article to represent the negative and distressing thoughts and feelings experienced by lesbians and gay men about their sexuality, and which are attributed to experiences of cultural heterosexism and victimization. For reasons discussed in the text, the author believes the term `homophobia' to be highly problematic; however, for the purposes of continuity and consistency with the majority of other publications on this issue, the term has been replicated here rather than using more accurate but obscure terminology.
Defining anti-gay and lesbian prejudice and its internalization
There has been considerable discussion within lesbian and gay academic circles about how best to conceptualize the nature of anti-gay and lesbian prejudice. Whilst the term `homophobia'2
Defined by Sprecher and McKinney (Sprecher and McKinney, 1993) as `negative and/or fearful attitudes about homosexuals or homosexuality'.
Used here as an inclusive term to include all social scientists working from an explicitly lesbian, gay, transgender or bisexual perspective.
See Logan (Logan, 1996) for a fuller discussion.
Alternatives that have been suggested include `homonegativism' (Hudson and Ricketts, 1980), which is a multidimensional construct that focuses more clearly on the belief and value systems of prejudiced individuals, and `heterosexism', which now features widely in gay and lesbian literature, and refers to an underlying belief that heterosexuality is the natural/normal/acceptable or superior form of sexuality.
In common with research into other forms of prejudice,5
For example, Clark and Clark's (Clark and Clark, 1958) study into young Black children's internalization of racial prejudice.
Shidlo (Shidlo, 1994) argues that the terms `internalized homonegativity' and `internalized homonegativism' are preferable to `internalized homophobia', but that these are conceptually similar enough to use interchangeably. The former of these has been used on occasion in this article.
Plumer [(Plumer, 1996), p. 89] also highlights the impact of stigma upon an evolving identity: `The awareness of stigma that surrounds homosexuality leads the experience to become an extremely negative one; shame and secrecy, silence and self-awareness, a strong sense of differentness—and of peculiarity—pervades the consciousness'.
The term has become widely used within lesbian and gay studies, especially within gay-affirmative psychotherapeutic models which understandably typically place the concept at the centre of explanations and interventions around mental health issues faced by lesbians and gay men. As a concept the idea of internalized homophobia strikes a chord within almost all gay men and lesbians, and a number of contemporary pieces of qualitative research provide evidence of consistent and coherent narratives of the phenomenon (Stokes and Peterson, 1998). Participants in Cody and Welch's ethnographic study of rural gay men frequently talked of having experienced intense feelings of shame and guilt. One participant had believed `I was the embodiment of all those nasty things that have been said about gay people' [(Cody and Welch, 1997), p. 60]. However, internalized homophobia has not proven to be such an easy concept to validate and operationalize as an acceptable theoretical and research-orientated concept, particularly for larger scale quantitatively oriented investigations. Wagner et al. [(Wagner et al., 1994), pp. 91–92] also discuss the lack of precision and reliability which has often accompanied research in this area. `Internalized homophobia as a psychological phenomenon has attracted little systematic research despite its destructive impact on the mental health of the gay community... As a result little is known about what predicts or precludes internalised homophobia.'
Validating and measuring internalized homophobia
As Shidlo argues in his seminal article on conceptual and empirical issues in measuring internalized homophobia [(Shildo, 1994), p.176], `the construct...can serve as a central organizing concept for a gay and lesbian affirmative psychology'. He suggests a number of major reasons for the significance of the construct which include the role played by internalized homophobia in psychological distress in lesbians and gay men, and the internalization of anti-gay and lesbian prejudice as a developmental event which is essential to understanding models of developing lesbian and gay identities. Internalized homophobia, he argues [(Shildo, 1994), p. 177], `can be a heuristic construct that organizes factors unique to lesbians and gay men in the areas of development, psychopathology, psychotherapy, and prevention'. Shidlo also argues that the generation of psychometrically efficient scales to measure levels of internalized homophobia are of use in examining the extent (and domains) of internalized prejudice, and therefore potentially assessing risk to the individual and evaluating the success of therapeutic or preventive interventions.
Few lesbian and gay academics, therapists or health professionals would dispute the importance of internalized homophobia. Indeed a number of valuable and sophisticated models exist which coherently outline the mechanisms and potential consequences of the internalization of anti-lesbian and gay oppression [e.g. (Bremner and Hillin, 1993)]. However, if approached from a more hard-nosed, empirical-based perspective, the conceptualization and operationalization of internalized homonegativity is less satisfactory. Traditionally, internalized homophobia was equated with an ego-dystonic7
Ego-dystonic homosexuality refers to an individual's persistent and profound distress with their lesbian or gay sexuality and is often associated with a desire to modify sexual orientation. The term has become particularly controversial after the inclusion of the term as a diagnostic classification within DSM-III (American Psychiatric Association, 1980). See Cabaj and Stein [(Cabaj and Stein, 1996), pp. 25–26] for more information. Bohan [(Bohan, 1996), p. 19] also offers a brief, lucid discussion of `ego-dystonia'.
The Nungesser Homosexuality Attitudes Instrument (NHAI) is lengthier, and comprises three subscales which are attitudes towards one own's sexuality, attitudes towards homosexuality per se and a disclosure scale which measures degree of comfort with others knowing one's gay or lesbian sexuality. Shidlo (Shildo, 1994) has revised the NHAI, removing or changing the least valid or most ambiguous items and adding a number of new items to one of the existing scales. He also suggests an optional extra 15-item subscale which measures internalized homonegativity in relation to issues around HIV and AIDS. Wagner et al. (Wagner et al., 1994) also combined nine items from the NHAI with 11 new ones to form their Internalized Homophobia Scale.
The NHAI has been shown to demonstrate convergent validity with Martin and Dean's scale (0.59, N = 159, P < 0.01) in a study by Sbordone (Sbordone, 1993) using a sample of gay fathers. Furthermore, research by Shidlo (Shidlo, 1987) using a sample of 59 found significant relationships with measures of a range of related concepts including self-esteem (–0.59, P < 0.01), depression (0.37, P < 0.01), stability of self (–0.35, P < 0.01) and loneliness (0.62, P < 0.001).8
Shidlo (Shidlo, 1994) offers a detailed overview of a number of studies which have evaluated the NHAI.
A more recent scale to measure internalized homophobia amongst gay men has been developed by Ross and Rosser (Ross and Rosser, 1996), and suggests four dimensions to the construct: public identification as being gay, perception of stigma associated with being gay, degree of `social comfort' with other gay men and beliefs regarding the religious or moral acceptability of homosexuality. Ross and Rosser (Ross and Rosser, 1996) report a series of significant associations for male participants between scores on the scale and a variety of other measures potentially relating to a healthy adjustment to a gay identity. These include affiliation to gay community groups, relationship satisfaction and duration, and disclosure in personal and work lives. The predictive validity of the public identification and social comfort subscales appears to be considerably stronger than for the other two. Overall this scale looks promising, but clearly needs further rigorous testing with regard to demonstrating psychometric credibility more fully.
Perhaps partly because research with lesbians has more typically adopted a feminist/qualitative paradigm, there appears to be no scales which have been developed for and widely used with female participants. Generally, scales have been validated on and typically include items which are directed towards the experience of urbanized, White gay men and may not adequately reflect heterogeneous experiences of being gay or lesbian.
Furthermore, establishing such measures and operationalizing internalized homophobia has proved difficult because of the considerable overlap with other relevant concepts (e.g. other aspects of self-esteem), and traditionally a lack of clear differentiation between internalized homophobia itself and certain intrapsychic or behavioural consequences (e.g. intimacy difficulties, depression, etc.). There is no clear consensus over the most salient aspects of homophobia and there may also be problems with scales which suggest that discomfort regarding disclosure is a valid measure of internalized homophobia. Whilst `coming out' is arguably the most salient and powerful process of developing a well integrated lesbian or gay identity,9
It would therefore appear that whilst these scales are clearly of potential interest and value to health and clinical psychologists, it remains highly questionable to what extent any of the scales can be considered sufficiently psychometrically robust and universally applicable for unqualified endorsement.
Nonetheless, the development of such scales allows for the measurement of internalized homophobia, and it becomes possible for social scientists to carry out quantitative analysis to add to the qualitative dimensions of assessing its role in explaining and preventing health difficulties.
Internalized homophobia as `minority stress'
A recent development in this area has been to conceptualize internalized homophobia as a component of minority stress. Following from the work of Brooks (Brooks, 1981) which conceptualizes minority stress as a psychosocial stress that is derived from membership of a low status minority group, studies by Meyer (Meyer, 1995) with gay men and DiPlacido (DiPlacido, 1998) with lesbians have found this to be a useful paradigm for the study of internalized homophobia and its relation to aspects of ill-health. In a detailed and articulate account, Meyer (Meyer, 1995) develops the conceptualization of minority stress within a social stress discourse, and drawing on aspects of conflict and societal reaction theories and social comparison and symbolic interactionist processes. He argues, `minority stress arises not only from negative events but from the totality of the minority person's experience in dominant society. At the centre of this experience is the incongruence between the minority person's culture, needs, and experience, and societal structures' [(Meyer, 1995), p. 35]. In Meyer's model, internalized homophobia represents one (although the `most insidious') of three aspects of minority stress and was operationalized through use of the Martin and Dean scale described above. The other dimensions of minority stress are perceived stigma and the experience of what he calls `prejudice events'. Meyer argues that each of these three aspects significantly impact upon psychological adjustment, but there is also an interaction which compounds what he calls the `psychologically-injurious effects'. Meyer provides evidence for his theory with a large-scale study of 741 gay men in New York recruited through a combination of network and snowball sampling techniques.10
It is worth noting that the majority of the sample (85%) were strongly affiliated to the `gay community' and open about their sexuality. As is a frequent problem with studies involving gay men, the sample here, although large and recruited through a number of sampling strategies, is most unlikely to be a good representation of the overall gay community.
DiPlacido's research on minority stress also emphasizes the role of internalized homophobia but her ongoing research is in many ways rather different to that of Meyer.11
This research is still being carried out. Discussion of the study is based upon the pilot study data published in 1998.
Internalized homophobia and HIV
Considerable amounts of research have investigated the role of internalized homophobia within HIV processes. These can largely be reduced into three areas—HIV prevention and safer sex decision-making processes, coping strategies of seropositive gay men, and whether internalized homophobia has any effect upon viral progression.
A relationship between internalized homophobia and riskier sexual acts seems logical for a number of reasons. Homonegative gay men are likely to be less affiliated with the gay community and may therefore have less access to safer sex information and resources. Furthermore, homonegativity correlates with lower self-esteem which may undermine the individual's desire to keep themselves safe. Finally, some studies have suggested that greater levels of homonegativity may be related to greater substance use and alcohol consumption (e.g. Finnegan and Cook, 1984, Glaus, 1988; Meyer and Dean, 1995) which may impair decision-making processes.
Despite widespread knowledge about the risks of HIV transmission, especially through unprotected receptive anal intercourse with multiple partners, rates of anal sex without condoms remain surprisingly high. Davies et al. (Davies et al., 1993) asked participants who enjoyed receptive anal sex to explain why. A small group of respondents explained that it helped reaffirm their gay identity. It is possible that this may apply to a greater extent to homonegative gay men who's sexual identity may be more fragile. In another qualitative study, Gold et al. (Gold et al., 1994) found that escapism is another motivating factor for engaging in (unprotected) anal intercourse. It may be that homonegative gay men feel a greater need for escapism than ego-syntonic13
Ego-syntonic—where the individual is accepting of lesbian or gay sexual identity. Being gay or lesbian represents a potential source of pride.
Interestingly, despite a range of potential theoretical explanations for a relationship between internalized homophobia and reduced condom usage, correlations in studies have typically been weak or inconsistent (Sandfort, 1995). Kippax et al. (Kippax et al., 1993) in their study of 535 Australian gay men found that two measures of gay community attachment (social engagement with other gay men and gay community involvement) were predictors of engaging in safer sex practices. Shidlo (Shidlo, 1994) found no relationship between internalized homophobia and engaging in anal sex without condoms, whilst Meyer and Dean (Meyer and Dean, 1995) found a small subgroup of gay men (6%) who appeared to be qualitatively different from their rest of the sample and who reported very high levels of internalized homophobia, substance use and participation in high-risk acts. A relationship between the two variables may not affect all gay men equally, but caution should be exercised before it is concluded that no relationship exists between homonegativity and high-risk sex. As Meyer and Dean (Meyer and Dean, 1998) suggest, other variables such as sexual difficulties and intimacy issues may interact with these processes. Furthermore, sampling techniques used in studies like these are likely to use gay-affirmative networks, commercial venues and community centres, and are most unlikely to recruit reasonable numbers of more closeted gay men who feel alienated from organized gay communities.
Other research has looked into seropositive men and how internalized homophobia might impact upon the selection and success of coping strategies.
Nicholson and Long (Nicholson and Long, 1990) in a study of 89 seropositive gay men correlated scores on the NHAI and a modified version of Folkman et al.'s Ways of Coping Scale (Folkman et al., 1986). High homonegativity was significantly related to avoidant coping strategies such as resignation or denial, whilst low negativity predicted proactive or active-behavioural coping strategies such as problem solving and resource seeking. As part of a longitudinal study in New York State, Wagner et al. (Wagner et al., 1996) asked a sample of gay men to fill in a battery of instruments including a modified version of NHAI, Lazarus and Folkman's (Lazarus and Folkman, 1984) version of the Ways of Coping scale and a number of measures of psychological distress. Participants14
Participants were heterogeneous in relation to HIV status, and included significant numbers of seronegative men, symptomatic and asymptomatic HIV-positive men, and a small number with an AIDS diagnosis.
Attrition rate of the sample was due to drop-out (N = 25) or death (N = 20). No participants had seroconverted during the two testing periods.
Research involving HIV and AIDS represents the largest body of knowledge for testing the relationship between conceptualizations and measures of internalized homophobia and illness. The research produced has had rather mixed results, suggesting that the concept may have some empirical value and predictive validity, but the relatively poor levels of reliability and replicability in studies, even allowing for difficulties with sampling strategies, should advise caution in those employing the construct in their research.
Internalized homophobia and other health issues
Some of the potential effects of internalized homophobia upon affect and psychological adjustment generally have been presented above. A number of other studies have also suggested that internalized homophobia may be a valid antecedent of a range of psychological problems. One area that has attracted particular interest is that of self-injurious behaviours including substance abuse, eating disorders, self mutilation and suicidality. A number of studies have shown the increased vulnerability of young lesbians and gay men to suicide generally (Remafedi et al., 1991), and research by Rofes (Rofes, 1983) amongst others suggests that internalized homonegativity may explain differences within lesbian and gay communities. Hammelman (Hammelman, 1993) found that young lesbians and gay men were at greater risk of attempted suicide if they discovered their same sex preference early in adolescence, experienced negative `coming out' reactions from significant others, experienced sexuality-orientated victimization, and used drugs and alcohol to cope with problems relating to their lesbian or gay identity. All of these findings are consistent with an internalized homophobia hypothesis. Teenagers who discover and disclose their sexuality earlier may be more isolated, cognitively embedded within heterosexist norms and values, and have less access to gay-affirmative organizations (e.g. at college, etc.) and individuals. Gonsiorek and Rudolph (Gonsiorek and Rudolph, 1991) refer to a narcissistic injury which is the significant blow to self-esteem that occurs when the individual is rejected through disclosing significant, personal information such as disclosing a lesbian or gay identity. Younger adolescents may be particularly vulnerable because they have less developed coping strategies and fewer coping resources—particularly given the overtly heterosexist culture of most secondary schools.16
Research also suggests that internalized homophobia may also be associated with more chronic forms of self-harm. Studies on alcoholism and substance abuse have already been mentioned. Research on gay men has indicated a strong relationship between measures of ego-dystonic gay sexuality and the NHAI with a number of measures of eating disturbance including the Eating Attitudes Test (Garner and Garfinkel, 1979) and Garner's (Garner, 1991) Eating Disorders Inventory #2 (Williamson and Hartley, 1998; Williamson, 1999). The relationship appears to be strongest with measures of bulimia and this may be consistent with a desire to punish the body for its same-sex urges. In America, work by Brown (Brown, 1987) has pointed to similar correlations amongst lesbian women.
Internalized homophobia may also affect health indirectly—especially when operating below consciousness. Work by Margolies et al. (Margolies et al., 1987) and Malyon (Malyon, 1982) suggest that internalized homophobia may affect intra-psychic functioning by generating various defence mechanisms. These may project themselves through difficulties with intimacy, commitment or other aspects of relationships. For example, a study by Rosser et al. (Rosser et al., 1997) reports significantly lower levels of sexual satisfaction in high scoring homonegative gay participants compared to low scorers. Equally, homonegativity may lead to the development of self-defeating personality traits which reflect internal representations of the stereotypical dysfunctional homosexual. These `secondary and tertiary adaptations' as Malyon labels them may have a profound impact upon the mental health of the individual and any interventions would need to be within a gay-affirmative therapeutic paradigm.
Internalized homophobia—balancing the `personal' and the `political'
Despite the widespread acceptance of internalized homophobia as a valid concept within lesbian and gay health and social scientific arenas, there are a number of concerns around an undiscriminating acceptance of the term as a research-orientated and theoretical paradigm. In addition to the conceptual and methodological difficulties raised earlier in this paper, there are also significant implications for lesbian and gay communities if we choose to explain and conceptualize health difficulties in this way. Kitzinger (Kitzinger, 1996, 1997) raises a number of objections to the concept of internalized homophobia. These include its emphasis on individual pathology rather than on institutional oppression. The danger of using terms like internalized homophobia or homonegativity is that being gay or lesbian is implicitly represented in pathological terms. `Instead of going to heterosexual therapists to be cured of our homosexuality, now lesbians and gay men are supposed to seek out lesbian and gay therapists to be cured of internalized homophobia' [(Kitzinger, 1997), p. 211]. The concept suggests weakness rather than the resilience demonstrated by lesbians and gay men, and keeps the focus away from the structures of inequality and oppression.17
Kitzinger goes on to argue that distress in the face of oppression is perfectly reasonable and that to see such `forms of unhappiness' as internalized homophobia (i.e. an example of individual pathology) is inaccurate. Because `homophobia' strongly suggests pathology, homonegativity again appears to be a preferable alternative.
Furthermore there is a clear need to balance interventions on an individual basis with more collective social action. There is clearly a need in most countries for specific legislation which addresses the civil rights of lesbian and gay peoples, and which addresses `hate crimes' specifically. In Britain, the repeal of Section 28 of the 1988 Local Government Act, which prohibits the `intentional promotion of homosexuality' in Local Education Authority run schools, is essential for teachers and school counsellors to feel free to fully support lesbian and gay students without fear of sanction.
Whilst cultural and institutional heterosexism remain largely unchallenged, it is difficult to provide nurturing and supportive environments for all lesbian and gay people, and particularly those who are particularly anxious, distressed or confused about their sexual identities.
Fortunately, there does appear to be an increase in the social/support infrastructure available to (particularly younger) lesbians and gay men in many areas. Cody and Welch's study (Cody and Welch, 1997) demonstrated the importance of social and community groups in working through issues around internalized homophobia and constructing `families of choice'. One participant in the study states; `The support group helped me to feel better about being gay...being happy and gay is not an oxymoron. You can have both' [(Cody and Welch, 1997), p. 62]. Interestingly a sense of community was achieved by men in this study through groups, friendship networks and particular partners rather than through participation in the commercial gay `scene' in neighbouring towns and cities. This may be particularly valid for women for whom there is often a dearth of commercial venues which are not (gay) male dominated. Such groups appear to often play a vital role in providing accurate information, discussing salient issues and working on skills (e.g. assertiveness, negotiating safer sex, etc.) and strategies both formally and informally within a safe and explicitly gay-affirmative environment.
This paper has aimed to provide a wide-ranging discussion of the role of internalized homophobia with health-related problems amongst lesbians and gay men. I have provided evidence that suggests that homonegativity merits consideration as a predisposing and perpetuating factor in many aspects of ill-health. As a concept, it may help to identify particularly vulnerable and at risk individuals, and should be considered in health education and disease prevention models. Internalized homophobia may relate to coping strategies and a willingness to access certain coping resources. There is as yet no clear evidence that homophobia impacts directly on the progression of illnesses, although it may interact with other factors (e.g. coping strategies) to produce important health-related consequences. I hope that the paper has re-emphasized the need for gay-affirmative models of health care and intervention. Whilst there is a valid explanatory role for internalized homophobia, there remains a number of integral inadequacies and inconsistencies in how the concept is defined and operationalized, and there is a significant need to refine the concept and improve the way that it is assessed, both qualitatively and quantitatively. Furthermore, there is also a need to be aware of the potentially harmful role of using the concept imprudently in terms of feeding into heterosexist pathological models and in underestimating both the heterogeneity and resilience of lesbian and gay communities. At worst, internalized homophobia represents a catch-all pseudo-explanation which colludes with anti-gay and lesbian models of ill-health. As social scientists and health professionals, we need to ensure that this does not happen but rather that the concept is used judiciously and helps to further develop models which validate the experiences of lesbians and gay men and provide adequate interventions to health problems.