Internalized Homonegativity: A Systematic Mapping Review of Empirical Research

ABSTRACT

Internalized homonegativity (IH) is an important variable affecting the wellbeing of lesbian, gay, and bisexual (LGB) persons. We included 201 studies in a systematic mapping review of IH. Most studies were conducted in North America and examined IH as a predictor of poor health. The primary focus of 14 studies was IH scale measurement, and, in total, these studies detailed nine distinct scales. Eighteen studies compared levels of IH in LGB populations, four described prevention programs, and one investigated IH using qualitative methods. Our review indicates that further research is needed, particularly qualitative research and ways to ameliorate IH.

Legal restrictions on same-sex relationships, hate crimes against gays and lesbians, bullying of gay teens, and familial and social rejection of homosexuals are examples of heterosexism that denigrate and devalue non-heterosexual forms of identities, behaviors, relationships, and communities. Internalized homonegativity (IH) refers to the process whereby lesbian, gay, and bisexual (LGB) persons internalize societal messages toward gender and sex—often unconsciously—as part of their self-image (Meyer, 1995). Herek (2000) argued that this incorporation can result in negative feelings toward oneself when a person recognizes his or her own homosexuality or bisexuality. This is because negative internalized beliefs create a psychological dilemma between romantic desires and negative beliefs about the self; the disjuncture can lead to feelings of guilt and shame, low self-esteem, and other emotional difficulties (Herek, 2007; Malyon, 1982; Meyer & Dean, 1998; Shidlo, 1994; Weinberg, 1973).

The concept is sometimes referred to as internalized homophobia. However, Herek (1994, 1995, 2004) has criticized conceptualizations of IH that are focused too narrowly on fear and avoidance of LGB persons and informed by a clinical language that pathologizes and stigmatizes particular identities and attitudes. Herek (2004) has commented that research has failed to detect fear and anxiety responses when heterosexuals view photographs of men having sex with men. Instead, disgust and anger appear to be central to heterosexual people’s negative responses and attitudes to LGB persons. It is important to recognize, as Hudson and Ricketts (1980), Mayfield (2001) and Szymanski and Carr (2008) have done, that internalized homonegativity is shaped by wider societal factors and is not simply a product of personal, subjective, and “irrational” fears. As Russell and Bohan (2006) and Herek (2007) have reasoned, IH is not an inherent personal response from individuals but is instead the product of social and political stigma and bias.

Thus, for example, Russell (2007) encouraged therapists to see IH as the result of a process of social and individual exchange and not as the result of a personal characteristic or trait. The social context of IH is emphasized, too, in Meyer’s (2003, 2007) theory of minority stress, which provides a coherent explanation for the impact of homonegativity, or heterosexism, on LGB people, arguing that the heightened vulnerability of LGB persons originates from their exposure to stressors that are unique to their minority status. Meyer (2003, 2007) identified five such stressor types (experiencing prejudicial events, expectations of rejection, hiding and concealing one’s sexual orientation, internalized homonegativity, and ameliorative coping processes), based on their proximity to the self: the internalization of societal heterosexist attitudes is seen as a key proximal stressor. Because this subjective stressor is formed in cultural contexts that are sometimes characterized by extreme stigma toward same-sex romantic, emotional, and sexual behavior, Meyer (2003) noted that IH is unique, chronic, and socially based. Inevitably, most LGB people are likely to have some level of IH related to their status as members of a stigmatized group.

Weinberg (1973) and Malyon (1982) are credited with initial systematic conceptualizations of IH. Subsequent influential IH works have maintained IH as a central clinical theme in psychotherapy with LGB clients (Davies, 1996; Shidlo, 1994). Carballo-Dieguez (1998) was among the first researchers to expand concerns about IH to include considerations of gender roles, doing so in the context of Latin America and particularly related to IH’s effects on HIV prevention. Other narrative reviews of IH have emphasized its potential negative impacts on health, including, for example, its effects on body image of lesbians (Pitman, 1999), on drug abuse and HIV (Williamson, 2000), on violence (Balsam, 2001), and on substance use (Brubaker, Garrett, & Dew, 2009). Newcomb and Mustanski’s (2010, 2011) meta-analytic reviews on the relationship between IH and mental health and between IH and sexual risk taking found largely small effect sizes for these associations. Recently, Szymanski and collaborators (Szymanski, Kashubeck-West, & Meyer, 2008a, 2008b) examined the psychometric properties of instruments for measuring IH. They concluded that there were at least five measures with adequate reliability and validity support, but that more research was needed related to the measurement of IH, particularly with regard to racial and ethnic minority LGB individuals.

Although much has been written about IH, the scope of the empirical work that has been undertaken on this subject so far remains unclear, and there is a need to establish a heuristic frame to maneuver through a rapidly expanding field of research on the issue. The purpose of this review, therefore, is to document the current status of empirical research on IH.

Methods

When conducting this systematic mapping review, we employed the methods and procedures developed by the Evidence for Policy and Practice Information and Co-ordinating Centre (Bates, Clapton, & Coren, 2007; Clapton, Rutter, & Sharif, 2009; Oakley, Gough, Oliver, & Thomas, 2005; Peersman, 1996). We also followed criteria applied to systematic reviews in searching for, identifying, and describing research (Petrosino, Borouch, Soydan, Duggan, & Sanchez-Meca, 2001; Petticrew & Roberts, 2006).

While a systematic mapping review is a useful product in its own right and describes the kinds of research that have been undertaken within a particular field of study, it also provides an overview of a research area, highlighting areas in which empirical research has been conducted and aiding the identification of knowledge gaps. Additionally, such reviews enable researchers to focus on one particular area of the map, identifying worthwhile parallel or subsequent in-depth review(s) with narrower criteria for inclusion. Thus a systematic mapping review provides a wider research context within which to interpret the findings of evidence syntheses (Bates et al., 2007; Clapton et al., 2009; Oakley et al., 2005; Shepherd et al., 2005).

We conducted a comprehensive and systematic search for relevant literature in seven international, commercially available, electronic databases: Embase, ISI Web of Knowledge, MEDLINE, PsycINFO, PubMed, Social Services Abstracts, and Sociological Abstracts. The search strategies were prepared and piloted by an information search specialist at the Norwegian Knowledge Center for the Health Services, in cooperation with one reviewer (RB), and executed in November 2012. The searches incorporated text words (in titles and abstracts), such as internalized homophobia and internalized gay stigma, that related to internalized homonegativity and the various descriptions thereof, among the target population of LGB persons. In MEDLINE, the search was (internali* adj5 [homonegativ* or homo-negativ* or homophob* or homo-phob* or heterosex* or hetero-sex* or heteronorm* or hetero-norm*]).tw. Methodology search filters were not applied, and we did not restrict the searches to any specific languages or publication periods.

Bibliographic records imported from the databases were stored in Reference Manager. Two reviewers (RB and HMK) then read all titles/abstracts from the literature search independently to assess their relevance, compared their judgments, and obtained full-text copies of the studies deemed relevant. After reading the full-text versions, the same two reviewers classified the studies as either meeting all the inclusion criteria (these were included) or not (these were excluded). For each screening level, predesigned inclusion evaluation forms were used; differences in opinion were resolved through a reexamination by both reviewers of the papers in question and further discussion.

Inclusion/exclusion criteria were applied when evaluating the relevance of the literature identified to the review question. The following types of literature that referred to internalized homonegativity were considered: systematic reviews, experimental designs, cohort studies, case-control studies, cross-sectional studies, and qualitative studies based on interviews or focus groups with LGB persons. Nonsystematic reviews, therapists’ case reflections, letters written to editors, and theoretical or similarly nonempirical reports were excluded. We considered studies that sampled individuals who self-identify as lesbian, gay, bisexual, men who have sex with men, or women who have sex with women in terms of sexual identity (i.e., sexual orientation). We used definitions suggested by Doctor (2004), in which a gay or lesbian person is a person whose primary sexual orientation is to members of the same gender or who identifies as a member of the gay community; a bisexual is a person whose sexual orientation is directed toward men and women, though not necessarily at the same time; while the term men who have sex with men (MSM) refers to any man who has sex with a man, regardless of whether he identifies as gay, bisexual, or heterosexual. Similarly, women who have sex with women refers in this study to any woman who has sex with a woman. For literature to be included in our review, the topic of IH had to be examined, though not necessarily as a substantial or primary emphasis, in a planned research study. For the purposes of this review, IH was defined as “the gay person’s direction of negative social attitudes toward the self, leading to a devaluation of the self and poor self regard” (Meyer & Dean, 1998, p. 161); work that was based on definitions that were substantially different to this was excluded.

Books, book chapters, unpublished reports, abstracts, and brief and preliminary reports were considered for inclusion using the same criteria applied to published reports. The financial limitations of our study did not permit us to include theses and dissertations. However, no language restrictions were applied to the literature that was considered. When reports were not in languages adequately understood by the authors of this study, the material was translated using Google Translate, or native speakers were consulted.

Extraction of data and analysis

Keywording is a term describing the process of extracting data from included studies for the purposes of providing a raw “map”; because it excludes an appraisal of data quality, the actual study findings are not considered (Clapton et al., 2009; Oakley et al., 2005). This task of extracting data from the literature sources was undertaken independently by two investigators, using a predesigned data recording form with a set of prespecified questions. The form was piloted on 20 studies, and the questions were slightly revised before a final version was agreed upon. Using a standard framework such as this enabled us to ensure that the process of data extraction was performed in a methodical, systematic, and consistent manner. After the first investigator had read through each study to extract data, the second investigator checked the results. Any differences in the data extracted were resolved by reexamination of the full text and a follow-up discussion.

Keywording allows the grouping of studies according to their chief characteristics. To facilitate the mapping of IH, we extracted the following types of data: publication details; the geographical distribution of the studies (i.e., where the studies had been conducted); the range of research methods used in this field; and the characteristic of the samples—including size, gender, sexual orientation, and age. Additionally, the investigators created a set of five a priori outcome categories in which to group the studies: the assessment of factors associated with IH; levels of IH in compared populations; the measurement of IH (e.g., psychometric assessment of scales); intervention efforts to prevent or reduce IH; and lived experiences. If a paper fitted into more than one category, it was placed in the category that more closely matched the overall objective of the paper. All extracted and coded data were entered into a database. After the papers had all been keyworded and the quality assurance process completed, descriptive analyses were carried out by running frequencies and cross-tabulations.

Results

Description of the literature

The database search identified 887 individual records (Figure 1), of which 432 were considered potentially relevant. Budget limitations meant that a full-text consideration could not be given to the 156 theses/dissertations identified, and one conference abstract could not be obtained (Cox, Dewale, & Vincke, 2005). A full-text evaluation of 275 publications was therefore undertaken. Of these, 201 studies met the inclusion criteria (the details are available upon request). Most publications (94%) were journal articles, but ten conference abstracts and two book chapters were also included. These 201 studies covered a time span of 23 years, from 1989–2012, and a rapid expansion of studies was evident from the late 2000s (Figure 2).

Figure 1.

PRISMA flow diagram of literature reviewing process.

Figure 2.

Year of publication of included studies, by country of origin.

Legend: US = Studies with US sample only, and US plus Canadian samples; Other = Studies with samples from countries other than North America.

The majority of the studies (n = 164, 82%) had as a primary emphasis the examination of the association of IH with other variable(s); 18 studies compared levels of IH in various populations; 14 studies concerned IH scale measurement; four described a program for the prevention/reduction of IH; and one qualitative study presented lived experiences of IH in a group of lesbian women. Over three quarters (79%) of the studies were carried out in North America, primarily in the United States. Eight studies were conducted in Australia, seven were from Italy, four each from Belgium and the Netherlands, and two each from Mexico, Portugal, and South Africa. We located one study from each of the following countries: Austria, China, Germany, Israel, Spain, Turkey, Uganda, and the United Kingdom (UK).

In total, the 201 studies included 77,663 participants. Nearly one half (47%) of the studies concerned gay, bisexual, and other MSM; 36% included both men and women; and 18% of the study samples were composed of women. The ages of the study participants ranged from 11 to 94 years, and the mean age was 33 years. One third of the studies did not specify the sexual orientation of the participants; 50 described their sample as having an exclusively gay sexual orientation; one study sampled only bisexual women. Across all studies reporting sexual orientation, about 85% of the participants characterized themselves as gay, and 15% as bisexual. Sixty percent of the studies included tables or statements about the racial/ethnic background of the participants. Approximately three quarters of the participants were European White/Caucasian. A handful of studies examined IH among only one racial/ethnic group: European White/Caucasian men and women (two studies), Black/African American men (four studies), Black/African American women (one study), Latin men (four studies), and Asian men and women (two studies).

Factors associated with IH

All in all, we categorized 164 studies as having as their primary emphasis to examine association among variables, with IH included as an independent (typically), moderator (occasionally), or dependent (infrequently) variable—the direction of relationship was not always indicated. These studies spanned the years 1989–2012, and 71% were from the United States. The 14% (n = 23) of this scientific literature that was not from the Anglophone world originated from China, Europe, Israel, and Mexico. With the exception of seven cohort- or similar longitudinal study design and two mixed-methods studies, all were quantitative cross-sectional studies with sample sizes in the 35–2,677 range. About one half of these association studies pertained to men (48%), 16% to women, and the remaining 36% sampled male and female gay and bisexual persons. The participants were of various ethnicities, and the mean age ranged from 17–69 years, with a median age of 35 (mode = 36 years).

As noted above, across all included studies the majority of research participants were adults, but among studies classified as having an association focus, there were 14 that focused on youth (all participants ≤25 years), and six that concerned adults aged 40 and older (mean ages 55–69). With regard to the youth studies, the mean age in these studies was 19, and the participants had a mix of sexual orientation, national, and ethnic backgrounds. These studies investigated the relationship between IH and acculturation, alcohol and other substance use or abuse, gay community connection and friendships, disclosure, mental health, religion, relationship, and sexual behaviors. As is evident from the list, the majority of studies had divergent interests. However, results from several studies suggested that among LGB youth, IH is positively associated with substance abuse and unsafe sexual behaviors and negatively associated with disclosure of sexual orientation. In contrast to the studies on LGB youth, the studies on older sexual minorities consisted almost exclusively of gay, ethnically White individuals from the United States or Canada. Further, the research pertained largely to aspects of social networks and support and mental health. Overall, the results indicated that among older gay and lesbian individuals, IH is negatively associated with close friendships and social support as well as mental health, although results regarding a link between IH and loneliness appeared equivocal.

Concerning perspectives or context, a full third of the association studies involved mental health aspects, generally suggesting that higher levels of IH are related to higher risks of depression, shame, guilt, and low self-esteem. Another approximately 20% of the association studies were in regard to (males’) sexual behaviors (typically, unprotected anal sex) and satisfaction, with mixed results suggesting both no association and a positive association between IH and sexual risk behaviors. The two research areas substance use and/or abuse and relationship characteristics (e.g., satisfaction, violence) each made up about 10% (n = 15–20) of the association studies. As with sexual behaviors, results concerning the association between IH and substance use were mixed. Findings with regard to relationship satisfaction indicated a negative link with IH, while results concerning intimate partner violence suggested a positive relationship between IH and abuse or aggression perpetration, among both male and female sexual minorities. All these above issues had been investigated among LGB persons of various backgrounds.

Last, we identified about half a dozen to a dozen studies as dealing with the following three topics: body image, health care use, and identity. We found that studies that addressed body image and eating disturbance, such as links between IH and body dissatisfaction, body ideal distress, and disordered eating involved men in Western countries. Results consistently documented an association between IH and greater eating disorder symptomatology among men residing in the United States, but not in Australia. Research articles examining associations between IH and health care use and needs originated largely from North America. The results indicated that IH negatively affects the likelihood of HIV testing, use of health care resources, and attendance in treatment programs. Several studies concerned identity formation, dynamics, and congruence, suggesting IH interferes with sexual minority identity formation. Last, there were several studies on level of openness about sexual orientation and disclosure, and the role of IH with regard to gay community connection and gay social support. All in all, these studies supported the importance of self-acceptance for the formation of social networks (and vice versa).

Levels of IH

There were 18 studies that compared levels of IH in various LGB populations. In these largely descriptive studies, degree or symptoms of IH were compared across personality types (e.g., “butch vs. femme”), sexual orientation, ethnicity/race, HIV-status, religious faith, partnership experience (regarding marriage, being in the first same-sex relationship, having children), residency (e.g., rural vs. urban, U.S. vs. UK), and living in a U.S. marriage amendment state or not. The majority of studies had divergent foci. However, three U.S.-based studies investigated IH differences among various ethnicities/races. Dube and Savin-Williams (1999) found that male ethnic groups (White, African American, Latino, Asian) did not differ in levels of IH, like Moradi and colleagues (2010), who identified no significant differences in IH between White and LGBs of color. In contrast, O’Leary and collaborators (2007) found that African American MSM exhibited higher levels of IH than European American MSM.

Scale measurement

There were 14 included studies that had as their primary focus to describe measurement of IH. We note that also a few other included studies included scale information, such as Nicholson and Long (1990) and Kahn (1991). The former applied the Nungesser Homosexual Attitudes Inventory. The inventory includes a subscale on homosexual individuals’ attitude toward one’s own homosexuality, which was first described in Nungesser’s (1979) thesis, and later expanded in his 1983 book Homosexual Acts, Actors, and Identities. However, the first factor analytic study of an IH scale that operationalized the construct based on both theoretical and clinical components was the 26-item Reactions to Homosexuality Scale (RHS) by Ross and Rosser (1996). Among our included literature were subsequent studies on the RHS showing: test of a shorter version of the RHS scale (Currie, Cunningham, & Findlay, 2004), validation of the scale in Portuguese gay men (Pereira & Leal, 2005), investigation of its cross-cultural properties in Ugandan men (Ross et al., 2010), and, finally, a revised and validated 7-item scale (Smolenski, Diamond, Ross, & Rosser, 2010). The results of these five studies show that the RHS evidence metric and scalar invariance, as well as language invariance (English, Spanish, Portuguese) with good internal consistency reliability (Cronbach’s alpha = 0.73).

Another identified scale was Mayfield’s (2001) 23-item Internalized Homonegativity Inventory, which documented a Cronbach’s alpha of 0.91. Gençöz and Yüksel (2006) presented the psychometric properties of a Turkish-language version of the Internalized Homophobia scale for gay men. The scale consisted of 10 items, and the Cronbach’s alpha coefficient was 0.82.

The first scale used to measure IH in lesbians specifically was the 52-item Lesbian Internalized Homophobia Scale, developed by Szymanski and associates (Szymanski & Chung, 2001; Szymanski, Chung, & Balsam, 2001). It showed Cronbach’s alpha coefficients of 0.93 and 0.94 in two studies among American lesbians (Szymanski & Chung, 2001; Szymanski, Chung, & Balsam, 2001). Its Italian translation and factorial structure among Italian lesbians were presented by Flebus and Montano (2009). Also in Italian, we identified, first, an 85-item internalized homophobia inventory for gay men and lesbians (the authors report no internal consistency reliability results) (Flebus & Montano, 2012), and, second, the Measure of Internalized Sexual Stigma for Lesbians and Gay men, which had 12 items and a Cronbach’s alpha coefficient of 0.808 (Lingiardi, Baiocco, & Nardelli, 2012). Last, a subscale of the Lesbian, Gay, and Bisexual Identity Scale measures IH with Cronbach’s alphas of 0.86–0.93 (Mohr & Kendra, 2011), and Steffens (2008) has tested the measurement of automatic homonegativity in gay men and lesbians, finding that it can be measured with association tests (the author reports no psychometric details).

Efforts to prevent or reduce IH

We identified four studies that examined diverse programs for the prevention and reduction of IH (Lin & Israel, 2012; Ross, Doctor, Dimito, Kuehl, & Armstrong, 2008; Vega, Spieldenner, DeLeon, Nieto, & Stroman, 2011; Yadavaia & Hayes, 2012).

The first program took place in Canada, included 55 LGB persons with depression, and was composed of modified cognitive behavioral therapy. Through group-based sessions on issues such as IH, it addressed external and internal gay stigma, which were believed to be underlying causes of depression (Ross et al., 2008). The next three programs were all set in the United States, and as the Canadian program, were explored in nonrandomized studies. Vega and colleagues (2011) delivered and implemented the program SOMOS (“we are”) among 113 Latino gay men in New York City. It included group sessions, social marketing, and community presentations, and it encouraged the participants to address IH in their communities. In contrast, Lin and Israel (2012) developed and tested an online IH-reduction intervention for same-sex attracted men (n = 290), which consisted of three interactive modules. Yadavaia and Hayes (2012) described an acceptance and commitment therapy program for self-stigma around same-sex sexual orientation. Five men and women took part in the study. All four programs were described as promising.

Lived experiences

We classified one study as concerning the lived experiences of IH. Boatwright and colleagues (1996) interviewed 10 U.S. lesbian women regarding four aspects of the effects of a lesbian identity on career trajectory. One aspect explored in depth was the effects of external and internalized homonegativity. The interviews revealed that the women’s feelings of IH increased their social isolation and self-monitoring of information in the workplace, as well as lowered their professional self-confidence.

Discussion

The purposes of this review were to map the empirical research that has been undertaken on IH, provide an overview of research activity in the area for different users of research, and inform decisions on what future research might usefully address.

This systematic mapping review identified a number of characteristics of the rich body of IH literature. First, the literature on IH is rapidly expanding. Second, the concept of IH has been examined from a variety of perspectives, including sociology, gender theory, health education, feminist theory, public health, and most commonly through a psychology frame, such as development perspective and minority stress theory. There is particularly strong interest in IH among North American researchers toward the association between IH and aspects of mental health. Indeed, the issue of IH has become mainstay in psychological literature dealing with LGB experiences. Given the considerable number of studies identified, it may now be appropriate conducting in-depth systematic reviews, potentially using meta-regression analyses, to examine links between IH and aspects of mental health, sexual behaviors, substance use, relationship violence, and body image/disordered eating. Two existing meta-analyses provide the best efforts today in quantifying the relationships between IH and mental health and sexual risk-taking (Newcomb & Mustanski, 2010, 2011). These analyses included 31 and 16 studies, respectively. However, the authors excluded non-English language studies, and subsequent studies on these links have been published, which may change the results of future updates.

We found that research on IH involves most aspects of LGB persons’ life, such as spiritual (e.g., religion), mental (e.g., depression, eating disorders), career (e.g., educational or career goals), physical (e.g., substance abuse), social (e.g., relation to LGB communities), and behavioral (e.g., violence perpetration, use of health care resources). While an examination of the strength of association between variables is beyond the scope of systematic mapping reviews, we found that there is growing empirical support for a relationship between experiencing IH and higher levels of psychological distress, with the minority stress model serving as a useful framework for explaining such effects, as well as a number of related health outcomes in LGB populations. Further, results from path-analyses models suggest that IH may lead to negative mental health through lower self-esteem and less perceived availability of social support. Primary and secondary research is needed to elucidate these connections.

Most of the included studies had an individualistic focus, generally examining IH as a predictor of poor health. This focus should be tempered by increased consideration of the impact of societal factors on IH. Examinations are warranted of the link between exposure to familial, community, and structural homonegativity and IH. Specifically, the impact of social determinants such as civil rights and equity on homonegative internalizations should be examined. Relatedly, a fruitful avenue of research is a detailed, comparative assessment of the nine identified scales for the measurement of IH. While the Reactions to Homosexuality Scale is the most extensively tested and amended scale for measurement of IH, other scales evidence higher internal consistency reliability, including the subscale of the Lesbian, Gay, and Bisexual Identity Scale, which may be the best scale for assessing IH among diverse LGB samples. (We note the existence of two other known scales not included here: the 25-item Internalized Homophobia Inventory described in an unpublished doctoral dissertation by Alexander (1986) and the 9-item Internalized Homophobia scale by Martin and Dean (1987).) Given the interest in and centrality of IH as a construct relative to LGB wellbeing, it is critical that investigators know and measure the construct with the best instrument possible. For such review endeavors, researchers may wish to adopt our literature search strategy as it proved to have excellent specificity with no apparent loss of sensitivity.

The few intervention studies related to IH suggest an important area of future research. It would be particularly interesting to investigate the effects of different community norms regarding sexual minorities on IH, adopted in research of longitudinal formats where participants’ levels of IH are assessed at different points in time to establish impact. None of the four programs we identified regarding efforts to prevent or reduce IH were randomized, and only one employed a control group. In fact, almost all (93.5%) of the 201 included studies had a cross-sectional design. Clearly, more cohort or similar longitudinal study designs as well as mixed-methods studies would be useful to detail influences on and levels of IH over time. The resource intensiveness of controlled and longitudinal studies compared to cross-sectional studies may partially explain the low priority given to such designs on the topic of IH. Similarly, we identified and included only one qualitative study that addressed IH. Future qualitative research could usefully inform the development and testing of programs for the prevention and reduction of IH. For example, the phenomenology of IH could be explored, including what it is like to live with a strong sense of gay self-stigma. Moreover, we found that the majority of participants in IH research have been male, White, and residing in Western countries; thus, more research is needed to capture the viewpoints of diverse LGB communities. Only three studies were from Africa, there were just a few from Asia, and no studies at all were from Eastern Europe and South America. One possible reason for the lack of traction regarding research on IH in these areas may be related to low acceptance of LGB individuals. In Western societies, from which most IH research emerges, legal protections and support of sexual minorities have increased considerably, in parallel with research on IH, in recent decades (Hellevik, 2002; Jaspers, Lubbers, & de Graaf, 2007). Examinations of IH may be perceived as less important and legitimate in more socially and politically homonegative climates.

Limitations to our systematic mapping review should be noted when considering its findings. First, theses and dissertations were not included. Although many eventually become available as peer-reviewed articles, we expect that a number of empirical works on IH published as dissertations only are not covered here. Second, despite comprehensive and systematic literature searches, it is likely that gray literature other than dissertations could have been missed. However, we employed broad inclusion criteria and a systematic process for identifying relevant studies in any publication form. Related, studies including IH as a variable without using IH or its analogues in the title, abstract, or keywords cannot be identified through electronic searches, only expert referral. Like all reviews, once new research is done, even systematic mapping reviews become outdated. However, in contrast to effectiveness reviews, for mapping reviews this is probably a minor problem as it is unlikely that a few additional studies would change the results of the map.

In summary, within a relatively brief period of time, there has been a substantial increase in research on IH. Most empirical research is in the area of psychosocial-health correlates of IH. Conversely, gaps exist with regard to qualitative research, studies that describe ways to ameliorate IH, and examination of IH among non-Western LGB communities. Research in these areas is encouraged.