Introduction

Internalized homophobia is defined as “anti-homosexual bias in the homosexual psyche” (Allen & Oleson, 1999). Homosexuals may possess attitudes and beliefs about their sexual orientation that reflect the views of the heterosexual culture. As a result, sexual minorities learn to believe that homosexuals are inferior and that they should be ashamed of their sexual tendencies (Nungesser, 1983; Shidlo, 1994). Thus, via internalized homophobia, homosexual people tend to reflect the negative attitudes of the society to themselves. Although this broad definition of internalized homophobia is common among mental health researchers and practitioners (Herek, Cogan, Gillis, & Glunt, 1997), internalized homophobia has been operationalized not only as a dislike of one's own homosexual feelings and behaviors, but also as hostile and rejecting attitudes toward other gay people, denigration of homosexuality, unwillingness to disclose one's homosexuality to others, perceptions of stigma associated with being homosexual, and acceptance of societal stereotypes about homosexuality (Kahn, 1991; Lima, Presto, Sherman, & Sobelman, 1993; Meyer, 1995; Nungesser, 1983; Ross & Rosser, 1996).

Numerous studies (e.g., Cass, 1979; Miranda & Storms, 1989; Troiden, 1989) have focused on internalized homophobia as the most important “roadblock” to the development of a positive homosexual identity. The dissonance between internalized homophobia and developing homosexual identity tends to create a conflict (Allen & Oleson, 1999). In that case, many homosexuals attempt to deny their emerging homosexual identity and they try to behave as heterosexuals in society (Bradford, Caitlin, & Rothblum, 1994).

It has been suggested that bias against homosexual persons may lead to feelings of shame and anger (Ratti, Bakeman, & Peterson, 2000). However, such anger may not be expressed toward the perceived oppressor, who is both powerful and idealized. Instead, this anger may be redirected against oneself and one's group, and expressed in the form of shame or self-hatred (Adam, 1978; Firestone, 1987). Allen and Oleson (1999) also showed that there was a significant relationship between shame and internalized homophobia. Friedman and Downey (1999) argued that during psychoanalysis even well adjusted homosexuals report antihomosexual feelings and shame about feelings of unmasculinity. Consequently, internalized homophobia appears to be associated with many psychological problems, including demoralization, guilt (Meyer, 1995), depression (Allen & Oleson, 1999; Herek et al., 1997; Lima et al., 1993), low self-esteem (Herek et al., 1997), and problems of anxiety and alcoholism (Lima et al., 1993). However, Lewis, Derlega, Griffin, and Krowinski (2003) failed to find an association between internalized homophobia and depressive symptoms, though they found associations of depressive symptoms with gay-related stress and stigma consciousness.

Psychopathology, particularly symptoms of depression and anxiety, were found to be closely associated with positive affect and negative affect (Gençöz, 2002). Thus, the association of homophobia with affective states also deserves attention. Watson and Tellegen (1985) defined two factors: positive affect and negative affect. Although these labels give the impression that these are two opposite ends of the same continuum, Watson and Tellegen suggested that they are orthogonal factors. Positive affect covers having pleasurable experiences (e.g., being excited, enthusiastic, elated) whereas negative affect refers to feelings of upset and unpleasant arousal (e.g., being distressed, fearful, and experiencing feelings like disgust and guilt). To our knowledge, there has been no study focusing on gay men's positive affect and negative affect using these orthogonal dimensions suggested by Watson and Tellegen (see also Watson, Clark, & Tellegen, 1988).

Hence, a measure of internalized homophobia was expected to have positive correlations with negative affect and psychological problems, particularly with depression and anxiety, and to have a negative correlation with self-esteem. On the other hand, its correlation with positive affect and attitudes of hostility, for example, should be low, revealing the discriminant validity of the measure of internalized homophobia, since these constructs are expected to be theoretically dissimilar.

Furthermore, cultural features may play an important role on the feelings of homophobic people. Turkey is in a transition period from collectivism to individualism (Kağıtçıbaşı, 1987), where collectivist attitudes are still appreciated. Thus, interpersonal relationships are quite close, and people are expected to behave in line with social norms. Thus, the probability of having homosexual tendencies may induce a threatening feeling, especially in such a culture, and these feelings are expected to be closely associated with psychological problems. Hence, homophobia emerges as a very critical concept in Turkey, and comparison of associated features of homophobia between collectivist and individualistic cultures gain importance to enlarge the window we are interpreting the homophobic feelings. As an initial step towards this aim, the present study focused on the examination of psychometric properties of the Turkish Internalized Homophobia Scale. Moreover, the association of internalized homophobia with psychological problems will be examined after controlling for the variance accounted for by self-esteem and positive and negative affects.

Method

Participants

The participants consisted of 132 men in romantic, intimate, or sexual relationships with men, and who identified themselves as homosexual (n=112) or bisexual (n=20). The majority of the participants were recruited from several gay, lesbian, and bisexual clubs in Ankara and Istanbul, two large cities in Turkey. Others were recruited by asking conveniently available people to inform others about the study and ask them to participate in it. All participants had at least primary school education, and 80% (n=106) had a university or higher education. As for the relationship status of the participants, 75% (n=99) reported that they were currently in a romantic relationship with a man or having one night relationships with several men; 23% (n=30) reported that they were not in a current relationship. The others reported that they were living with a woman partner (n=2) or did not report his relationship status (n=1). The participants ranged in age from 17 to 49 years, with a mean age of 26 (SD=5.6).

Measures

Internalized homophobia Scale (IHS; Herek et al., 1997)

The IHS assesses the degree to which gay individuals feel discomfort with their sexual orientation. The scale originally consisted of 9 items; however, an item was added for the Turkish version (see below). All items were rated on a 5-point Likert type scale, ranging from 1 (strongly disagree) to 5 (strongly agree). The IHS items were originally derived from the diagnostic criteria for ego-dystonic homosexuality, described in the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1980).

For the Turkish translation of the scale, three independent translators translated the scale into Turkish. Afterwards, based on a consensus of opinion, the final form of the IHS was determined. Following this procedure, back translation of this Turkish version of the scale was shown to be consistent with the original scale.

Before finalizing the scale, interviews were conducted with a few gay men and these interviews revealed that Turkish gay men were quite anxious about being labeled as homosexual and being associated with a gay community. Considering this information, for the Turkish version of the scale, an item which assesses this anxiety (i.e., “I avoid being seen and being involved with effeminate gay men”) was included in the scale. Thus, the Turkish version of the scale consisted of 10 items.

Positive and Negative Affect Schedule (PANAS; Watson et al., 1988)

The PANAS consists of 20 items, 10 of which measure Positive Affect (PA) and 10 which measure Negative Affect (NA). Participants were asked to rate the extent to which they experienced the affective states described by the PANAS adjectives in the last two weeks. Items were rated on a 5-point scale from 1 (not at all) to 5 (very much) (for reliability and validity information, see Watson, 1988; Watson et al., 1988). PA measures the extent to which a person feels enthusiastic, active, and alert; NA measures the extent to which a person experiences subjective distress, such as anger, disgust, guilt, and fear. The PANAS was adapted into Turkish by Gençöz (2000) with sufficient reliability and validity coefficients.

Brief Symptom Inventory (BSI; Derogatis, 1992)

The BSI is a shortened form of the SCL-90-R. It consists of 53 items, which are chosen from 90 items that loaded on 9 factors. It consists of 9 subscales: somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety disorder, hostility, phobic anxiety, paranoid ideation, psychosis, and additional items. The items are rated on a 5-point scale (0=none to 4=a lot/too much). The BSI was adapted into Turkish by Şahin and Durak (1994).

Rosenberg's Self-Esteem Scale (RSE; Rosenberg, 1965)

The scale is composed of 10 items that are rated on a 5-point Likert-type scale from 1 (strongly disagree) to 5 (strongly agree). Higher scores reflected higher self-esteem. The scale was standardized into Turkish by Çuhadaroğlu (1986).

Results

Means and SDs for the measures are provided in Table 1. Psychometric properties for the Turkish version of the IHS are given below.

Table 1 Means and SDs for the measures
Full size table

Reliability

Considering the reliability of the IHS, its internal consistency was examined both through Cronbach's alpha reliability coefficient and split-half reliability. Cronbach's alpha coefficient of the IHS was .82. The split-half reliability coefficient was .82 with the two halves having Cronbach's alpha coefficients of .68 and .72, respectively.

Validity

Construct validity

All items of the IHS were subjected to a principle component factor analysis. A scree-plot (Cattell, 1966) suggested a single factor, which had an eigenvalue of 4.34 and accounted for 43.4% of the total variance.

Criterion-related validity

Criterion-related validity of the scale was studied based on the correlation of the IHS with other scales and it was examined through both convergent and discriminant validity procedures (see Table 2).

Table 2 Correlations between internalized homophobia and the other measures
Full size table

As for convergent validity, the correlation of the IHS with the BSI measure was .39, p < .001. Furthermore, among the subscales of BSI, internalized homophobia correlated significantly with depression (r=.32, p < .001) and anxiety (r=.35, p < .001). Moreover, the IHS had a significant negative correlation with self-esteem (r=−.25, p < .005), and a significant positive correlation with Negative Affect (r=.30, p < .001).

Regarding discriminant validity, IHS did not show a significant correlation with Positive Affect (r=.09) and only a weak correlation with the Hostility subscale of the BSI (r=−.19, p < .05).

Table 3 Variables associated with psychological problems
Full size table

Difference between high and low scorers of internalized homophobia on psychological problems

As in previous studies (see Herek et al., 1997), while classifying the participants as high and low scorers of internalized homophobia, those participants who rated at least one of the items of IHS as “strongly agree” or “agree” were considered as the high scorers, whereas the others were considered as the low scorers. As a result of this classification, 72 participants with a mean internalized homophobia score of 13.03 were in the low internalized homophobia group, and 60 participants with a mean internalized homophobia score of 21.63 were in the high group. Consistent with expectation, the high IHP scorers reported more psychological problems (i.e., total score of BSI) (M=62.48) than the low IHP scorers (M=49.00), t(130)=−2.08, p < .05.

The association between internalized homophobia and psychological problems

In order to examine the association of internalized homophobia with psychological problems through a more robust statistical analysis, a multiple regression equation was formulated. In the first step of this analysis, self-esteem, Positive Affect, and Negative Affect measures were forced to enter, and, in the second step, the Internalized Homophobia measure was entered. As can be seen from Table 3, first step variables explained 64% of the total variance, F(3,128)=74.27, p < .001, and, among this set variables, Self-esteem, t(128)=−4.06, p < .001; pr=−.30, and Negative Affect, t(128)=10.20, p < .001; pr=.64, were significantly associated with the psychological problems. With the entrance of Internalized Homophobia, the total explained variance increased to 66%, F change (1,127)= 7.96, p < .01, and after controlling for the large variance accounted for by the first step variables, Internalized homophobia by itself revealed a significant correlation with the psychological problems, t(127)=2.82, p < .01; pr=.24.

Discussion

The current study investigated the psychometric properties of the Turkish version of IHS among gay men living in Turkey. The reliability and validity of the Turkish version of IHS revealed quite promising outcomes. Furthermore, the association of internalized homophobia with psychological problems was revealed even after controlling for the variance accounted for by the variables that were closely associated with internalized homophobia.

The present study showed that the IHS had acceptable psychometric characteristics. Considering the reliability of the scale, internal consistency coefficients, as measured by both Cronbach's alpha coefficient and split-half reliability coefficient, were quite strong. Herek et al. (1997) reported a very similar internal consistency for the scale.

As for the validity of the scale, consistent with the original measure, construct validity supported the single factor structure of the scale. Criterion validity was studied based on the correlation of the IHS with other scales, and both convergent and discriminant validity information was sought. As expected, internalized homophobia was found to be associated with psychological problems (as assessed by BSI), particularly with symptoms of depression and anxiety. Furthermore, internalized homophobia had moderate positive correlations with negative affect and a moderate negative correlation with self-esteem. These findings parallel the work of others (e.g., Allen & Oleson, 1999; Herek et al., 1997; Lima et al., 1993; Meyer, 1995).

Regarding discriminant validity, internalized homophobia scores revealed very low correlations with hostility and positive affect. Though shame and anger are frequently experienced feelings by gay men (Allen & Oleson, 1999; Friedman & Downey, 1999; Ratti et al., 2000), it is hardly expressed towards others but usually directed to one's own self or group in the form of self-hatred (Adam, 1978; Firestone, 1987). Thus hostility attitudes are expected to be independent from the feelings of internalized homophobia. Similarly, positive affect was not expected to be correlated with internalized homophobia since positive affect represents happy, active, and enthusiastic mood of the people, and one's positive affect is theoretically independent from one's negative affect (Watson et al., 1988; Watson & Tellegen, 1985). Thus, these low correlations were in line with the expectations of the discriminant validity.

Possible differences between low and high scorers of internalized homophobia were examined in terms of their psychological problems, and high scorers were found to be suffering from more psychological problems than the low scorers. Similarly in Herek et al.'s study (1997), high scorers were found to report more depressive symptoms than the low scorers.

Therefore, these reliability and validity findings support the use of the Turkish version of the IHS in Turkey. In the present study, though the average score of IHS reflects that the participants have been experiencing low level of internalized homophobia, in Meyer's (1995) and Herek et al.'s (1997) studies, the average score (i.e., total score divided by number of items) for the scale was very similar for the gay men (in the present study, the average score was 1.69, whereas in Meyer's study it was 1.52, and in Herek et al.'s study it was 1.64). One possible reason for these low average scores may be that the participants had been comfortable with their sexual identity, and the social and sexual roles expected from this identity. Supporting this argument in the present study, most of the participants were currently in a romantic relationship with a man or they were having one night relationships with several men.

Finally, in order to underline the importance of internalized homophobia, its association with psychological problems was examined. In this examination, the role of self-esteem and positive and negative affects were ruled out though they totally explained a large variance of psychological problems (64%). After controlling for the large variance accounted for by these variables, internalized homophobia still revealed significant association with the psychological problems. Hence, internalized homophobia emerges as a critical concept for the researchers of mental health, particularly in Turkey, where interdependent relationships are largely respected (Kağıtçıbaşı, 1990). Similar studies conducted in other cultures are strongly encouraged, preferably with homosexual men who have higher scores on IHS.