Negative attitudes toward lesbians and gay men are common in all parts of the world and are widespread in multiple spheres, including the mass media, politics, public institutions, pseudo-scientific contexts, and interpersonal relationships, which can be very distressing and sometimes unbearable for the victims (Herek and McLemore 2013; Poteat and Russell 2013). Similar to sexism, anti-semitism, and other ways of hating in the first-person plural (Moss 2003), it is necessary to gain a comprehensive understanding of these negative attitudes—often called “homophobia” or “homonegativity”—in order to develop prevention policies and practices (Lingiardi and Nardelli 2014a).Footnote 1

Several studies showed that higher levels of homonegativity are related to demographic characteristics, such as male gender (Herek 1988, 2002; Kerns and Fine 1994; Kite 1984; Kite and Whitley 1996; Morin and Garfinkle 1978; Parrott et al. 2002), older age (Baiocco et al. 2013; Herek 2002; Steffens and Wagner 2004), lower education levels (Ohlander et al. 2005; Shackelford and Besser 2007), political conservatism (Haddock and Zanna 1998; Pacilli et al. 2011; Whitley and Ægisdόttir 2000), and greater involvement in religion (Jäckle and Wenzelburger 2014; Linneman 2004; Whitley 2009; Wylie and Forest 1992). Moreover, several studies indicate that the lack of personal knowledge of lesbians and gay men is positively correlated to homonegativity (Herek and Capitanio 1996; Smith et al. 2009; Walch et al. 2010). This latter association could be explained by the “contact hypothesis” (Allport 1954): if, on the one hand, interpersonal contact with minority group members is an effective way to reduce prejudices against them, on the other hand, a lack of interpersonal contact may be associated with greater hostility against them (Costa et al. 2014).

Many researchers have also investigated the relationship between homonegativity and some personality and psychological characteristics. In general, sexual prejudice toward lesbians and gay men is associated with higher levels of (1) right-wing authoritarianism or the tendency to adhere to conventional values, comply with established and legitimate authority, and harbor distasteful attitudes toward those who may violate conventional norms (Basow and Johnson 2000; Cramer et al. 2013; Goodman and Moradi 2008; Pacilli et al. 2011; Rubinstein 2003); (2) social dominance orientation or the propensity to maintain the superior position of their in-group relative to out-groups that have lower perceived social status (Altemeyer 1998; Whitley 1999; Whitley and Ægisdόttir 2000; Whitley and Lee 2000); (3) traditional gender role attitudes or beliefs about a clear and traditional distinction between the sexes and between female-role norms and male-role norms (Herek 1988; Lingiardi et al. 2005; Whitley 2001); and (4) being closed to experience or being conventional in beliefs and attitudes, conservative in tastes, dogmatic, and rigid, as well as behaviorally set and emotionally unresponsive (Barron et al. 2008; Cullen et al. 2002; Ekehammar and Akrami 2007; Miller et al. 2012; Shackelford and Besser 2007).

Focusing on the issue that is the subject of this paper, the study of homonegativity has been neglected by Italian social scientists for a long time. Historically, Italian people have been quite ambivalent toward homosexuality. On the one hand, homosexuality is considered a “sin” or a “deviation” from normal development, yet, on the other hand, it is considered a private matter as well as an “artistic” personality trait (Baiocco et al. 2013; Capozzi and Lingiardi 2003; Lingiardi and Capozzi 2004; Rossi Barilli 1999). This ambivalent attitude has produced a typical “don’t ask, don’t tell” attitude toward homosexuality. In other words, Italian lesbian, gay, and bisexual (LGB) people are better “tolerated” if they do not publicly assert their sexual identity or engage in political activities to obtain equal rights, including the right to marry and to adopt children (Baiocco et al. 2014a; Baiocco and Laghi 2013; Lingiardi et al. 2012).

The climate of silence that existed regarding homosexuality until the 1990s has been replaced by increasingly determined political and civil demands from LGB organizations, as well as from lesbian and gay citizens. These civil demands are partly the result of the unchanging positions on homosexuality held by the Catholic Church that homosexuality is “objectively disordered” (e.g., Catechism of the Catholic Church 1993, pp. 2357–2358), a view that is inconsistent with contemporary science. From the Catholic Church’s perspective, LGB people and their families pose a threat to the cultural institution of the family, which is considered the foundation of public morality. The traditional Catholic perspective is that a family consists of a man and a woman whose sexual union is focused on procreation (e.g., Catechism of the Catholic Church 1993, pp. 2332–2333). Same-gender unions, with or without children, are inconsistent with this perspective (Baiocco et al. 2013). The Italian situation is unique because of the location of the Vatican State in Italy, which makes it highly unlikely that Italy will soon follow the example of Spain and Portugal, which legalized same-gender marriage respectively in 2005 and 2010, despite their strong Catholic cultural traditions. Recently, Pope Francesco Bergoglio made a small opening (Spadaro 2013). Nevertheless, the president of the Italian Episcopal Conference (CEI), Msgr. Bagnasco, stated that the new forms of family are aimed at confusing the people and are a kind of “Trojan horse” that will weaken the only worthy form of family, consisting of both a man and a woman (Bagnasco 2014). Largely, due to the connections between political and clerical power, the recognition of civil rights for lesbians and gay men may still progress slowly in Italy. However, it is also true that some Italian mayors do their best to change and take action so that things change. For example, some Italian mayors (e.g., in Rome, Milan, Naples, etc.) have begun to recognize same-sex unions and to promote educational projects within schools, in spite of the obstructionism and attacks by part of the Italian government (Ioverno et al. in press; see also http://lecosecambiano.roma.it).

Lesbians, gay men, and bisexual women and men constitute a sexual minority that is subjected to stigma and minority stress (Meyer 1995, 2003). As with other stigmatized minorities, episodes of discrimination and violence can have a very strong emotional impact that transcends their direct effects, affecting not only the victims directly, but also those who cannot avoid thinking that episodes of violence and discrimination might happen to them too. But, unlike other minorities, they cannot always rely on the recognition and support of their families, because the family may often be an additional source of discomfort and stress (Lingiardi 2012). Moreover, institutional discrimination toward sexual minorities, as well as the encouragement and justification of negative attitudes toward them (Turner-Frey 2014), has a direct negative impact on their well-being. Sexual minorities could internalize the social hostility against them (Drescher 1998) or develop posttraumatic stress symptoms (Dragowski et al. 2011; Roberts et al. 2010), partner relational problems (Green 2008; Lingiardi and Nardelli 2012), and other issues related to belonging to a sexual minority (Meyer and Northridge 2007; Nadal et al. 2011). On the other hand, several studies have shown that lesbians and gay men who live in contexts where there are laws against homophobia, religious organizations are more accepting, and the right to marry is not denied, resulting in significantly lower levels of distress and substance abuse (Hatzenbuehler 2009; Hatzenbuehler et al. 2010, 2012). These topics, showing the noncompliance of Italy with the EU guidelines, make the study of negative attitudes toward homosexuality in Italy very challenging and timely (ILGA Europe 2014).

The present research is a continuation and extension of a previous examination of homonegativity in Italy (Lingiardi et al. 2005). In that study, Italian versions of the Modern Homophobia Scale (MHS; Raja and Stokes 1998) and the 16 Personality Factors (16 PF) Questionnaire (Cusin and Novaga 1962; Majer and Bargellini 1986) were used to explore the relationship between homonegativity and personality characteristics and between homonegativity and gender. Analyses of the 16 PF scores showed that individuals scoring higher in homonegativity were lower in what Cattell called “ego strength” (i.e., they were more emotionally unstable and more easily upset), as well as in “submission-dominance,” indicating that they were more conformity-oriented and more likely influenced by others’ views. Regarding gender differences, the study noted that male participants had significantly higher homonegativity scores than females.

Furthermore, homonegativity was significantly associated with a lack of personal knowledge of lesbians and gay men. However, the study had two main limitations: (1) the sample comprised only young adults (age 20 to 30 years) and (2) it was not large (N = 200 participants). Nevertheless, the findings are consistent with those of the previous studies (e.g., Herek 1988, 1991, 1994, 2002; Kerns and Fine 1994; Kite 1984; Kite and Whitley 1996; Morin and Garfinkle 1978).

We addressed these limitations in the present study using an Italian sample of young adults and adults (age range 15–74) and with a greater number of participants (N = 731). As in the earlier research, as well as in some studies conducted in other European countries (Davies 2004; Steffens and Wagner 2004; Stulhofer and Rimac 2009) and non-European countries (Herek and Gonzalez-Rivera 2006; Kan et al. 2009; Yen et al. 2007), we sought to understand the impact of demographic and personality characteristics on the homonegativity once again, especially nowadays, when the political and cultural debate is in ferment.

Based on the outcomes of a literature review and our previous study, both summarized above, we hypothesized that homonegativity scores are higher among males (H1), older (H2), and less educated (H3) people. We also expected that those participants who are more politically conservative and more religiously involved show stronger homonegativity toward lesbians and gay men (H4). Regarding the personality (based on the 16 PF), we hypothesized that more homonegative individuals have particular personality characteristics, such as higher personal insecurity (O Confident-Guilt Prone), lower ego strength (C Low-High Ego Strength), higher super-ego rigidity (G Low-High Super-Ego Strength) and submission to authority (E Submission-Dominance), social conformity (M Conformism-Eccentricity), and conservatism (Q1 Conservatism-Radicalism) (H5). Finally, according to the “contact hypothesis,” we hypothesized that knowledge of lesbians and gay men is related to a more positive attitude toward homosexuality (H6).

Method

Participants and Procedures

Participants were recruited through the “snowball” recruiting technique (people were approached through advertising or direct contact and asked to help to identify other recruits). Respondents were eligible to participate in the study if they declared their sexual orientation as heterosexual and were Italian citizens. Moreover, in recruiting participants, data on the national distributions of age groups, gender, and geographical provenance provided from the National Institute of Statistics in Italy (2012) have been taken into particular account in order to approximate the demographic characteristics of research sample to the Italian population. For this purpose, the recruitment was constantly monitored: at a larger stage during the data collection, instructions for the snowball recruitment were based on more strict inclusion criteria in order to ensure that the sample’s composition remained similar to the larger population. The participation was completely voluntary and anonymous. Participants also were requested to provide their informed consent. The total sample size was 731 participants (378 women and 353 men). The mean age of the entire sample was 45.85 (SD = 16.85); among women, it was 46.82 (SD = 17.07); and among men, it was 44.81 (SD = 16.58). There was no significant statistical difference between women and men in terms of age: F(1, 729) = 2.62, p = 0.11. Table 1 describes the distribution of respondents by age, gender, and geographical area of residence. The spread of our sample in terms of age, gender, and geographical provenance is similar to the data from the last national census of the Italian population (National Institute of Statistics 2012) except for the youngest and oldest age groups. Overall, in the sample demographic, characteristics of participants from Central Italy are the most overlapping with the national data.

Table 1 Distribution of participants by age, gender, and geographical area of provenance (percentage values)
Full size table

Measures

Religious Involvement and Political Ideology

Participants completed five questions about their religious involvement and political ideology, such as the following: “How involved do you feel in religion?” or “How do you define your political stance?” With regard to the religion variable, the questions were assessed using a 5-point Likert-type scale, ranging from 1 (not at all) to 5 (extremely). Higher scores (obtained by adding the scores of each question about religion) indicated a greater religious involvement. With regard to political beliefs, these were assessed using a 5-point Likert-type scale, ranging from 1 (left wing) to 5 (right wing). Thus, higher scores of this variable (obtained as above) indicated a politically conservative stance; conversely, lower scores indicated a politically liberal stance. Similar measures are have already been used in other studies (e.g., Baiocco et al. 2013; Lingiardi et al. 2012).

Contact with Lesbians and Gay Men

Personal knowledge of lesbians and gay men was determined by asking whether participants had ever known any lesbians or gay men (“yes” or “no”). If the participant answered “yes,” she or he was then asked a question concerning the nature of these relationships through a 3-point Likert-type scale (superficial vs. close). To obtain a measure of the participants’ degree of contact with lesbians and gay men, these two questions were merged into a 4-point Likert-type scale coded as follows: no knowledge of lesbians/gay men = 0; superficial acquaintances = 1; acquaintances but not close friendships = 2; and close friendships = 3. Additionally, participants were asked how many lesbians or gay men they knew. In the present study, 71 % of participants knew lesbians or gay men (17 % knew only one person; 54 % knew more lesbians and gay men).

Homonegativity

We assessed homonegativity after adapting the MHS (Raja and Stokes 1998) to the Italian social and cultural context. The original MHS includes two subscales: one concerning attitudes toward lesbians (MHS/L) and the other concerning attitudes toward gay men (MHS/G). They provide scores on three dimensions: (1) deviance, the degree to which the respondent views homosexuality as deviant, pathological, and changeable (e.g., “Physicians and psychologists should strive to find a cure for female homosexuality”; “Male homosexuality is a psychological disease”); (2) personal discomfort (or socializing), the urge to avoid personal contact with gay men and lesbians due to discomfort caused by their presence (e.g., “I wouldn’t mind working with a lesbian”; “I wouldn’t mind going to a party that included gay men”); and (3) institutional homophobia (or rights), opposition to efforts aimed at decreasing institutional discrimination against lesbians and gay men, such as in professional organizations or religious organizations (e.g., “Employers should provide health-care benefits to the partners of their lesbian employees”; “I don’t mind companies using openly gay male celebrities to advertise their products”).

As in the original scale, the adapted form used in this study (MHS-R) includes two subscales, one concerning lesbians (MHS/L-R with 21 items) and the other concerning gay men (MHS/G-R with 22 items). Both scales were assessed using a 5-point Likert-type scale, ranging from 1 (do not agree) to 5 (strongly agree). Items were coded so that high scores reflected more homonegative attitudes.

It was necessary to develop a revised form of the original MHS in order to modify or eliminate some items that were not congruent with the Italian context. For instance, the item “Lesbians should be allowed to be leaders in religious organizations” is meaningless in Italian culture, which is closely linked to the Catholic religion, because a religious vocation and homosexuality are incompatible in the Catholic doctrine; also, it is unlikely that women would assume leadership roles in the Catholic Church. Another example is the item “I would remove my child from class if I found out the teacher was gay,” because in Italy, it is very rare that parents change the class of their sons or daughters, and it is very unusual for a teacher to reveal his or her sexual orientation to students, families, or colleagues. Instead, four new items, which are particularly important in Italian culture, were added: relationship between homosexuality and the family (“Homosexuality represents a threat to the family as a social institution”), adoption (“I think that the law should allow a gay/lesbian couple to adopt a son or daughter”), relation between homosexuality and the Catholic Church (“The Catholic Church is correct to assume homosexuality is a sin”), and confusion between homosexuality and pedophilia (“Male homosexuality and pedophilia are the same thing”).

To identify the factor structure of the MHS-R, an exploratory factor analysis (EFA) was carried out on the data of about half participants (N = 370) randomly selected from the complete sample. The three factors identified in the MHS-R included the new items, but they were very similar to the dimensions of the original MHS; thus, the same labels were retained. Confirmatory factor analysis (CFA) was performed on the data of the remaining sample of participants (N = 361) to determine the appropriateness of MHS-R’s three-factor model. The analyses were conducted via Lisrel 8.80. Lisrel computes many fit measures for the estimated model. Most of these depend on the chi-square test for the model. The chi-square test provides a reasonable measure of fit for models with about 75 to 200 cases. For models with more cases, the chi-square test is almost always statistically significant because it is affected by the size of the correlations in the model. For these reasons, alternative measures of fit have been considered. We decided to use the following fit indices: (1) the Root Mean Square Error of Approximation (RMSEA) and the confidence interval for RMSEA, (2) the Standardized Root Mean Square Residual (SRMR); (3) the Comparative Fit Index (CFI), and (4) the non-normed fit index (NNFI). To determine which values of these indices represented the boundary between the acceptance and rejection of the model, we considered the cutoff thresholds recommended by Hu and Bentler (1999): RMSEA ≤ 0.06, SRMR ≤ 0.08, CFI ~ 0.95, NNFI ≥ 0.95. All the fit indices showed an acceptable model fit for gay men, χ 2 (206) = 890.14, p < 0.01; RMSEA = 0.067, confidence interval for RMSEA = 0.063–0.070; SRMR = 0.052, CFI = 0.97; NNFI = 0.97, and for lesbians, χ 2 (186) = 747.46, p < 0.01; RMSEA = 0.064; confidence interval for RMSEA = 0.061–0.068; SRMR = 0.046, CFI = 0.97; NNFI = 0.97. The two forms of the MHS-R (MHS/L-R and MHS/G-R) were highly correlated at r = 0.88. In this study, the scores of the MHS-R factors were used. They showed excellent internal reliability (Streiner 2003). In fact, Cronbach’s alpha values were the following: deviance = 0.88; personal discomfort = 0.86; institutional homophobia = 0.87 concerning the MHS/L-R; Deviance = 0. 85; personal discomfort = 0.84; institutional homophobia = 0.84 concerning the MHS/G-R.

Personality Characteristics

To measure personality characteristics, we used the Italian form C of the 16 PF (Cusin and Novaga 1962; Majer and Bargellini 1986), which assesses 16 independent personality traits: Schizothymia–Cyclothymia; Dull–Bright; Low–High Ego Strength; Submission–Dominance; Desurgency–Surgency; Low–High Super-Ego Strength; Caution–Boldness; Hardness–Sensitivity; Trustful–Suspicious; Conformism–Eccentricity; Naiveté–Shrewdness; Confident–Guilt Prone; Conservatism–Radicalism; Group Adherence–Self-Sufficiency; Weak Willed–Self Disciplined; and Low–High Ergic Tension. In this study, we decided to administer the same version previously used in the research of Lingiardi et al. (2005) to compare results after about 10 years. The scales have six items each, except for the Dull–Bright subscale, which have eight items. The participants responded to 105 statements using three or more response options depending on the scale, such as “Yes,” “I don’t know,” and “No.” The range of scores for each subscale was from 1 to 10. Cronbach’s alphas of the different scales in previous studies in the Italian context ranged from 0.75 to 0.95 (Lingiardi et al. 2005). In the present study, 16 PF scales showed excellent internal consistency (Streiner 2003). The following values of Cronbach’s alpha coefficients were obtained: Schizothymia–Cyclothymia, 0.83; Dull–Bright, 0.87; Low–High Ego Strength, 0.85; Submission–Dominance, 0.78; Desurgency–Surgency, 0.82; Low–High Super–Ego Strength, 0.93; Caution–Boldness, 0.87; Hardness–Sensitivity, 0.90; Trustful–Suspicious, 0.84; Conformism–Eccentricity, 0.93; Naiveté–Shrewdness, 0.85; Confident–Guilt Prone, 0.80; Conservatism–Radicalism, 0.85; Group Adherence–Self-Sufficiency, 0.83; Weak Willed–Self Disciplined, 0.85; and Low–High Ergic Tension, 0.87.

Data Analysis

To conduct bivariate and multivariate analyses relating to independent variables, we used the Statistical Package for the Social Sciences (SPSS). The internal consistency of the overall scale and subscales was measured using the Cronbach’s alpha coefficient. Pearson correlations were preliminarily performed to examine the association between the three dimensions (personal discomfort, institutional homophobia, and deviance) of homonegativity toward lesbians and gay men, demographic and sociocultural variables, and personality variables. The values of the Pearson product correlation coefficient’s effective size were categorized as small, medium, and large when correlation coefficients were smaller than 0.3, between 0.3 and 0.5, or equal to or larger than 0.5, respectively (Cohen 1992). Gender differences in attitudes toward lesbians and gay men were analyzed using analysis of covariance (MANCOVA) with age and level of instruction as covariates. Hierarchical multiple regressions were performed to examine the associations between all the dimensions (personal discomfort, institutional homophobia, and deviance) of homonegativity toward lesbians and gay men, demographic and sociocultural variables, and personality traits. In the regression, we entered age, gender, and geographical area of provenance in the first step. Relationships with religion, political ideology, and level of schooling were entered in the second step; dimensions related to personal knowledge of gay and lesbian people were input in the third step. The 16 dimensions of the 16 PF were entered as a fourth step. This specific order allows for verifying the impact of each step on the three dimensions of homonegativity toward lesbians and gay men, providing the strictest test for personality characteristics (in fact, all of the previous variables are controlled for). We took into account that personality derives from the convergence of constitutional, developmental, interpersonal, and sociocultural factors. Thus, using this order, we intended to consider the specific contribution of personality, net of the variability explained by all the previous predictors. The changes in R 2 were considered to evaluate how much predictive power all the variables entered into each step had, while the total R 2 was considered as a measure of variability accounted for by all the predictors of the four steps.

Results

Relationship between Homonegativity, Demographic Characteristics, and Sociocultural Factors

The relationships between the three subscales of MHS-R for lesbians and gay men, demographics, and sociocultural variables associated with homonegativity are presented in Table 2.

Table 2 Correlations between the three subscales of MHS-R for lesbians and gay men and variables correlated to homonegativity (N = 732)
Full size table

To investigate hypothesis H1, we conducted two MANCOVA analyses using gender as the independent variable and the subscales (personal discomfort, institutional homophobia, and deviance) of MHS-R assessing homonegativity toward lesbians and gay men as dependent variables. Age and years of education were used as covariates, and the results remained significant whether the covariates were included or not; to reduce redundancy, we only reported the results with age and years of education as covariates. As the covariates, age, and years of education had significant effects both on attitudes toward lesbians (Wilks’s LambdaAge = 0.80, F(3, 724) = 59.55, p < 0.001, η2 = 0.20; Wilks’s LambdaEducation = 0.97, F(3, 724) = 7.13, p < 0.001, η2 = 0.03) and gay men (Wilks’s LambdaAge = 0.85, F(3, 724) = 43.59, p < 0.001, η2 = 0.15; Wilks’s LambdaEducation = 0.97, F(3, 724) = 8.05, p < 0.001, η2 = 0.03). A significant effect was found for gender both for attitudes toward lesbians (Wilks’s Lambda = 0.95, F(3, 724) = 11.99, p < 0.001, η2 = 0.05) and gay men (Wilks’s Lambda = 0.97, F(3, 724) = 8.70, p < 0.001, η2 = 0.03). Regarding attitudes toward lesbians, the gender variable was significant only for the deviance, F(1, 727) = 4.72, p < 0.05, η2 = 0.006, and personal discomfort MHS/L-R subscales, F(1, 727) = 7.13, p < 0.01, η2 = 0.010. As we expected (H1), female participants (M Deviance = 2.19, SD = 0.73; M Personal discomfort = 2.49, SD = 0.89) reported lower levels of homonegativity than male participants (M Deviance = 2.25, SD = 0.75; M Personal discomfort = 2.30, SD = 0.82). With respect to attitudes toward gay men, the gender variable was significant for all the MHS/G-R subscales, F Deviance(1, 727) = 6.42, p < 0.01, η2 = 0.009; F Personal discomfort(1, 727) = 24.15, p < 0.001, η2 = 0.03; F Institutional homophobia(1, 727) = 10.67, p < 0.001, η2 = 0.01. According to the hypotheses, female participants (M Deviance = 2.31, SD = 0.99; M Personal discomfort = 2.23, SD = 0.79; M Institutional homophobia = 2.69, SD = 0.78) reported lower levels of homonegativity than male participants (M Deviance = 2.43, SD = 1.02; M Personal discomfort = 2.46, SD = 0.87; M Institutional homophobia = 2.84, SD =0 .88).

Predicting Homonegativity toward Lesbians and Gay Men

To evaluate the relevance of demographic and sociocultural variables, as well as the personality traits, in predicting negative attitudes toward lesbians and gay men, a hierarchical multiple regression was conducted. Predictors were chosen based on the effect sizes from preliminary bivariate results and on issues related specifically to Italian culture. In the regression, we entered age and gender in the first step. Religious involvement, political ideology, and education were entered in the second step; dimensions related to personal knowledge of gay and lesbian people were input in the third step. The 16 factors of the 16 PF were entered as a fourth step (Tables 3 and 4).

Table 3 Hierarchical regression analyses for variables predicting homonegative attitudes toward lesbians
Full size table
Table 4 Hierarchical regression analyses for variables predicting homonegative attitudes toward gay men
Full size table

In the initial step, the demographics were found to significantly predict homonegativity: male participants are less negative toward lesbians than toward gay men. Nevertheless, irrespective of gender, older participants have homonegative attitudes both toward lesbians and toward gay men (H2). The second step involved the addition of the variables of education, political, and religious ideology to the hierarchical regression: consistent with our hypotheses, data revealed that a negative evaluation of gay men and lesbians was related to more years of education, even though the effect of this variable is not statistically significant (H3); meanwhile, political conservatism and religious involvement predicted higher levels of homonegativity (H4). In particular, religious involvement was the strongest predictor of these models. The third step showed that negative attitudes toward gay men and lesbians were significantly associated with a lack of personal knowledge of gay and lesbian people. Finally, the last step entered all 16 PF dimensions into the models simultaneously: the Low–High Super-Ego Strength (G) dimension was the strongest predictor of all the dimensions of homonegativity toward lesbians and gay men. Moreover, lower levels in the Low–High Ego Strength (C) and Hardness–Sensitivity (I) factors were associated with higher scores in negative attitudes toward lesbians. At the same time, lower levels in Dull–Bright (B), Hardness–Sensitivity (I), and Conservatism–Radicalism (Q1) factors were related to higher scores in negative attitudes toward gay men. The Confident–Guilt Prone predicted negative attitudes toward gay men, but not toward lesbians. Contrary to our expectations, Submission–Dominance (E) and Conformism–Eccentricity (M) dimensions were not significant in predicting homonegative attitudes. Finally, half of the variance of the scores of the MHS/L-R’s subscales regarding lesbians was accounted for by all the variables used. Similar results were found for the scores of the MHS/G-R’s subscales regarding gay men.

Discussion

This study is the first to provide a quite comprehensive investigation about the homonegativity within a reasonably large number of participants in Italy where sexual minorities and attitudes against them are in fact frequently ignored. As expected, the incremental R 2 associated with the first and second steps of the hierarchical regression analysis emphasized the fundamental importance of demographic variables in predicting homonegative attitudes. Older and male participants appeared to be more homonegative than younger and female participants (H1 and H2). Nevertheless, hypothesis H1 can be accepted partially, showing a complex gender dynamic associated with homonegativity. The gender dimension was considered from the viewpoint of both the participant’s and the target’s gender. Consistent with the literature (e.g., Herek 2000; Kite and Whitley 1996), results suggest that heterosexual men develop higher homonegativity toward gay men than toward lesbians. This result may confirm that male homosexuality could evoke the fear of passivity in heterosexual men and be perceived by them as a threat to their sense of masculinity. The very “gaze of a homosexual male is seen as contaminating, because it says ‘You can be penetrated,’” wrote by Nussbaum (2010, p. 114). Several studies evidenced that antigay prejudice significantly correlated with misogyny and hypermasculinity (Barron et al. 2008; Bartoş et al. 2014; Goodnight et al. 2014; Herek 1986; Theodore and Basow 2000).

Education level is also another important variable associated with homonegativity (H3), although considering it together with other variables, the hierarchical regression shows that it has less weight in predicting attitudes toward lesbians and gay men. Perhaps, implementing better education regarding LGB issues (i.e., at schools or within the media; Herek 1991) could function as a protective factor against homonegativity. Several studies demonstrate that homonegativity can significantly decrease through educational interventions (e.g., Rogers et al. 2009; Van de Ven 1995; Wright and Cullen 2001). Regarding politics and religion, more politically conservative participants and especially those involved with religion showed stronger homonegativity toward lesbians as well as toward gay men (H4).

Furthermore, incremental R 2 from the third and fourth steps indicate the function of some personality characteristics in predicting homonegative attitudes (H5). Above all, consistent with the previous study (Lingiardi et al. 2005), participants with higher Super-Ego Strength—who conform to social rules and strict moral codes and have a strong sense of duty—showed the strongest homonegativity. Moreover, our results showed that prejudicial attitudes were higher among individuals with high levels of Hardness—they tend to be utilitarian and tough-minded.

These findings seem to confirm that right-wing authoritarianism and social dominance orientation (Altemeyer 1998)—both social attitudes linked to some personality traits like Dull, Low Ego Strength, High Super-Ego Strength, and Hardness—imply prejudice toward sexual minorities and other minority groups (Pearte et al. 2013; Sibley and Duckitt 2008). Indeed, in all MHS-R subscales, negative attitudes toward gay people were significantly predicted by higher levels in the Confident–Guilt Prone dimension (people higher in this dimension are apprehensive, self-doubting, and worried). In contrast, lower levels in the Conservatism–Radicalism dimension, people who typically are associated with low openness to experience, predicted higher personal discomfort in the presence of gay people and higher institutional homophobia. These results give support to the Dual-Process Cognitive-Motivational Model (Sibley and Duckitt 2008), suggesting that personality traits with higher levels of conservatism identify the existing social order as a normative referent, tend to be more sensitive to threats and perceive out-groups as threatening in-group norms. In contrast, people who are open to new experience tend to be more accepting of out-group norms and more open to information that is inconsistent with existing social norms. Taken together, the results are in line with the previous studies including lower openness among the established correlates of homonegative attitudes (Miller et al. 2012).

The results also suggest an association between Hardness and negative attitudes toward lesbians and gay men. According to the literature (Sibley and Duckitt 2008), right-wing authoritarian personalities tend to show aspects of Hardness related to dimensions of tough-mindedness: Specifically, they tend to view the world as potentially dangerous and to make salient the intergroup superiority. The association between Hardness and the personal discomfort dimensions on MHS/L-R and MHS/G-R suggests that the perception of the world as unsafe becomes particularly relevant and visible when thinking about contact with members of out-groups (in this case, sexual minorities may represent out-groups for heterosexual people).

These aspects, Hardness, Guilt Prone, and Conservatism, seem to be more salient in predicting negative attitudes toward gay men compared to lesbians. These findings are consistent with earlier studies showing that more negative attitudes toward gay men are generally related to the representation of male homosexuality as a visible threat to the traditional gender norms and are intended to ensure that only the traditional gender-conforming, heterosexual behavior is socially acceptable (Herek 2000; Parrott, 2002).

Noteworthy, Ego Strength has a significant role in influencing attitudes toward lesbians. Ego Strength implies the ability to control emotions and behavior; thus, individuals high in this dimension have greater control over negative emotions, such as hostility. Emotional instability and lack of control over emotions seem to be underlying discriminatory behaviors toward lesbians. These results provide support to previous studies suggesting that hostility toward women and anti-lesbian attitudes is inherently linked because both serve the function of reinforcing traditional female gender norms (Wilkinson 2008).

Results also show that lower levels on the Dull–Bright dimension are associated with social discomfort in the presence of lesbians or gay men. The participants low on this dimension are concrete-thinking, have lower general mental capacity, and are unable to handle abstract problems. These findings support previous studies that suggest a relationship between cognitive functioning and conservative beliefs (Jost et al. 2003): persons having low levels of motivation to process information would be more likely to perceive predictability as beneficial and to support conservative ideologies because these rely on societal stability. We may assume that the desire for predictability leads people to avoid situations perceived as uncertain and that require information processing, such as social contact with lesbians or gay men.

Contrary to our expectations, Submission–Dominance and Conservatism–Radicalism dimensions were not significant in predicting homonegative attitudes. Dominant personalities tend to enforce the status quo of their in-group’s power position and to maintain a superior position relative to out-groups, while conformist personalities tend to identify themselves with the existing social order and to value order and stability. Thus, according to the literature (Altemeyer 1998; Whitley and Lee 2000), we expected an association between dominant and conformist personalities and negative attitudes toward lesbians and gay men. However, as Goodman and Moradi (2008) suggested in the previous study, the association may be subsumed by overlap with other personality traits related to authoritarian personalities (High Super-Ego Strength, Conservatism, and Guilt Prone).

Finally, another variable strongly associated with homonegativity is the lack of personal knowledge of lesbians and gay men (H6), suggesting that contact with them and education about sexual minorities may eradicate stereotypes and reduce antigay hostility (Allport 1954; Lingiardi and Nardelli 2014a; Smith et al. 2009).

There are some limitations to the study. Although our sample was carefully designed to try to approximate the demographic characteristics of the Italian population, it was not a probability sample. Future research could extend the use of the MHS-R to larger samples. Furthermore, the MHS-R is a self-report instrument and may be subject to social desirability (Westen and Weinberger 2004). Finally, MHS-R examines cognitive and behavioral attitudes toward homosexuality about which the respondent is aware; it does not assess homonegativity that could occur at implicit levels and is often linked to emotions of disgust (Inbar et al. 2009, 2012; Nussbaum 2010).

Conclusion

This research was conducted because there are few empirical studies on homonegativity in Italy. One impediment was the lack of a reliable measure that is appropriate for the Italian context. We hope that the MHS-R will be a useful instrument for a better understanding of negative attitudes toward homosexuality among Italians and worldwide, using our revised MHS to carry on with the research on negative attitudes toward LGB.

Consistent with other studies, our findings demonstrate that negative attitudes toward lesbians and gay men are associated with a lack of personal knowledge of them (Costa et al. 2014; Herek and Capitanio 1996; Herek and Glunt 1993). However, interpersonal contact is not the only way to decrease prejudice. Studies evidence that even indirect contact—sometimes called “vicarious contact” or “parasocial contact”—can reduce prejudices through exposing heterosexuals to the narratives of lesbians and gay men and fostering empathy toward them (e.g., through the media; Garretson 2014; Pettigrew et al. 2011; Schiappa et al. 2005, 2006). Thus, it would seem important to develop initiatives (such as antidiscrimination policies, social issues advertising, educational projects, and granting civil rights to all citizens, including same-sex marriage) aimed to promote knowledge of LGB issues. Such initiatives could highlight the lack of recognition of LGB committed relationships and how this legislative void increases minority stress (Brooks 1981; Hatzenbuehler 2009; Hatzenbuehler et al. 2010, 2012; Meyer 1995, 2003; Meyer and Northridge 2007). A focus upon reactions of parents to their LGB sons and daughters is especially important because many parents are not prepared for their children’s disclosure of their unexpected sexual orientation (Ryan et al. 2009).

Finally, another possible strategy is the development of educational projects promoting respect for individual differences and discouraging discrimination and victimization of LGB people. Because of the pervasiveness of homonegativity in school settings, such projects could be situated first of all in schools (Birkett et al. 2009; D’Augelli et al. 2002; Kosciw et al. 2012; Poteat and Russell 2013; Prati et al. 2010; Rivers 2011).

In the media, an interesting project is “It Gets Better” (Savage and Miller 2011; see also http://www.itgetsbetter.org) that in Italy has been renamed “Le cose cambiano” (it could be translated in English as “things are changing”) (Savage and Miller 2013; see also http://www.lecosecambiano.org). Its purpose is to communicate to sexual minority youth that things will improve over time and to create and inspire the changes needed to make it better for them. Sadly, there are many victims of homonegativity; for example, sexual minority youths reported significantly higher rates of victimization, depression, and suicidal feelings than their heterosexual counterparts (Baiocco et al. 2014b, c; Evans and Chapman 2014; Marshal et al. 2011). The project could also help heterosexuals to understand that homosexuality is not an illness or deviation. It will get better, we hope, also among psychologists, social workers, teachers, parents, and religious authorities; this research could contribute to them recognizing their prejudices (Drescher 2014; Lingiardi and Nardelli 2014b; Lingiardi et al. 2015a), developing more positive attitudes toward lesbians and gay men (Bartlett et al. 2009; Lingiardi and Capozzi 2004; Lingiardi et al. 2015b) and promoting acceptance of sexual minority children (Shpigel et al. 2015) and friendship between sexual minorities and heterosexual people (Baiocco et al. 2012, 2014d).