Sexual and gender minority (SGM) is an umbrella term capturing populations who identify as gay, lesbian, bisexual, transgender, queer, intersex, and other nonnormative identities that are highlighted by the use of the “ + ” sign (LGBTQI+). SGM individuals often experience sexual- and gender-based prejudices, or rather negative attitudes, due to their nonadherence to societal norms in terms of sexual orientation and/or gender identity (e.g., Herek, 2004; Hill, 2003). These negative attitudes toward SGM individuals rely on a sociocultural system defined as heterosexism (Herek, 1990), which is described as an implicit set of beliefs that considers heterosexuality and cisgender identity as natural and normal, entailing negative feelings and attitudes toward sexual and gender minorities.

Notwithstanding, the ways in which heterosexism can manifest itself are diverse and can range from more violent and hostile expressions to more benevolent and less aggressive attitudes, and this difference can depend on various social conditions. Walls (2008) developed the Multidimensional Heterosexism Inventory (MHI) to assess the different manifestations and dimensions of heterosexism, and to our knowledge, the MHI is currently the only scale that can comprehensively assess the different nuances of heterosexist attitudes, including the more subtle and stereotypically benevolent ones.

The current study is aimed at validating an Italian version of the MHI (Walls, 2008) in a sample of Italian cisgender heterosexual adults. In the following sections, we provide a comprehensive definition of the main psychosocial construct of the study (heterosexism), discussing its relations with other crucial related constructs, i.e., sexism, authoritarianism, and social dominance orientation (SDO). Then, we provide a brief presentation of the construct of benevolent heterosexism—as outlined in the work of Walls (2008)—moving from the construct of benevolent sexism and showing its theoretical and empirical significance for psychosocial research with SGM.

Sexism, Authoritarianism, Social Dominance Orientation

The characteristics and manifestations of homonegativity in Western societies have undergone considerable change in recent decades, even though the ideological system that feeds them, namely, heterosexism, appears relatively stable in its basic assumptions. Heterosexism has been originally defined by Herek (1990) as “an ideological system that denies, denigrates, and stigmatizes any nonheterosexual form of behavior, identity, relationship, or community” (p. 316). As an attitude or prejudice that can serve different psychological functions for different people, heterosexism is present in various individuals’ attitudes and actions, is influenced by various psychological (i.e., cognitive, motivational, behavioral) and situational factors, and can result in violence against SGM (Herek, 1992, 1995). Accordingly, like other ideologies of oppression such as racism and sexism, heterosexism can be regarded through a cultural lens, insofar as it is manifested in social norms and institutions (e.g., religion and the legal system), or through a psychological lens, as it is exhibited in individuals’ attitudes and behavior. In this sense, heterosexism is a measure of aversive or negative attitudes toward SGM, and as such it has been mainly employed in the scales measuring homo-/trans-negative prejudice and stigma.

Numerous empirical and theoretical studies have emphasized the strong association between sexism and heterosexism (Bochicchio et al., 2020; Brett et al., 2023; Kilianski, 2003; Scandurra et al., 2019a). Both sexism and heterosexism refer to negative attitudes, beliefs, and types of behavior that are aimed at devaluing, denigrating, stigmatizing, or restricting women and female-related characteristics on the one hand and LGBTQI+ persons or nonheterosexual forms of behavior on the other (Szymanski & Moffitt, 2012). Furthermore, both can be expressed on an individual, familial, institutional, and sociocultural level. Whereas sexism includes thinking that women are inferior to men, endorsing behaviors such as workplace sexual harassment, gender inequity in salaries, and media depictions of women that emphasize their body parts and sexual readiness, heterosexism includes believing that LGBTQI+ persons are disgusting, deserving rejection and condemnation, and denying rights, protections, and benefits associated with marriage (Szymanski & Carr, 2008) or civil unions (Scandurra et al., 2019b) to same-sex couples. Both sexism and heterosexism are often associated with right-wing authoritarianism (RWA) and SDO in research scholarship (e.g., Christopher & Wojda, 2008; Rollero et al., 2021; Sibley et al., 2007). Indeed, findings from the literature show that RWA is associated with prejudice toward gay individuals (Crawford et al., 2016; Stefurak et al., 2010) and that SDO is similarly related to both old-fashioned and modern heterosexism (Eldridge & Johnson, 2011). Even though originally conceptualized as an intra-group and an intergroup phenomenon, respectively (Altemeyer, 1998; Pratto et al., 1994), RWA and SDO both refer to attitudes supporting group-based hierarchy and inequality (Lehmiller & Schmitt, 2007). Accordingly, individuals embracing higher levels of SDO tend to feel threatened by progresses toward group equality and are thus more prone to exhibit high levels of RWA. Finally, both hostile and benevolent sexism are associated with hierarchical beliefs like those supporting RWA- and SDO-related attitudes (Lee, 2013).

Altemeyer (1998) originally regarded RWA as a fixed personality construct comprising three general characteristics, namely, authoritarian submission, conventionalism, and authoritarian aggression. More recently, Jugert and Duckitt (2009) viewed these three dimensions as expressing specific motivational goals or values. RWA has been variously conceptualized as a latent, context-specific disposition (Feldman, 2003), as a general attitude comprising dogmatism, conformist behaviors, punitiveness toward real or perceived adversaries, and a strong desire for social hierarchy (Manson, 2020), and as a group phenomenon that tends to be associated with negative opinions toward marginalized groups (Stellmacher & Petzel, 2005). Therefore, RWA-related norms are formed on the basis of both the individual’s personality and socially informed worldview beliefs. Furthermore, as RWA-related behaviors tend to increase along with societal threats (Bochicchio et al., 2021; Mezzalira et al., 2023b) and since threat hinders perceived control of events, RWA seems to increase with lower perceived control, leading individuals to support social ingroups against the outgroups (Fritsche et al., 2011). As another type of ideological orientation, SDO reflects the endorsement of a hierarchical order of social groups within society (Pratto et al., 1994). Individuals high in SDO tend to be reluctant to distribute resources and give power to subordinate groups (Halkjelsvik & Rise, 2014). Accordingly, SDO reflects the individuals’ belief that social groups are not equal, resulting in the belief in one group’s superiority over other groups. SDO has been found to be a predictor of hostile sexism toward women and tolerance of sexual harassment as well (Feather & McKee, 2012; Russell & Trigg, 2004; Sibley et al., 2007).

At the end of the twentieth century, the concept of sexism—traditionally encompassing hostility toward women and their confinement to roles that accorded them a lower status and power than men—has been disputed within the existing scholarship. Glick and Fiske (1996, 1997) proposed that the concept of sexism is intrinsically ambivalent, insofar as it subjectively involves both benevolent and hostile feelings toward women. Hostile and benevolent sexism are two sexist ideologies based on male structural power in society and female “dyadic power,” namely, a power stemming from dependency in interpersonal relationships (Guttentag & Secord, 1983). Whereas hostile sexism is intended to justify male power, traditional gender roles, and men’s exploitation of women as sexual objects, benevolent sexism is based on gentler justifications of male power and dominance and prescribed rigid gender roles (Glick & Fiske, 1996, 1997). Accordingly, benevolent sexism is an attitude perceived as positive for the sexist, in that it recognizes male dependence on women (women’s dyadic power) and is based on a romanticized conceptualization of sexual relationships with women, thus involving feelings of protectiveness and affection toward them. The two forms of sexism (i.e., hostile and benevolent) share three elements, namely, male power over women, gender differentiation (along with the endorsement of traditional gender roles), and sexuality (viewed within a patriarchal social structure) (Glick & Fiske, 1996, 1997).

In Italy, which is the context of this study, research confirms the data reported in the international scientific literature regarding the associations between sexism, heterosexism, RWA, and SDO (e.g., Callahan & Loscocco, 2023; Lingiardi et al., 2016; Rollè et al., 2022; Scandurra et al., 2017, 2020; Trappolin, 2022). Indeed, various studies have confirmed that Italy is still a highly heteronormative country, where sexual and gender minorities are strongly discriminated due to their nonheterosexual orientation or gender nonconformity (Baiocco et al., 2013; Bochicchio et al., 2019; Lingiardi et al., 2016; Mezza et al., 2023; Scandurra et al., 2020). Furthermore, Italy lacks a specific legislation protecting gender and sexual minorities from homo- or trans-phobic hate crimes, thus exposing these individuals to negative outcomes in terms of mental health and well-being (Hatzenbuehler, 2010, 2014), also due to the gender pressure they are subject to in their everyday life (Egan & Perry, 2001; Mezzalira et al., 2023a).

Benevolent Heterosexism and the Tool for Assessing It: The Multidimensional Heterosexism Inventory (MHI)

The strong association between sexism and heterosexism has been theoretically well explained by Kilianski (2003), who proposed the theory of exclusively masculine identity. The author argued that the assumption of a stereotypical male hetero-cisgender identity can entail bias and hostility toward all individuals that are not hetero/cisgender males, namely, cisgender women and lesbian/gay individuals, as well as transgender persons. This means that sexism and heterosexism stem from a common root, namely, the exclusively masculine identity, and can share common forms of expression and manifestation. Just as sexism expresses an ambivalent dimensionality ranging from overt hostility toward women to benevolent and paternalistic—and, therefore, devaluing—attitudes (Glick & Fiske, 1996), so heterosexism can be described as a multidimensional construct as well.

Assuming that heterosexism can also express a benevolent and paternalistic dimension, Walls (2008) theorized heterosexism as a multidimensional construct and proposed to capture its dimensions through the Multidimensional Heterosexism Inventory (MHI). According to the multidimensional conceptualization of heterosexism, the latter can be divided into four subdomains: (1) paternalistic heterosexism, defined as “subjectively neutral or positive attitudes, myths and beliefs that express concern for the physical, emotional or cognitive well-being of nonheterosexual persons while concurrently denying, denigrating, stigmatizing and/or segregating any nonheterosexual form of behavior, identity, relationship, or community” (Walls, 2008, pp. 27–28); (2) positive stereotypic heterosexism, including “subjectively positive attitudes, myths and beliefs that express appreciation of stereotypic characteristics often attributed to lesbians and gay men which function by denying, denigrating, stigmatizing and/or segregating any nonheterosexual form of behavior, identity, relationship, or community” (Walls, 2008, p. 28); (3) aversive heterosexism, defined as “attitudes, myths, and beliefs that dismiss, belittle, or disregard the impact of sexual orientation on life chances by denying, denigrating, stigmatizing and/or segregating any nonheterosexual form of behavior, identity, relationship, or community” (Walls, 2008, p. 46); and (4) amnestic heterosexism, that is “attitudes, myths and beliefs that deny the impact of sexual orientation on life chances by denying, denigrating, stigmatizing and/or segregating any nonheterosexual form of behavior, identity, relationship, or community” (Walls, 2008, pp. 46–47).

The MHI, therefore, is proposed as a tool to assess the different forms of heterosexism in those contexts where the openly hostile homonegativity is discouraged, contrasted, or even punished, like Italy and some other Western countries. The presence of heterosexist beliefs and attitudes, in those contexts, can indeed take a benevolent form, like the paternalistic and the positive stereotypic heterosexism, that can still represent a bias fostering discrimination, inequalities, and prejudice toward lesbians, gays, and bisexuals.

In the last decades, indeed, a vast array of research literature has been based on or mentioned the MHI for research purposes. For instance, Katz et al. (2019) found that endorsing amnestic heterosexist attitudes reduces feelings of personal responsibility to address anti-gay bullying. Van der Toorn et al. (2020) showed that threat reactions to nonheteronormative behavior reinforced the individual’s heteronormative beliefs. In addition to being included as a measure of attitudes toward gay men (Grey et al., 2013) and more generally toward sexual orientation (Ryan & Blascovich, 2015), the MHI (Walls, 2008) has been also quoted among the measures of affirmation and discrimination in LGBTQI+ individuals (e.g., Peterson et al., 2017). Seelman and Walls (2010) found that higher levels of perceived incongruence with social work values in a graduate program were associated with significantly higher levels of hostile, aversive, and paternalistic heterosexism, as well as with higher levels of RWA and SDO. Henry et al. (2022) demonstrated that religious individuals endorsing fundamentalist belief systems exhibit higher levels of heterosexism. Heterosexist attitudes using Walls’s (2008) multidimensional construct have been widely used in educational and academic settings (Clarke, 2019; Gredig & Bartelsen-Raemy, 2021; Katz & Federici, 2018; López-Sáez et al., 2020). Lastly, the multidimensional nature of heterosexism according to Stones and Glazzard (2019) has been addressed along with the minority stress theory proposed by Ilan Meyer (2003) and, more generally, in its relationship with anti-LGB microaggressions (Nadal et al., 2016; Platt & Lenzen, 2013; Spengler et al., 2016; Vishwanathan, 2022).

The literature thus points to the significance of the MHI in the research of prejudice and discrimination against lesbians, gays, and bisexual people based on the various facets of heterosexism. Therefore, our study is aimed at validating an Italian version of the MHI as proposed by Walls (2008). This is particularly important since Italy is not exempt from heterosexist attitudes that can foster serious harm to SGM, and our research can provide a useful instrument to assess the levels of heterosexism, and particularly the benevolent heterosexism, in the Italian population.

The Current Study

The current study is aimed at assessing the psychometric characteristics of the Italian version of the MHI in a national sample of Italian people by evaluating the model fit and different validities (i.e., predictive, convergent, and discriminant). Specifically, we hypothesized that (1) four subscales of the MHI have good fit indices in the Italian sample (i.e., model fit; Hypothesis 1); (2) social dominance orientation (SDO) and right-wing authoritarianism (RWA) will correlate positively with each of the subscales of the MHI (i.e., predictive validity; Hypothesis 2); (3) each of the subscales of the MHI will correlate positively with attitudes toward lesbians and gay men and ambivalent sexism (i.e., convergent validity; Hypothesis 3); (4) following Kazdin’s (2003) recommendations on the conceptual distinction between constructs, correlations between MHI and all other scales will be below 0.60 (i.e., discriminant validity; Hypothesis 4).

The literature has historically shown that men exhibit more negative attitudes than women toward homosexuality (Kite & Whitley, 1996). In a more recent study, Montgomery and Stewart (2012) found that women scored higher than men on both heterosexual privilege awareness and resistance to heteronormativity, and both these variables were associated with engagement in gay rights activism. Other studies confirmed that greater biases exist against sexual minorities among men compared to women (e.g., Andersen & Fetner, 2008; Mange & Lepastourel, 2013). Furthermore, Walls (2008) found that having any LGBTQI+ friend (as opposed to the number of such friends) was associated with higher levels of paternalistic heterosexism and that seculars were lower in apathetic heterosexism than conservative Protestants. Finally, Habarth et al. (2020) found that higher education was associated with lower heteronormativity among women.

Based on this literature, in addition to the hypotheses concerning the psychometric characteristics of the MHI, we also hypothesized that men, less educated participants, religious people, and those not having LGBTQI+ relatives and friends are generally higher in all subscales of MHI than their counterparts.

Method

Procedures

Translation of the MHI

The MHI was translated into Italian through the back-translation method (Behling & Law, 2000). Thus, 5 steps were implemented, as follows: (1) items were independently translated from English into Italian by three experts in the fields of gender studies and psychology by obtaining 3 Italian versions of the scale; (2) these versions were then compared to reach an agreement on a final unique Italian version; (3) this last Italian version was then translated into English by a native speaker with an excellent proficiency in the Italian language; (4) this new English version of the MHI was compared with the original English version and no substantial differences were found; and (5) three Italian researchers expert in gender studies and psychology participated in an online survey to evaluate independently the contents and comprehensibility of each item of the Italian version of the MHI, by answering the question “How clear are the contents of the following items?” Raters had to answer on a five-point Likert scale, from 1 (“not at all clear”) to 5 (“completely clear”). The average of all items was 4.81. The Italian version of the MHI is reported in the Appendix.

Survey Procedures

The data were collected through a web-based survey inserted in Google. Participants were recruited via internet in Italy between May 2022 and October 2022. To ensure wide and diverse participation, we strategically utilized popular social networks such as Facebook and Instagram as dissemination platforms. This approach enabled us to reach a good community sample comprising individuals from various backgrounds and demographics, comparable in size to the sample used in Walls’ (2008) first study. In recruiting participants, our goal was to engage individuals with diverse characteristics, aiming for a balanced representation of both progressive and conservative groups. To achieve this, we reached out to administrators of specific online groups, requesting them to circulate the survey among their members. Additionally, we implemented a snowball sampling procedure, encouraging all potentially interested participants to share the survey within their personal networks.

In the advertisements, it was specified that we were looking for people who (1) were aged ≥ 18 years, the Italian age of consent; (2) spoke Italian; (3) have been living in Italy at least in the last 10 years; and (4) were heterosexual and cisgender (i.e., not LGBTQI+). Participants were reached through a snowball sampling procedure, asking them to share the survey with their contacts.

By clicking on the link provided, participants could read the informed consent form, the objectives, information about the researchers, and benefits and risks of the study. They were informed about the anonymity of the survey, as well as about their right to withdraw from it if needed. All questions were mandatory to avoid missing data.

The study was approved by the Ethical Committee of the University of Calabria (date of approval: March 3, 2022), designed in respect of the principles of the Declaration of Helsinki, and conducted in accordance with the EU General Data Protection Regulation.

Participants

A total of 201 individuals (129 women and 72 men) participated in the survey. Participants ranged in age from 18 to 81 years (M = 36.42, SD = 12.56) and most of them had an educational level ≥ college (n = 145; 72.1%). Most participants declared to not have LGBTQI+ relatives (n = 152; 75.6%), but to have LGBTQI+ friends (n = 161; 80.1%), and were not religious (n = 124; 61.7%).

Measures

Sociodemographic Characteristics

The sociodemographic characteristics assessed in the current study included gender (women, men, and other), age, level of education (1 ≤ high school; 2 ≥ college), having LGBTQI+ relatives (yes vs. no) or friends (yes vs. no), and being religious (yes vs. no).

Multidimensional Heterosexism

The MHI (Walls, 2008) is a 23-item questionnaire assessing 4 dimensions of heterosexism on a 7-point Likert scale, from 1 (“strongly disagree”) to 7 (“strongly agree”). For the first 7 items only, the Likert scale provides in addition the option 0, indicating “I am OK with a gay son/lesbian daughter.” The measure consists of 4 subscales, as follows: (1) aversive heterosexism, containing items such as “Gay men should stop shoving their lifestyle down everyone’s throat”; (2) amnestic heterosexism, containing items such as “Discrimination against lesbians is virtually nonexistent in today’s society”; (3) paternalistic heterosexism, containing items such as “I would prefer my daughter NOT be a homosexual because she would face unfair discrimination”; (4) positive stereotypic heterosexism, containing items such as “Lesbians are better than heterosexual women at physically defending themselves.” Scoring and statistical information of the scale are reported in the Appendix, where the Italian version is also included.

Social Dominance Orientation

Social dominance attitudes were assessed using the Social Dominance Orientation Scale (SDO; Pratto et al., 1994) in its short version proposed by Ho et al. (2015). The short version of the SDO is an 8-item scale intended to measure “the extent to which one desires that one’s in-group dominate and be superior to outgroups” (Pratto et al., 1994, p. 742). Response options ranged from 1 (“strongly disagree”) to 7 (“strongly agree”). The SDO is subdivided into four domains: (1) pro-trait dominance (example item: “Some groups of people are simply inferior to other groups”); (2) con-trait dominance (example item: “No one group should dominate in society”); (3) pro-trait antiegalitarianism (example item: “It is unjust to try to make groups equal”); (4) con-trait antiegalitarianism (example item: “We should do what we can to equalize conditions for different groups”) (Ho et al., 2015). The alpha coefficient in the current sample was 0.72.

Right Wing Authoritarianism (RWA)

RWA was assessed through the 15-item version of the Right-Wing Authoritarianism Scale (Altemeyer, 1998; Giampaglia & Roccato, 2002), a scale assessing the degree to which people adhere to established authorities, exhibit aggression toward outgroups based on authorities sanctioning that aggression, and support traditional values endorsed by authorities (Saunders & Ngo, 2017). Response options ranged from 1 (“strongly disagree”) to 4 (“strongly agree”), with higher scores indicating greater RWA. An example item is “The only way our country can get through the crisis ahead is to get back to our traditional values, put some tough leaders in power, and silence the troublemakers spreading bad ideas.” The alpha coefficient in the current sample was 0.82.

Sexism

Sexism was assessed through the Ambivalent Sexism Inventory (ASI; Glick & Fiske, 1996; Rattazzi et al., 2008), a 22-item scale measuring sexist attitudes and feelings on a 6-point Likert scale, from 0 (“disagree strongly”) to 5 (“agree strongly”). The ASI consists of two subscales: (1) hostile sexism, which assesses the extent to which people endorse negative stereotypes of women rejecting traditional female roles and behaviors (example item: “Women exaggerate problems they have at work”), and (2) benevolent sexism, which assesses positive feelings about stereotypes associated with women who embrace traditional female roles (example item: “A good woman should be set on a pedestal by her man”). Higher scores on both subscales indicate greater sexism. The alpha coefficient in the current sample was 0.94 for hostile sexism and 0.90 for benevolent sexism.

Attitudes Toward Lesbian and Gay People

Negative attitudes toward lesbian and gay people (i.e., heterosexism intended as a unifactorial construct) were assessed through the short version of the Attitudes toward Lesbians and Gay Men (ATLG) scale (Herek, 1988), a 10-item scale measuring individuals’ attitudes toward SGM people. Response options ranged from 1 (“strongly disagree”) to 9 (“strongly agree”), with higher scores indicating greater heterosexism. An example item is “Male homosexuality is a perversion”. The alpha coefficient was 0.71.

Statistical Analyses

To test Hypothesis 1 (i.e., model fit of the MHI), a confirmatory factor analysis (CFA) with Version 2.3.0. of Jamovi using robust weighted least square estimation was performed. Specifically, we evaluated model fit by extracting the estimated loadings and evaluating the following indices: chi square/degrees of freedom (χ2/df), root mean square error of approximation (RMSEA), standardized root mean square residual (SRMR), comparative fit index (CFI), and Tucker-Lewis index (TLI). We followed the recommendations by Kline (1998), according to whom values of χ2/df < 2, RMSEA and SRMR < 0.08, TLI and CFI > 0.95 can be considered indicative values of a good fit with the data. Furthermore, the internal consistency reliability of the four subscales of the MHI was assessed through Cronbach’s alpha.

To test the hypotheses from 2 to 4 (i.e., predictive, convergent, and discriminant validity), a series of correlations between MHI subscales, social dominance orientation, RWA, attitudes toward lesbians and gay men, and ambivalent sexism was conducted using the Pearson’s correlation coefficient.

Lastly, to test potential MHI differences concerning sociodemographic factors (i.e., gender, level of education, having LGBTQI+ relatives and friends, and religiousness), independent samples t-tests were performed.

Results

Confirmatory Factor Analysis (CFA)

The original 4-factor model proposed by Walls (2008) demonstrated an adequate fit with the data obtained from the Italian sample, confirming our first hypothesis. Specifically, the following indices were found: χ2/df = 1.38, RMSEA = 0.039 (confidence intervals [CI] = 0.030, 0.048), SRMR = 0.068, CFI = 0.961, and TLI = 0.942. Internal consistency reliability was adequate for each subscale as Cronbach’s alpha ranged from 0.78 to 0.96. Full model statistics (i.e., Cronbach’s alphas, range, mean, standard deviations, standardized factor loadings, and standard error) are reported in Table 1.

Table 1 Confirmatory factor analysis of Multidimensional Heterosexism Inventory
Full size table

Predictive, Convergent, and Discriminant Validity of the MHI

Correlational analyses for hypotheses 2 to 4 are reported in Table 2. Regarding the predictive validity of the MHI, both SDO and RWA correlated positively with each of the subscales of the MHI, confirming Hypothesis 2. Similarly, regarding the convergent validity of the MHI, each of the subscales of the MHI correlated positively with negative attitudes toward lesbians and gay men, hostile sexism, and benevolent sexism, confirming Hypothesis 3.

Table 2 Correlations between Multidimensional Heterosexism Inventory, Social Dominance Orientation Scale, Ambivalent Sexism Inventory, right-wing authoritarianism, and attitudes toward lesbians and gay men
Full size table

Finally, regarding the discriminant validity of the MHI, Hypothesis 4 was partially confirmed, as not all correlations related to hypotheses 2 and 3 were below 0.60. Indeed, the subscale of the MHI concerning the aversive heterosexism showed correlation coefficients > 0.60 with hostile sexism, benevolent sexism, and RWA, thus showing a relatively high overlap between constructs.

Sociodemographic Differences in MHI Dimensions

The Student t-test was performed to compare potential differences between MHI dimensions based on gender identity and demonstrated that men presented higher levels of aversive and amnestic heterosexism, as well as general multidimensional heterosexism, than women. On the contrary, no gender differences were found with respect to either paternalistic and positive stereotypic heterosexism.

Results concerning levels of education, having LGBTQI+ friends, and religiousness were more robust. Indeed, with the exception of the absence of difference in means on paternalistic heterosexism regarding educational level, all other differences resulted statistically significant. Specifically, less educated participants, those with no LGBTQI+friends, and religious people were higher in all MHI subscales than their counterparts. On the contrary, no differences between participants with LGBTQI+ relatives and participants without LGBTQI+ relatives were found on any of the subscales. All results are reported in Table 3.

Table 3 Independent sample t-test concerning gender identity, level of education, LGBTQI+ relatives, LGBTQI+ friends, and religiousness on MHI dimensions
Full size table

Discussion

The current study is aimed at assessing the psychometric properties of the MHI in a sample of Italian cisgender and heterosexual individuals. The results obtained by CFA showed a good fit to the data, confirming the original 4-factor model of the scale. Moreover, the results showed that the MHI has adequate predictive and convergent validity in the recruited Italian sample, but that discriminant validity is not fully achieved due to overlaps of multidimensional heterosexism with other theoretical constructs (i.e., hostile sexism, benevolent sexism, and RWA). Finally, results showed some significant gender differences with respect to aversive and amnestic heterosexism, but not with respect to paternalistic and positive stereotypic heterosexism. To our knowledge, this is the first available measure to comprehensively assess the different dimensions and manifestations of heterosexist attitudes and beliefs in an Italian context, and in particular, the MHI is currently the only measure that can assess the extent of benevolent sexism in the Italian population.

Regarding predictive validity, we found significant associations of all MHI subscales with SDO and RWA. These results seem to confirm previous literature that found that both RWA and political conservatism are associated with certain negative attitudes toward same-sex sexuality (e.g., Whitley & Lee, 2000). Specifically, RWA has been shown to have a consistently positive association with prejudice toward gay individuals (Stefurak et al., 2010). As to the specific forms of multidimensional heterosexism, Walls (2008) had already found that higher levels of conservative political orientation predicted higher levels of apathetic heterosexism in an earlier version of the MHI. Eldridge and Johnson (2011), in turn, found strong associations between SDO and heterosexism (both old-fashioned and modern), likely because RWA and SDO refer to beliefs that support group-based domination and inequality (Lehmiller & Schmitt, 2007) and group-based hierarchy (Eldridge & Johnson, 2011). Indeed, individuals with high SDO tend to feel threatened by advances toward group equality and are therefore more predisposed to RWA. In short, our results are consistent with the literature in showing that all four domains of heterosexism, as theorized by Walls (2008), are associated with both RWA and SDO.

Regarding convergent validity, we found significant associations of all MHI subscales with negative attitudes toward lesbians and gay men, hostile sexism, and benevolent sexism. Glick and Fiske (1996) conceptualized benevolent sexism as consisting of three subcomponents: heterosexual intimacy (i.e., intimate relationships between men and women), protective paternalism (i.e., an individual’s unelicited interference or assistance in another person’s behavior or decision-making processes deemed beneficial to that person), and gender differentiation (i.e., conventional norms and beliefs that distinguish men and women according to their status differences). Walls’s (2008) conceptualization of paternalistic heterosexism and positive stereotypic heterosexism partially traces back and mirrors these aspects of sexism; the author also found a “significant, positive relationship between hostile sexism and amnestic heterosexism” (p. 52). Although hostile and benevolent sexism are interdependent, complementary, and cross-culturally prevalent ideologies (Glick & Fiske, 1996), they have also been considered as distinct constructs (Lee, 2013). On the one hand, hostile sexism is associated with behaviors such as sexual harassment (Begany & Milburn, 2002). On the other hand, benevolent sexism predicts paternalistic behaviors such as men’s chivalry toward women (Viki et al., 2003). The two forms of sexism (hostile and benevolent) are not only associated with higher levels of gender inequality (Glick et al., 2000), but are also positively associated with general hierarchical beliefs as held by individuals who embrace SDO and RWA (Lee, 2013). Specifically, Lee (2013) found that endorsement of SDO more strongly predicted the endorsement of hostile sexism than that of benevolent sexism, whereas the endorsement of RWA more strongly predicted endorsement of benevolent sexism than that of hostile sexism. The rationale underlying such difference lies in the different sexist ideologies supported by different worldviews, as posited by Duckitt’s dual-process model (Duckitt, 2001; Duckitt & Sibley, 2017). Accordingly, individuals high in SDO tend to endorse hostile sexism because they hold a “competitive” worldview and thus support male dominance over women. Conversely, individuals high in RWA tend to endorse benevolent sexism because they hold a “dangerous” worldview and adhere to social traditions that value women who conform to gender norms.

On the other hand, discriminant validity was partially confirmed as the aversive heterosexism subscale of the MHI was found to be overly associated with hostile sexism and RWA, indicating some degree of overlap between these theoretical constructs. In our view, the overlap between hostile sexism and aversive heterosexism is unavoidable to some degree, as they have a common origin that has been well explained theoretically by Kilianski (2003) through the theory of “exclusively masculine identity.” According to this theory, all identities that deviate from the hetero-cisgender male identity tend to generate aversion and hostility, which can be expressed at individual, familial, social, and institutional levels (Szymanski & Moffitt, 2012). Indeed, defining aversive heterosexism as a set of attitudes that dismiss or disregard the impact of sexual orientation on life chances by denying, stigmatizing, or marginalizing any nonheterosexual form of existence seems to incorporate the core definition of ambivalent sexism, as it includes the endorsement of traditional gender roles (Glick & Fiske, 1997).

We have also found that aversive heterosexism is strongly associated with RWA, perhaps because those who fall more on the “authoritarian specter” (Altemeyer, 1996) tend to perceive SGM as an affront and threat to the social order and are therefore treated with hostility and contempt. Indeed, recent research on the psychosocial impact of the COVID-19 pandemic has shown that societal threats tend to increase the levels of RWA (Bochicchio et al., 2021), which is strongly associated with higher levels of aversion and hostility toward SGM (Pacilli et al., 2022).

Lastly, accordingly to the previous literature, we found that MHI subscales were generally higher in religious people (Cragun & Sumerau, 2015; Jäckle & Wenzelburger, 2015; Roggemans et al., 2015), in less educated people (e.g., Bickmore, 2002), and in those who had no LGBTQI+ friends (e.g., Bochicchio et al., 2019). Interestingly, with respect to gender difference, we found that only the subscales of aversive and amnestic heterosexism were higher in men if compared to women, but not the subscales of paternalistic and positive stereotypic heterosexism. Herek (1988) had already shown that heterosexual men tend to have more negative attitudes toward SGM than heterosexual women, and recent research on this topic tends to confirm that gender difference affects attitudes toward homosexuality, with men being less accepting of and more aversive toward same-sex sexuality than women (Hildebrandt & Jäckle, 2023). Nevertheless, our study seems to refine the previous literature with regard to the different domains of multidimensional heterosexism by showing that only the two forms of aversive heterosexism and amnestic heterosexism are higher in men than in women. This finding is particularly interesting because it represents a further element of convergence between benevolent sexism and benevolent heterosexism. As Becker (2010) has shown, women tend to be more likely to hold benevolent sexist attitudes than men because of the internalization of sexist beliefs. Thus, it appears that benevolent expressions of sexual and gender prejudice—both sexism and heterosexism—are more acceptable to women than hostile ones. This may explain why the level of aversive heterosexism is lower for women, but the level of benevolent heterosexism is not affected by gender difference.

Limitations

Our study has significant limitations that should be considered. The cross-sectional nature of the study did not allow a complete investigation of the predictive and convergent validity of the Italian version of the MHI. In addition, the nonprobabilistic nature of the sample limited the external validity of our findings. Furthermore, the sample was relatively small, not representative of the Italian population, and unbalanced in terms of gender (129 vs. 72 men). Future studies should consider examining multidimensional heterosexism in a larger and gender-balanced sample and conduct a longitudinal study design to reevaluate our conclusions in more diverse samples and assess predictive validity through causal statistical analyses. In addition, it may be extremely interesting to include SGM individuals in future samples, as in the current study the sample consists exclusively of cisgender and heterosexual individuals, because it would be useful to understand whether SGM individuals hold benevolent heterosexist attitudes toward their own identity group as do women with benevolent sexist beliefs. Indeed, while our study captured heterosexist attitudes endorsed by heterosexual individuals, it is clear that lesbian and gay people themselves can have heterosexist beliefs and behaviors toward their own social minority group. In fact, lesbian and gay people are embedded in the same society where heterosexual individuals live, and they likely go through similar socialization processes. For this reason, it might be of great value to test the MHI among nonheterosexual individuals, also to assess group and individual differences among these partially different populations. This might also help highlight the levels of internalized heterosexism among nonheterosexual individuals (Amadio, 2006).

Social Policy Implications

Despite these limitations, the MHI could be considered an important resource for Italian researchers who need to evaluate heterosexism taking into account its multidimensional nuances. As we reported in the introduction of this paper, heterosexism is rooted in social ideologies produced and maintained in social institutions. Therefore, using the MHI can help researchers to assess not only explicit and aversive forms of heterosexism that are often sanctioned by legal systems (e.g., explicit workplace discrimination, hate speech, etc.), but also benevolent and subtle forms of heterosexism that tend to insidiously and implicitly perpetuate the heterosexist system, with significant consequences in terms of health, disparities, and inequality among SGM people. The various dimensions of heterosexism, maintained in the Italian validation of the MHI, can help assess the various forms in which heterosexism can manifest itself in the Italian population, thus aiding in shaping effective social policies that can contrast the discrimination and marginalization of nonheterosexual individuals. We agree with Hatzenbuehler (2010) in considering social policies as fundamental in fostering social contexts that can substantially compromise the mental health and well-being of the nonheterosexual population. Accordingly, social policies can be regarded as actual “health” policies, in that they can negatively impact and perpetuate mental health problems among nonheterosexual individuals by exposing them to specific minority stressors and rendering it difficult for them to access health-enhancing resources. As a result, monitoring the levels of heterosexism in the general population can be of great value to evaluate the social and health policies that can be effectively put into effect to reduce the stigmatization and discrimination of lesbian and gay people. Since the current study is aimed at providing an essential tool to assess the levels of heterosexism in the Italian population, it represents a resource that we deem as valuable in order to achieve a deeper understanding of the phenomenon and its social implications, thus helping foster actual policies aimed at reducing the marginalization of nonheterosexual individuals.

Conclusions

Our study is aimed at validating the Italian version of the MHI and also at assessing the associations between the different dimensions of heterosexism and some related constructs, i.e. sexism, RWA, and SDO. The results suggest that assessing the different nuances of heterosexism is crucial in contemporary society, as the decline of old-fashioned aversive and hostile heterosexism may give way to more subtle and implicit forms of prejudice, such as benevolent heterosexism in its positive stereotypic and paternalistic dimensions. And as the case of benevolent sexism has shown, this sociocultural process can lead to new forms of discrimination and inequality against SMG individuals. Future research should better assess the impact of benevolent heterosexism on health, well-being, and quality of life of SGM people, also examining how sociocultural and political dynamics may predict the development of heterosexist beliefs and attitudes.