Minority Stress and Mental Health in Italian Bisexual People

: Bisexual people are a strongly stigmatized population experiencing health disparities caused by social stigmatization. The predominant framework helping to understand these health disparities and the impact of stigma on mental health of social groups belonging to a sexual minority identity constitutes the minority stress theory. In Italy, studies assessing this model in bisexual populations are very limited. Within this framework, the current study aimed at assessing in 381 Italian bisexual individuals (62 men and 319 women) the e ﬀ ects of anti-bisexual discrimination, proximal stressors (i.e., anticipated binegativity, internalized binegativity, and outness), and resilience on psychological distress. The results suggested that only anti-bisexual discrimination and internalized binegativity were positively associated with psychological distress, and that resilience was negatively associated with mental health issues. Furthermore, the results suggested that internalized binegativity mediated the relationship between anti-bisexual discrimination and mental health problems. No moderating e ﬀ ect of resilience was found. This is the ﬁrst study to have thoroughly applied minority stress in Italian bisexual people, providing Italian clinicians and researchers with an outline of the associations between minority stress, stigma, resilience, and psychological distress within this population.


Introduction
Bisexuality may be defined as "the potential to be attracted-romantically and/or sexually-to people of more than one sex and/or gender, not necessarily at the same time, not necessarily in the same way, and not necessarily to the same degree" (Eisner 2013). Due to its wide range, this definition may encompass many individuals, making bisexual people the largest group inside the lesbian, gay, bisexual, and transgender (LGBT) population (Laumann et al. 1994;Herbenick et al. 2010). Gates (2011) estimated that 1.8% of the US population (2.2% of women and 1.4% of men) self-identifies as bisexual, and that in other countries (Canada, Australia, United Kingdom, and Norway) bisexual population ranges from 0.5% to 1.2%. However, these estimates may be even larger, as they were performed on single population-based surveys (Feinstein and Dyar 2017). In Italy, that is the context of the current study, the only demographic study was performed by the National Institute of Statistic (2011), which reported that, among 7725 Italian people ranged in age from 18 to 74 years old, 2.4% of the sample declared to be gay or bisexual. However, this study did not disaggregate gay and bisexual people.
Bisexual people are a strongly stigmatized population living systematic discrimination due to their minority identity and experiencing health disparities and psychological distress caused by social stigmatization (Feinstein and Dyar 2017). The predominant theoretical frameworks helping to understand these health disparities and the impact of stigma on mental health of social groups belonging to a sexual minority identity constitute the minority stress theory (MST; Meyer 2003) and the psychological mediation framework (PMF; Hatzenbuehler 2009), as a recent extension of the MST. Previous studies have applied MST and PMF to bisexual populations (Dyar et al. 2014;Katz-Wise et al. 2017;Brewster et al. 2013), confirming the validity of these frameworks in understanding the relationships between stress, stigma, and mental health. However, while Italian lesbian, gay (LG) (e.g., Amodeo et al. 2018b;Baiocco et al. 2010;Lingiardi et al. 2012;Scandurra et al. 2017b) and transgender (e.g., Amodeo et al. 2018a;Scandurra et al. 2017aScandurra et al. , 2018aScandurra et al. , 2019dScandurra et al. , 2020 populations have been widely investigated in previous research, very few Italian studies to date have been conducted on bisexual people and, to our knowledge, no Italian studies have thoroughly tested the MST in this population. Thus, within MST and PMF, the current study aimed at exploring the effects of anti-bisexual prejudice on psychological distress, assessing the mediating role of proximal stressors and the moderating role of resilience. MST postulates that social stigma originates within a stigmatizing environment, eliciting a chronic and unique stress which, in turn, increases the risk of developing mental health problems. MST suggests that stigma affects health through the action of two different stressors: distal stressors, that are objective stressors independent of individuals (e.g., rejection and violence), and proximal stressors, that are subjective stressors dependent on individuals' perceptions and beliefs (e.g., anticipated stigma, internalized stigma, and concealment of one's sexual orientation). However, MST highlights that sexual minority groups might develop efficient psychological strategies to buffer the negative effects of stigma on health and that resilience represents a functional personal adaptive strategy able to promote social adjustment, negotiating with the social context, and generating access to personal and social resources to overcome difficulties. A recent extension of the MST (i.e., PMF) sheds light on how stressors related to stigma lead to negative mental health outcomes, positing that distal stressors would affect mental health through the mediating role of the proximal stressors. Summarizing, MST assumes that stress is a mediator of the relationship between social status and mental health, while PMF posits that stress is the initial point of departure leading to mental health problems through the mediating role of proximal stressors. Despite some inconsistencies (Rosario et al. 2002), previous longitudinal studies seem to generally confirm the causal relationships between variables in LGBT populations, finding that minority stressors represent significant predictors of negative health outcomes (e.g., Dyar et al. 2019;English et al. 2018). For instance, within a sample of cisgender bisexual women, Dyar and London (2018) found that higher levels of discrimination were associated with greater internalized binegativity and sexual identity uncertainty, and that higher levels of internalized binegativity were associated with depression and anxiety through the action of changes in identification and visibility management.
A great amount of research conducted in the past decade has detected a significant presence of minority stress and related negative health outcomes (e.g., anxiety and depression) in bisexual populations (Bostwick et al. 2010;Ross et al. 2018;Taliaferro et al. 2018). Furthermore, previous studies revealed that both heterosexual and LG individuals express stigmatizing attitudes towards bisexual people because of the stereotypical beliefs that bisexuality is not a constant sexual orientation and that bisexual people are promiscuous (Brewster and Moradi 2010;Feinstein and Dyar 2017). In particular, stigma coming from the LG community may cause in bisexual people a low community connectedness (San Francisco Human Right Commission 2011;Rankin et al. 2015). Research has also largely demonstrated that bisexual individuals are higher in negative mental health outcomes compared to monosexual (lesbian, gay, and heterosexual) individuals (Bostwick et al. 2010;Semlyen et al. 2016).
Regarding distal stressors, there is evidence that bisexual people represent a target of stigma, victimization, and abuse (Katz-Wise and Hyde 2012). For instance, both Breiding et al. (2010) and Katz-Wise and Hyde (2012) reported that bisexual individuals experience higher levels of victimization than monosexual individuals. Similarly, Walters et al. (2013) found that rates of rape are higher in bisexual women (46.1%) than heterosexual (17.4%) and lesbian (13.1%) women.
Regarding proximal stressors, previous studies found that bisexual people are more likely to conceal their bisexual orientation for fear of discrimination (Balsam and Mohr 2007). To this end, although disclosing one's sexual orientation may have advantages, previous studies found that in bisexual populations being open about sexual orientation and being involved in the LGBT community were associated with a greater likelihood of being exposed to binegativity and, as a consequence, to increased risk of developing negative mental health outcomes (Feinstein et al. 2016). Dyar et al. (2014) found that when bisexual women are in a same-gender relationship they are often perceived as lesbians, thus invalidating their sexual orientation; on the contrary, when they are in a different-gender relationship they have a tendency to internalize binegativity which, as a consequence, can lead to depressive symptoms (Paul et al. 2014) and uncertainty regarding one's sexual orientation (Dyar et al. 2014). This uncertainty has been found to be greater in bisexual populations than in LG people (Balsam and Mohr 2007). Thus, due to the high levels of binegativity and victimization, bisexual people are at high risk of internalizing negative attitudes towards bisexuality, as well as anticipating hostility and rejection from others (Feinstein and Dyar 2017). Internalized binegativity leads bisexual people to develop negative feelings about their own bisexual identity, as well as the belief that bisexuality is not a stable or valid sexual identity. Instead, anticipated binegativity may lead bisexual people to expect that others will dismiss and invalidate their bisexual identity, perceiving them as promiscuous and confused. Both internalized and anticipated binegativity resulted in being associated with depression and psychological distress (Paul et al. 2014). In regard to gender differences, Katz-Wise et al. (2017) found that bisexual men reported less bisexual-specific minority stress than bisexual women, highlighting the importance of considering gender differences when scholars assess minority stress in sexual and gender minority social groups.
Previous research also showed that bisexual people are able to protect themselves from the detrimental impact of stigma on health through specific resilience factors. For example, identity affirmation and self-acceptance provide bisexual people with a sense of pride and comfort towards their sexual identity (Paul et al. 2014) and with greater levels of outness, insight, and awareness (Rostosky et al. 2010). A similar positive effect is played by social support, which is able to buffer the negative effects of stigma on mental health (Meyer 2015;Lewis et al. 2010).
There have been very few studies in Italy addressing stigma and health in Italian bisexual people, and none of them has been specifically focused on assessing minority stress. For example, Pistella et al. (2016), in a sample of 291 Italian LGB adolescents and young adults, found that bisexual people experience greater difficulties in coming out to family members compared with LG counterparts. Petrocchi et al. (2020), in a sample of 327 Italian LGB participants, found that bisexual individuals had lower levels of authenticity, self-awareness, intimacy, community, and social justice than LG counterparts. Verrastro et al. (2016), in a sample of 468 Italian LGB youths, found that bisexual youths were more likely to be at risk of alcohol abuse than LG participants. These studies, as well as other studies addressing stigma and health in Italian LG (Lingiardi et al. 2016;Petrocchi et al. 2020;Picariello et al. 2019) and transgender (Amodeo et al. 2015;Anzani et al. 2020;Bochicchio et al. 2019;Prunas et al. 2015;Scandurra et al. 2019bScandurra et al. , 2019dVitelli et al. 2017) individuals, showed that Italian socio-cultural context is not very accepting of these social groups. Indeed, the Italian legal system does not contemplate a law prohibiting hate crimes based on sexual orientation and only in 2016 recognized same-sex civil unions, differentiating them from marriage and not allowing same-sex parents to adopt children. Furthermore, bisexual identity is not mentioned in any law. Additionally, the almost complete lack of organizations specifically designated to bisexual people, as well as the scarcity of online communities dedicated to bisexuality, may be considered as indicators of the low levels of bisexual visibility in the country.
This study aimed at filling a gap in the literature by applying MST-and its extension (i.e., PMF)-to Italian bisexual people to increase understanding of psycho-social factors contributing to psychological distress within the Italian bisexual population. Informed by MST, we hypothesized that distal stressors (i.e., prejudice events) and proximal stressors (i.e., anticipated binegativity, internalized binegativity, and concealment of one's bisexual identity) were positively associated with psychological distress (i.e., depression and anxiety), and that resilience was negatively associated with psychological distress (Hypothesis 1). Informed by PMF, we hypothesized that each proximal stressor mediated the relationship between distal stressors and psychological distress (Hypothesis 2). Then, based on MST and research addressing resilience in bisexual populations (Paul et al. 2014;Rostosky et al. 2010), we hypothesized that resilience buffered the relationship between distal/proximal stressors and psychological distress (Hypothesis 3). We also hypothesized that the indirect effect of distal stressors on psychological distress through proximal stressors was moderated by resilience (Hypothesis 4). Finally, as bisexual-specific minority stress may be different based on gender identity (Conron et al. 2010;Conron et al. 2010;Katz-Wise et al. 2017), we tested whether the hypothesized relationships were significantly different in bisexual men and women. Similarly, as socio-demographic dimensions influence both stressors and psychological distress (Scandurra et al. 2019a), we considered the potential confounding effects of the following factors: age, ethnicity, education level, monthly income, having one or more partners, type of community (urban vs. non-urban), LGBT activism, and religious education.
The tested moderated-mediation model is reported in Figure 1. Soc. Sci. 2020, 9, x FOR PEER REVIEW 4 of 17 This study aimed at filling a gap in the literature by applying MST-and its extension (i.e., PMF)-to Italian bisexual people to increase understanding of psycho-social factors contributing to psychological distress within the Italian bisexual population. Informed by MST, we hypothesized that distal stressors (i.e., prejudice events) and proximal stressors (i.e., anticipated binegativity, internalized binegativity, and concealment of one's bisexual identity) were positively associated with psychological distress (i.e., depression and anxiety), and that resilience was negatively associated with psychological distress (Hypothesis 1). Informed by PMF, we hypothesized that each proximal stressor mediated the relationship between distal stressors and psychological distress (Hypothesis 2). Then, based on MST and research addressing resilience in bisexual populations (Paul et al. 2014;Rostosky et al. 2010), we hypothesized that resilience buffered the relationship between distal/proximal stressors and psychological distress (Hypothesis 3). We also hypothesized that the indirect effect of distal stressors on psychological distress through proximal stressors was moderated by resilience (Hypothesis 4). Finally, as bisexual-specific minority stress may be different based on gender identity (Conron et al. 2010;Conron et al. 2010;Katz-Wise et al. 2017), we tested whether the hypothesized relationships were significantly different in bisexual men and women. Similarly, as socio-demographic dimensions influence both stressors and psychological distress (Scandurra et al. 2019a), we considered the potential confounding effects of the following factors: age, ethnicity, education level, monthly income, having one or more partners, type of community (urban vs. non-urban), LGBT activism, and religious education.
The tested moderated-mediation model is reported in Figure 1.

Participants
Data from 381 Italian bisexual individuals (62 men and 319 women) were analyzed in this study. Participants' age ranged from 18 to 60 years (M = 25.16, SD = 6.75). Inclusion criteria to participate in the online survey were: (1) Self-identifying in the bisexual spectrum (bisexual, polysexual, pansexual, etc.); (2) Being at least 18 years old; and (3) Living in Italy for at least 10 years. The original sample was constituted of 385 participants, among which 4 participants self-identified as transgender with a bisexual orientation. Due to the low representativity of this group, transgender participants with bisexual orientation were removed from the final sample. Socio-demographic features are reported in Table 1. Based on Fritz and MacKinnon (2007), the final sample size was considered sufficient for achieving an adequate statistical power (0.80 ;Cohen 1990).

Participants
Data from 381 Italian bisexual individuals (62 men and 319 women) were analyzed in this study. Participants' age ranged from 18 to 60 years (M = 25.16, SD = 6.75). Inclusion criteria to participate in the online survey were: (1) Self-identifying in the bisexual spectrum (bisexual, polysexual, pansexual, etc.); (2) Being at least 18 years old; and (3) Living in Italy for at least 10 years. The original sample was constituted of 385 participants, among which 4 participants self-identified as transgender with a bisexual orientation. Due to the low representativity of this group, transgender participants with bisexual orientation were removed from the final sample. Socio-demographic features are reported in Table 1. Based on Fritz and MacKinnon (2007), the final sample size was considered sufficient for achieving an adequate statistical power (0.80; Cohen 1990). Note: M = mean; SD = standard deviation; Group differences related to age were tested through Student's t test; Group differences in other factors were tested through the χ2 test.

Procedures
This study constated in a cross-sectional online survey was distributed through the social networks (e.g., Facebook) on the main Italian bisexual online groups and pages. The main Italian association promoting LGBT rights was contacted and invited to perform a snowball sampling recruitment procedure to spread the survey to their personal contacts. Such a recruitment procedure allowed us to spread the survey at a national level. Advertisements reported that participants could take the survey if they satisfied the inclusion criteria. No rewards were provided to participate in the survey.
After clicking the link of the survey, participants were directed to a page containing the informed consent, and information about researchers (affiliations, e-mails, and phone numbers of the principal investigator) was provided. Participants were informed about objectives, potential risks, benefits, and their rights to stop the survey if they desired. At the end of the first page, participants had to give their consent to take part in the survey by clicking on the button "I accept to participate in the survey". At the second page of the survey, participants had to answer the questions related to the inclusion criteria. If all criteria were satisfied, participants could start the survey.
The study was approved by the Ethical Committee of the University of Naples Federico II (project identification code: 31b/2019; date of approval: 12th September 2019), as well as designed according to the EU General Data Protection Regulation and the principles of the Declaration of Helsinki on Ethical Principles for Medical Research Involving Human Subjects. Privacy was guaranteed in accordance with Italian law 196/2003, through the use of secure gateway to which only the principal investigator could access. Furthermore, before sharing the dataset with other scholars, the principal investigator deleted the IP addresses of participants.

Distal Stressors
Two scales were used to measure distal stressors. General discriminations were assessed through the Experiences of Discrimination Scale (EDS1) (Bartos and Baban 2010; Montano and Andriola 2011), an 8-item scale measuring four prejudice events: avoidance, verbal abuse, victimization, and unequal treatment (e.g., "I happened to be marginalized because of my sexual orientation" or "I have been insulted, offended or ridiculed for my sexual orientation"). Options ranged from 1 ("never") to 5 ("often"), with higher scores indicating higher levels of prejudice events. The Cronbach's alpha was 0.81.
Everyday discriminations were assessed through the Everyday Discrimination Scale (EDS2) (Meyer et al. 2008), a 9-item scale measuring the frequency of 9 types of everyday discriminations (e.g., being treated with less respect, less courtesy, and as not smart, receiving poorer services, etc.). Each item is linked to each one's own bisexual orientation. The response options ranged from 0 ("never") to 3 ("often"). The Cronbach's alpha was 0.84.

Proximal Stressors
We assessed anticipated binegativity, internalized binegativity, and concealment of one's bisexual identity as proximal minority stressors. To assess anticipated and internalized binegativity, we used Anticipated Binegativity (AB) and Internalized Binegativity (IB) subscales of the Bisexual Identity Inventory (BII) (Paul et al. 2014), respectively.
The AB scale comprised 5 items assessing concerns and fears about how others respond to one's bisexual identity (e.g., "People might not like me if they found out that I am bisexual"). The IB scale consisted of 5 items assessing negative feelings and attitudes related to one's bisexual identity as a result of the internalization of negative societal attitudes regarding bisexuality (e.g., "My life would be better if I were not bisexual"). Items were translated into Italian by two native Italian psychologists fluent in English. Two different versions were obtained and compared in order to achieve a final agreement. Then, an English native speaker translated the new Italian version into English, confirming that the translation was accurate. Options ranged from 1 ("strongly disagree") to 7 ("strongly agree"), with greater scores indicate higher AB and IB. The confirmatory factor analyses confirmed the factor structure of both AB, χ2 (5)  Finally, we evaluated concealment of one's bisexual identity through the Outness Inventory (OI) (Mohr and Fassinger 2000;Lingiardi et al. 2012), an 11-item measure assessing the level of openness about one's sexual orientation, considering 11 groups or individuals (e.g., father, mother, siblings, straight friends, work peers, etc.). This measure consisted of three subscales (i.e., family, world, and religion), and allowed the calculation of an overall score of outness. Options ranged from 1 ("person definitely does not know about your sexual orientation status") to 7 ("person definitely knows about your sexual orientation status, and it is openly talked about"), with higher scores indicating greater outness. The Cronbach's alpha for the overall score of outness was 0.70.

Resilience
Resilience was assessed through the Resilience Scale (RS) (Wagnild and Young 1993;Peveri 2009), a 10-item scale measuring resilience as a personal feature moderating the impact of stress on health and promoting personal and social adjustment (e.g., "I usually manage one way or another"). The response options ranged from 1 ("strongly disagree") to 7 ("strongly agree") and the Cronbach's alpha was 0.89.

Psychological Distress
We assessed depression and anxiety as psychological distress dimensions. Depression was assessed through the Center for Epidemiologic Studies Depression Scale (CES-D) (Radloff 1977;Fava 1981), a 20-item scale measuring depressive symptoms during the previous week (e.g., "I felt that I could not shake off the blues even with help from my family or friends."). The response options ranged from 0 ("rarely or none of the time-less than 1 day") to 3 ("all of the time-5-7 days"). The Cronbach's alpha was 0.81. The cut-off of 16 was used in the Italian normative sample to identify individuals at risk of clinical depression (Fava 1981).
Anxiety was assessed through the Beck Anxiety Inventory (BAI) (Beck et al. 1988;Sica et al. 2006), a 21-item measure evaluating anxious symptoms (such as difficulty in breathing or fear of losing control) during the previous month. The response options ranged from 0 ("not at all") to 3 ("severely"). The Cronbach's alpha was 0.94. The cut-off of 13 was used in the Italian normative sample to identify individuals with high anxiety symptoms (Sica et al. 2006).

Statistical Analyses
To test the univariate normality of data, we examined the distribution of each observed variable for skewness and kurtosis. No variables approached skewness > |3|or kurtosis >|10|, indicating that data followed a normal distribution (Weston and Gore 2006). We considered several confounding variables in the study: age, ethnicity, education level, monthly income, having partners, type of community, LGBT activism, and religious education. As monthly income, type of community, and religious education did not show any significant associations with the other study's variables (see Table 2 for bivariate correlations), they were excluded from further analyses. Furthermore, we decided to exclude from the analyses also ethnicity of participants because of the very low number of non-Caucasian individuals. The remaining confounding variables were included in the model by regressing all the independent and dependent variables on them. Note: SD = standard deviation. * p < 0.05, ** p < 0.01, *** p < 0.001.
The hypotheses were tested through a structural equation modeling (SEM) approach in Mplus 8 (Muthén and Muthén 2017) with maximum likelihood (ML) estimator. Two latent factors were estimated: "anti-bisexual discrimination", indicated by observed measures of everyday discrimination and general discrimination, and "psychological distress", indicated by observed measures of anxiety and depression. The equivalence of structural parameters across groups (men vs. women) was tested considering two nested models: a baseline model, where all structural parameters were freely estimated for men vs. women, and a fully constrained model, where the paths were constrained to be equal across groups. We used the chi-square difference test (∆χ 2 ) to test the relative fit of nested models. After rejecting the more constrained model (p < 0.05), we tested a less restrictive model of partial invariance in which, based on modification indices, equality constraints on one or more parameters were relaxed until the change in fit was no longer significant.
We performed the analyses in two steps. In regard to the first hypothesis, the main effects of distal and proximal stressors, as well as of resilience, on psychological distress were initially tested (Hypothesis 1); then, the mediating role of proximal stressors in the relationship between anti-bisexual discrimination and psychological distress was examined (Hypothesis 2). We assessed the goodness of the model's fit by employing the following criteria: χ 2 likelihood ratio statistic, the Comparative Fit Index (CFI), and the Root Mean Square Error of Approximation (RMSEA) with associated confidence intervals. The significance value of χ 2 was not considered a relevant criterion as it is sensitive to large sample sizes. Thus, CFI values greater than 0.95 and RMSEA values lower than 0.05 were accepted (Kline 2016). As a second step, the moderating role of resilience in the relationship between anti-bisexual discrimination and psychological distress, as well as between all proximal stressors and psychological distress, was tested (Hypotheses 3). Finally, we assessed if the indirect effects of anti-bisexual discrimination on psychological distress were moderated by resilience, by computing the indices of moderated mediation for every hypothesized moderated path (Hypothesis 4), as suggested by Hayes (2015). Moderation analysis was performed by including the interaction terms in the model. Mediation (Hayes and Scharkow 2013) and moderated mediation (Hayes 2015) were assessed through bias-corrected bootstrap confidence intervals based on 5000 resamples, with confidence intervals not including zero as the indication of a significant indirect effect.

Descriptive Statistics and Bivariate Correlations
Descriptive statistics and bivariate correlations between variables are reported in Table 2. The results showed a positive correlation between discrimination experiences and both anticipated and internalized binegativity. Furthermore, except for outness, both distal and proximal stressors were positively associated with anxiety and depression.
Resilience was negatively associated with both dimensions of binegativity and both anxiety and depression, whereas a positive correlation was found between resilience and outness. Anticipated binegativity was positively associated with internalized binegativity, and negatively with outness. With respect to control variables, LGBT activism was positively correlated with anti-bisexual discrimination, outness, and resilience. Anticipated binegativity was higher in women, compared with men, and in younger participants. Finally, old age, high educational level, and having one or more partners were linked to lower levels of psychological distress.
Finally, a very high percentage of participants met the clinical cut-off for both depression (78.7%; N = 300) and anxiety (68.2%; N = 260).

Multiple-Group Structural Equation Modeling
The unconstrained model with all parameters freely estimated across men and women indicated an adequate fit to the data, χ 2 (43) = 88.17, p < 0.001; CFI = 0.95; RMSEA = 0.07, 90%C.I. [0.05, 0.09]. When comparing the fit of the unconstrained model with that from the model constraining correspondent parameters to the same values across groups, the χ 2 difference test revealed that there were no significant differences between the two models, ∆χ 2 (15) = 22.45, p = 0.10, χ 2 (58) = 110.615, p < 0.001, CFI = 0.94, RMSEA = 0.07 90%C.I. [0.05, 0.08]. Significant paths and standardized coefficients for the final model are displayed in Figure 2. Overall, we found that anti-bisexual discrimination and internalized binegativity were positively associated with psychological distress and that resilience was negatively linked to mental health issues. Furthermore, we found that anti-bisexual discrimination had a significant positive effect on both anticipated and internalized binegativity. The mediation analysis highlighted one indirect effect linking anti-bisexual discrimination with psychological distress through internalized binegativity (β = 0.02, p < 0.05, 95%C.I. [0.01, 0.04]. Soc. Sci. 2020, 9, x FOR PEER REVIEW 9 of 17 Figure 2. Results from the mediational model. Note. Solid and dashed lines represent significant and nonsignificant paths, respectively. The coefficients are referred to standardized estimates. For simplicity, associations with control variables are omitted. **p < 0.01, ***p < 0.001.

The Moderating Role of Resilience
To test the moderating role of resilience in the relationships linking distal and proximal stressors with mental health, four interaction terms predicting psychological distress were included in the model: anti-bisexual discrimination X resilience, anticipated binegativity X resilience, internalized binegativity X resilience, outness X resilience. The results highlighted no significant interaction effects, suggesting that the association of both distal and proximal stressors with psychological distress did not depend on levels of resilience (all ps > 0.05). Similarly, the moderated mediation indices showed that no mediation effect was moderated by levels of resilience (all ps > 0.05).

Discussion
Informed by MST and PMF, the current study explored the relationships between minority stressors, resilience, and psychological distress in a group of Italian bisexual individuals. In general, our findings indicated that MST can be applied to such a population, thus contributing to the scientific literature aimed at understanding the minority stress processes experienced by bisexual people and how they affect their psychological distress. To our knowledge, this is the first study which thoroughly assesses the MST in Italian bisexual people, providing Italian clinicians and researchers with an outline of the associations between stress, stigma, resilience, and psychological distress within this population.
First, regarding the descriptive analyses, results of the current study highlighted different associations between sociodemographic factors, minority stressors, psychological distress, and resilience. Indeed, minority stressors were positively related to psychological distress, while resilience was negatively associated with both minority stressors and mental health problems, confirming the strong relationship between stigma, stress, and psychological distress in bisexual With respect to control variables, LGBT activism was associated with high levels of anti-bisexual discrimination (β = 0.33, p < 0.001) and outness (β = 0.50, p < 0.001), whereas having one or more partners was associated with higher levels of resilience (β = 0.32, p < 0.001).

The Moderating Role of Resilience
To test the moderating role of resilience in the relationships linking distal and proximal stressors with mental health, four interaction terms predicting psychological distress were included in the model: anti-bisexual discrimination X resilience, anticipated binegativity X resilience, internalized binegativity X resilience, outness X resilience. The results highlighted no significant interaction effects, suggesting that the association of both distal and proximal stressors with psychological distress did not depend on levels of resilience (all ps > 0.05). Similarly, the moderated mediation indices showed that no mediation effect was moderated by levels of resilience (all ps > 0.05).

Discussion
Informed by MST and PMF, the current study explored the relationships between minority stressors, resilience, and psychological distress in a group of Italian bisexual individuals. In general, our findings indicated that MST can be applied to such a population, thus contributing to the scientific literature aimed at understanding the minority stress processes experienced by bisexual people and how they affect their psychological distress. To our knowledge, this is the first study which thoroughly assesses the MST in Italian bisexual people, providing Italian clinicians and researchers with an outline of the associations between stress, stigma, resilience, and psychological distress within this population.
First, regarding the descriptive analyses, results of the current study highlighted different associations between sociodemographic factors, minority stressors, psychological distress, and resilience. Indeed, minority stressors were positively related to psychological distress, while resilience was negatively associated with both minority stressors and mental health problems, confirming the strong relationship between stigma, stress, and psychological distress in bisexual people (Bostwick et al. 2010;Brewster et al. 2013;Ross et al. 2018;Taliaferro et al. 2018). Other associations showed that being involved in activism was positively associated with anti-bisexual discrimination, outness, and resilience, highlighting that a greater visibility contributes to resilience factors but, at the same time, increases the risk of being victimized (Feinstein et al. 2016). Furthermore, women and younger participants resulted in being higher in anticipated binegativity than men and older participants. Higher levels of expectation of rejection in women may be explained through the evidence that women who are not conforming to societal heteronormative expectations are less stigmatized than male counterparts (Habarth 2015); as a result, the lesser visibility of women living in a sexist and heteronormative society, such as the Italian one (Di Napoli et al. 2019;Lingiardi et al. 2012;Pacilli et al. 2011;Procentese et al. 2019a;Di Napoli et al. 2019;Lingiardi et al. 2012;Pacilli et al. 2011;Procentese et al. 2019b), may lead them to expect that, once their bisexual identity becomes visible, they may be rejected, heightened vigilance to social rejection. Instead, regarding the higher levels of anticipated binegativity in younger participants, it is plausible to hypothesize that older bisexual individuals have had more time to integrate their identity within the self-concept, developing competencies that allow them to buffer the negative effects of stigma on health (Van Wagenen et al. 2013). Furthermore, we found that being older, more educated, and being in intimate relationships were associated with a better mental health, confirming previous studies highlighting the protective role against psychological distress of such factors (e.g., Fredriksen-Goldsen et al. 2014;Van Wagenen et al. 2013;Whitton et al. 2018). Finally, we found that a very high percentage of participants met the clinical cut-off of both depression (78.7%) and anxiety (68.2%). These percentages are broadly higher than those detected in the general Italian cisgender people within the European Study on the Epidemiology of Mental Disorders, which were 11.1% for depression and 11.2% for anxiety (De Girolamo et al. 2005). Our findings showed that psychological distress experienced by bisexual participants was partly due to the minority stress processes, in particular to anti-bisexual discrimination and internalized binegativity.
Regarding our first hypothesis, we found that only anti-bisexual discrimination and internalized binegativity were positively related to mental health problems, and that resilience was negatively associated with psychological distress. Thus, our hypothesis was partially confirmed as not all proximal stressors resulting were associated with psychological distress. Notwithstanding that, our findings confirm that both distal and proximal stressors have a detrimental effect on mental health of bisexual people, as well as that resilience may buffer the risk of developing mental health problems. It means that our results are consistent with previous studies on MST among bisexual populations (Dyar et al. 2014;Katz-Wise et al. 2017), as well as with those highlighting the adaptive psychological strategies buffering the negative effects of minority stress on health (Paul et al. 2014). Interestingly, the finding that only internalized binegativity was associated with mental health problems may represent further evidence of Meyer (2007) theorization about the internalized stigma, defined as the most insidious subjective minority stressor. To this end, prior research found associations between internalized binegativity and depressive symptoms (Paul et al. 2014). Furthermore, this result may even be due to the very high levels of bisexual invisibility in Italy, which may contribute to increase the risk of self-stigmatizing one's bisexual identity by internalizing societal negative attitudes about bisexuality, making internalized binegativity the most significant minority stressor in Italian bisexual people. However, future research should qualitatively explore all other proximal stressors in Italian bisexual populations, providing possible justifications of the weaker effects of anticipated binegativity and concealment on psychological distress.
Regarding the second hypothesis, we found support for internalized binegativity as a mediator in the relationship between anti-bisexual discrimination and psychological distress. As before, our hypothesis is partially confirmed, as not all proximal stressors resulted as mediators in the relationship between social stigma and psychological distress. Notwithstanding, such a result represents further evidence for the PMF as a theoretical framework shedding light on psycho-social pathways through which stressors related to stigma are linked to psychological distress in the bisexual population (Brewster et al. 2013;Hatzenbuehler 2009). This result may inform clinical practice with bisexual clients, by suggesting that clinically working on negative feelings about one's bisexual identity and affirmatively validating such a sexual identity may reduce the negative effects that social rejection and prejudice may have on mental health (e.g., Amodeo et al. 2017;Scandurra et al. 2018bScandurra et al. , 2019c. However, our findings revealed also that being an activist and in intimate relationships may represent protective factors against internalized binegativity. It is plausible to hypothesize that bisexual participants with these socio-demographic characteristics have a more integrated bisexual identity than their counterparts, as both dimensions entail a greater visibility that is, in turn, probably due to a greater self-acceptance. Finally, we did not find a moderating effect of resilience in the relationships between distal and proximal stressors with psychological distress (Hypothesis 3 and 4). Surprisingly, these findings are in contrast with previous studies finding support for the buffering role of resilience factors in the relationships between minority stressors and mental health (e.g., Brewster et al. 2013), as well as those reporting minority stress differences based on gender identity and sexual orientation (e.g., Conron et al. 2010). We believe that future Italian studies should use more sensitive measures than those used in the current study. Indeed, Resilience Scale (Wagnild and Young 1993) is a general measure of resilience not specifically created for bisexual populations. Thus, future studies should consider assessing bisexual-specific protective factors, such as identity affirmation, identity pride, community connectedness, or self-acceptance. Furthermore, we did not find statistically significant differences between bisexual men and women in the hypothesized relationships. This is probably due to the sample composition, that is quite unbalanced in terms of gender (62 men vs. 319 women). Future studies should recruit more balanced samples, as well as qualitatively investigate potential gender differences in terms of minority stress.
This study has some limitations through which the results should be read. First, the cross-sectional design did not allow us to disentangle how minority stressors influence changes in psychological mediation processes and how both impact changes in health outcomes over time. Thus, the results of the current study can only rely on previous findings based on longitudinal designs supporting the psychological mediation framework (e.g., Dyar and London 2018). Future studies should overcome such a limit, analyzing longitudinal data through the use of cross-lagged panel models (Hamaker et al. 2015) and considering at least three measurements of both dependent and independent variables. Second, participants of the current study were cisgender and prevalently Caucasian. Future studies should expand the sample, including transgender people with a bisexual orientation and non-Caucasian individuals. Finally, as previously reported, future studies should assess bisexual-specific protective dimensions, considering both individual (e.g., identity affirmation) and group level (e.g., social support) factors.

Conclusions
This is the first study to assess thoroughly minority stress, stigma, resilience, and psychological distress in Italian bisexual people. The findings suggest that internalized binegativity increases the risk that anti-bisexual discrimination leads to anxiety and depression. Furthermore, although resilience did not buffer such relationships, it resulted in being negatively associated with mental health problems. These results are in line with the MST, confirming that this framework can be profitably applied to the Italian bisexual population. This study highlights the need to ameliorate internalized binegativity to reduce the risk in Italian bisexual people to develop mental health problems. Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest.