Minority Stress and Intimate Violence Perpetration Among Men Who Have Sex With Men in China: Hazardous Alcohol Use as a Meditator

A study in China found that stigma and alcohol use significantly contributed to intimate partner violence perpetration among men who have sex with men.

Abstract

Background

Intimate partner violence (IPV) has become a global public health issue, including in men who have sex with men (MSM). This study, based on minority stress theory, examines the relationship between minority stress and IPV perpetration among Chinese MSM, emphasizing the role of hazardous alcohol use. We propose 3 hypotheses: (1) enacted stigma, internalized stigma, identity concealment, and hazardous alcohol use contribute to IPV perpetration; (2) hazardous alcohol use mediates the effect of minority stress on IPV perpetration; and (3) proximal stressors (internalized stigma and identity concealment) serve as mediators between enacted stigma and IPV perpetration.

Methods

A total of 915 valid questionnaires were collected in China via snowball sampling for analysis, using structural equation modeling to examine the relationships between variables.

Results

Among 915 MSM, the IPV perpetration prevalence was 18.6% in the past year. The structural equation model showed that enacted stigma (β = 0.414, P < 0.001), internalized stigma (β = 0.179, P < 0.01), and hazardous alcohol use (β = 0.245, P < 0.001) significantly contributed to IPV perpetration. Although identity concealment (β = −0.134, P < 0.01) had a protective effect. Hazardous alcohol use mediated the relationship between minority stress and IPV perpetration, except for internalized stigma. Proximal stressors also partially mediated the link between enacted stigma and IPV perpetration.

Conclusions

Findings suggest that enacted stigma, internalized stigma, identity concealment, and hazardous alcohol use are associated with IPV perpetration. Reducing stigma toward MSM and addressing hazardous alcohol use are crucial for IPV perpetration.

Intimate partner violence (IPV) refers to behaviors in which one partner inflicts physical, emotional, controlling, or sexual harm on their intimate partner, whether in the past or present.¹ Intimate partner violence has emerged as a significant global public health issue that necessitates collective societal effort to address it.² Furthermore, IPV also has a significant impact on men who have sex with men (MSM).³ According to a national survey in the United States, MSM experienced a high rate of lifetime IPV, with bisexual men at 37.3% and gay men at 26%.⁴ Research conducted in various cities in China indicates that IPV is prevalent among MSM; for example, the IPV victimization rate among MSM in the Chengdu area of Sichuan Province was 35.5%, whereas the perpetration rate was 27.6%.⁵ Current research primarily focuses on the relationship between experiencing IPV and adverse health outcomes, such as substance abuse, unsafe sexual behaviors, HIV infection, and depression.⁶ However, there is a lack of attention, particularly in China, to IPV perpetration and its antecedents.

The minority stress theory posits that the MSM experience additional stress due to their marginalized sexual orientation.⁷ This discrimination and the stress related to sexual orientation not only impact individual health but also seem to increase the risk of IPV perpetration.⁸ Minority stress encompasses both distal stressors, such as enacted stigma, which refers to overt acts of discrimination and ostracism directed toward MSM,⁹ and proximal stressors, including internalized stigma and identity concealment. The former refers to MSM who have endorsed and internalized negative attitudes toward them, leading to feelings of shame and guilt.⁹ The latter refers to MSM who have adopted not to disclose their sexual minority identity during interpersonal interactions.⁷ Multiple studies have shown that both proximal and distal stressors are associated with IPV among MSM. Research conducted among MSM in the United States indicates that high rates of discrimination, internalized homophobia, and IPV perpetration are correlated.¹⁰ Furthermore, a study has concluded that a higher level of identity concealment is associated with an increased risk of physical IPV perpetration in current relationships among American MSM.¹¹

The relationship between remote minority stress and IPV perpetration may be mediated by proximal minority stress and alcohol use.¹² Two studies found that minority stress could facilitate IPV perpetration through hazardous alcohol use in lesbians.^13,14 MSM who have encountered higher levels of sexual orientation discrimination exhibit significantly more severe alcohol-related problems than those facing lower or no discrimination in the United States.¹⁵ However, different forms of minority stress may not exert similar influences. For instance, identity concealment serves as a coping strategy that can generate the stress of hiding while also providing protection against the stress of discrimination. A review indicates that less identity concealment is associated with more substance use problems.¹⁶ Increasing alcohol use is significantly linked to various types of IPV perpetration among MSM in the United States.¹⁷ Similarly, internalized homophobia, as well as the partial or complete disclosure of sexual orientation, significantly increase the risk of alcohol use among Chinese MSM.¹⁸ In addition, a study has shown that factors contributing to IPV perpetration include substance abuse and minority stress among Chinese MSM.¹⁹

Given the significant intersection between IPV and various health issues, coupled with the fact that much of the research on IPV has been conducted in the United States, there remains a notable lack of data in other regions.²⁰ Therefore, this study constructed a structural equation model to explore the linkages between minority stress, alcohol use, and IPV perpetration, based on minority stress theory (see Fig. 1). In this model, we have included 2 proximate stressors. A previous study has indicated that identity concealment and internalized stigma would interact with one another.²¹ Specifically, we hypothesized that: (1) enacted stigma, internalized stigma, identity concealment, and hazardous alcohol use contribute to IPV perpetration; (2) the impact of minority stress on IPV perpetration is mediated by hazardous alcohol use; and (3) proximal stressors (internalized stigma and identity concealment) act as mediators between enacted stigma and IPV perpetration.

Figure 1.

The research hypothesis model of minority stress, hazardous alcohol use, and IPV perpetration.

METHODS

Participants and Recruitment

From December 2023 to May 2024, we used a snowball sampling method. We collaborated with social service organizations. During events such as free HIV testing campaigns for MSM, volunteers informed participants about the survey, obtained their consent, and distributed the questionnaire via WeChat for completion. In addition, we recruited some participants through peer recommendations on Tencent QQ, a social media platform in China. The inclusion criteria were (1) aged between 18 and 65 years and (2) self-reported engagement in same-sex sexual behavior in the past year. A total of 1024 questionnaires were collected. After excluding 4 respondents who were younger than 18 years or older than 65 years, 35 with invalid logic reviews and 70 duplicate submissions, we ultimately had 915 qualified respondents. Upon completing the questionnaire, participants received a monetary reward of 20 RMB (approximately $2.8) or an equivalent gift, such as wired headphones or a data cable. This study obtained informed consent from each participant and received approval from the Ethics Committee of Chongqing Medical University (2019001).

Measurements

Background Information

The sociodemographic characteristics included age, ethnicity, household registration (indicating the place of birth, including rural and urban areas), educational level, employment, marital status, personal monthly income, and sex orientation.

Intimate Partner Violence

We used 6 items from the dating violence questionnaire designed for gay, lesbian, and bisexual adolescents.²² We modified the questionnaire to inquire whether participants had engaged in IPV perpetration against their partner in the past year.²³ The forms of IPV perpetration assessed included control, emotional, scared for safety, physical, and sexual and threats of disclosing MSM/HIV status. For example, one question asked, “Have you attempted to control most of your partner's daily activities in the past year, such as where they cannot go and whom they cannot talk to?” The response options for each statement were as follows: (A) no, (B) have done the above to same-sex partners, (C) have done the above to opposite-sex partners, and (D) have done the above to both same-sex and opposite-sex partners. Participants who selected (A) were identified as non-IPV perpetrators, and those who selected (B/C/D) were identified as IPV perpetrators. We then tallied each response to determine the number of perpetrators for each specific type of IPV perpetration. In addition, we defined IPV perpetration for those who selected (B/C/D) in any of the 6 IPV statements.

Alcohol Use

The Alcohol Use Disorders Identification Test (AUDIT-10) was used to assess alcohol use in the past year.²⁴ Questions 1 to 3 assess alcohol consumption and frequency, 4 to 6 probe drinking behaviors, and the remaining questions focus on alcohol-related problems. The total score ranges from 0 to 40. According to the classification criteria, a score of 1 to 7 indicates low-risk drinking, 8 to 14 signifies harmful drinking, and a score greater than 15 is indicative of moderate-severe alcohol use disorder. In our analysis, AUDIT was treated as a binary variable, with a score of 8 or higher denoting hazardous alcohol use for the structural equation model analysis in this study. This binary categorization allowed us to illustrate the impact of hazardous alcohol use on IPV in comparison to participants who engage in non-harmful drinking participants. In this study, Cronbach α coefficient was calculated to be 0.774.

Minority Stress

We referred to relevant literature to develop a measure of minority stress.^25–27 Four questions were designed to assess enacted stigma, with response for each item ranging from 1 to 4 (never to many times). Three items were used to evaluate internalized stigma, whereas 4 items were used to measure identity concealment. Responses for both constructs were collected using a 5-point Likert scale. Table 1 shows the minority stress constructs and their corresponding items. Cronbach α was 0.745 for enacted stigma, 0.765 for internalized stigma, and 0.854 for identity concealment.

TABLE 1.

Minority Stress Constructs and Corresponding Items in the Research

Constructs	Measures	Value Assignment
Enacted stigma	How often have you been teased or called names because you are gay/bisexual?	1 = never, 2 = once or twice, 3 = a few times, 4 = many times
	How often have you been hit or beaten up for being gay/bisexual?
	How often have you lost your friends because of being gay/bisexual?
	How often has your family not accepted you because of being gay/bisexual?
Internalized stigma	If I had the chance to be a complete heterosexual, I would accept that opportunity.	1 = strongly disagree, 2 = disagree, 3 = not sure, 4 = agree, 5 = strongly agree
	For me, being gay/bisexual is a personal flaw.
	I would like to get professional help in order to change my sexual orientation from gay/bisexual to straight.
Identity concealment	I limit what I share on social media, or who can see it, because I do not want others to know I am gay/bisexual.	1 = never do that, 2 = rarely do that, 3 = sometimes do that, 4 = often do that, 5 = always do that
	I change my mannerisms or speech because I do not want others to think I am gay/bisexual.
	I avoid talking about my romantic life because I do not want others to know I am gay/bisexual.
	I avoid telling people about certain things in my life that might imply I am gay/bisexual.

Statistical Analysis

This study used R Studio for data compilation and descriptive analysis, describing categorical variables with frequencies and percentages. Subsequently, we conducted confirmatory factor analysis and structural equation modeling (SEM) analysis by Mplus 8.3, using weighted least squares with mean and variance adjusted) to estimate the parameters. Specifically, we assessed the reliability and validity of the measurement model through confirmatory factor analysis. After that, SEM analysis and bootstrap method was performed to test the hypotheses.

Composite reliability (CR) and Cronbach α, with values exceeding 0.70, indicating adequate internal consistency. Average variance extracted (AVE), with a value exceeding 0.50 considered sufficient for convergence validity. If the standardized correlation coefficient between different factors is lower than the square root of the AVE value, it indicates adequate discriminant validity. The model fit was evaluated based on the following criteria: the chi-squared and degree of freedom ratio (χ²/df) < 5, Tucker-Lewis index (TLI) > 0.90, comparative fit index (CFI) > 0.90, root mean square error of approximation (RMSEA) < 0.08, and standardized root mean square residual (SRMR) < 0.08.²⁸ When a P value < 0.05, it was determined to have a significant difference. In addition, the indirect effects and the 95% confidence intervals (CIs) were calculated using the bootstrap method.

RESULTS

Descriptive Results

A total of 915 MSM were involved in the analysis. Among them, 48.0% were aged 18 to 28 years, most (94.5%) were Han nationality, 64.9% resided in urban areas, 54.5% held a bachelor's degree, 70.1% were employed, 86.2% were unmarried, and 41.7% reported a monthly disposable income of 5000 to 10,000 RMB (approximately $682–$1364). Most participants identified as gay (77%). In terms of alcohol use, 36.1% of MSM had hazardous alcohol use. One hundred seventy (18.6%) MSM reported committing any form of IPV against their partners in the last 12 months. Among all forms, control IPV was the most prevalent (12.3%), whereas threat of disclosing MSM/HIV status occurred less frequently (6.2%). Intimate partner violence perpetration was more common among same-sex partners, with 7.2% of participants reporting any form of IPV against opposite-sex partners. The descriptive results are presented in Table 2.

TABLE 2.

Descriptive Results of the Study (N = 915)

Variables	Classification	N (n = 915)	%
Age, y	18–28	439	48.0
	29–39	371	40.5
	≥40	105	11.5
Ethnicity	Han nationality	865	94.5
	Other ethnic minorities	50	5.5
Household registration*	Urban areas	594	64.9
	Rural areas	321	35.1
Educational level	High school and below	180	19.7
	Junior college	236	25.8
	College and above	499	54.5
Employment status	Employed	641	70.1
	Retired or unemployed	130	14.2
	Students	144	15.7
Marital status	Unmarried	789	86.2
	Married	63	6.9
	Divorced/widowed	63	6.9
Personal monthly income	1000 RMB† or less	63	6.9
	1001–3000 RMB	215	23.5
	3001–5000 RMB	256	28.0
	5001–10,000 RMB	278	30.4
	10,000 RMB or more	103	11.3
Sex orientation	Gay	705	77.0
	Bisexuality/Other	210	23.0
Alcohol use	Low risk alcohol use	585	63.9
	Harmful alcohol use	245	26.8
	Moderate-severe alcohol use	85	9.3
IPV perpetration	Yes	170	18.6
	No	745	81.4
Form of IPV perpetration	—	—	—
Control	Yes	113	12.3
	No	802	87.7
Emotional	Yes	92	10.1
	No	823	89.9
Scared for safety	Yes	83	9.1
	No	832	90.9
Physical	Yes	73	8.0
	No	842	92.0
Sexual	Yes	73	8.0
	No	842	92.0
Threat of disclosing MSM/HIV status	Yes	57	6.2
	No	858	93.8
IPV perpetration to whom	To same-sex partner	164	17.9
	To opposite-sex partner	66	7.2
	To same and opposite-sex partner	60	6.6

*Usually captures the place of birth in China, including rural and urban areas. Some people with rural household registrations arrive in cities to live after growing up.

^†1 RMB equal to 0.1364 US dollars.

Measurement Model

In Table 3, the factor loading, CR, and AVE for each construct, as well as discriminant validity between constructs, are presented. All factor loadings were greater than 0.6 (P < 0.001), which is considered acceptable. Both reliability and validity met the established criteria. The measurement model fit indices indicate that the model is acceptable (χ²/df = 217.116/41, CFI = 0.980, TLI = 0.973, SRMR = 0.041, RMSEA = 0.069).

TABLE 3.

Factor Loading, Composite Reliability, Convergent Validity, and Discriminant Validity

		Item Reliability	Composite Reliability	Convergence Validity	Discriminant Validity
Dim.	Items	STD.LOADING	CR	AVE	IS	IC	ES
IS	3	0.673–0.869	0.829	0.620	0.787
IC	4	0.684–0.896	0.891	0.675	0.492***	0.822
ES	4	0.677–0.856	0.861	0.609	0.357***	0.436***	0.780

The bold fonts in the leading diagonals are the square root of AVEs. Off-diagonal elements are correlations among constructs.

*P < 0.05.

**P < 0.01.

***P < 0.001.

EC indicates enacted stigma; IC, identity concealment; IS, internalized stigma.

Structural Model

The results of the SEM are shown in Figure 2. The model fit values indicate that the model fit the data well (χ²/df = 266.436/59, RMSEA = 0.063, CFI = 0.978, TLI = 0.970, SRMR = 0.044). Experienced more enacted stigma (β = 0.411, P < 0.001) and high internalized stigma (β = 0.179, P < 0.001) were directly associated with IPV perpetration. Higher identity concealment (β = −0.134, P < 0.05) was linked to lower risk of IPV perpetration. Hazardous alcohol use (β = 0.245, P < 0.001) also directly and significantly contributed to IPV perpetration. The model explained 32.8% of the variance of IPV perpetration in MSM, close to the critical value of medium explanatory value (0.33).²⁹

Figure 2.

The path diagram of structural model. Note: The dotted line indicates that the coefficient is not statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001.

We conducted 5000 bootstrapping tests to assess the indirect effect and the 95% CIs, and presented in Table 4. Enacted stigma facilitates to IPV perpetration by promoting hazardous drinking (β = 0.093, P < 0.001), with mediated effects account for 18.83% of the total. In addition, identity concealment reduces the risk of IPV perpetration by decreasing hazardous drinking (β = −0.036, P < 0.05), with the mediated effects representing 21.18% of the total. However, hazardous alcohol use did not serve as a mediating factor in the relationship between internalized stigma and IPV perpetration, which is inconsistent with our hypothesis. Meanwhile, both internalized stigma (β = 0.065, P < 0.01) and identity concealment (β = −0.058, P < 0.05) play partial mediating roles between enacted stigma and IPV perpetration. Moreover, the chain mediation of enacted stigma on IPV perpetration through identity concealment and hazardous alcohol use was also significant (β = −0.016, P < 0.05).

TABLE 4.

Standardized Indirect and Total Effects of Variables on IPV Perpetration

Path	Estimate (95% CI, P Value)
	Standardized Indirect Effect	Standardized Total Effect
Enacted stigma on IPV perpetration through		0.494 (0.402 to 0.583)*
Hazardous alcohol use	0.093 (0.048 to 0.153)*
Internalized stigma	0.065 (0.025 to 0.113)†
Identity concealment	−0.058 (−0.122 to −0.005)‡
Identity concealment to hazardous alcohol use	−0.016 (−0.034 to −0.005)‡
Total indirect effect	0.084 (0.009 to 0.069)‡
Identity concealment on IPV perpetration through		−0.170 (−0.295 to −0.047)†
Hazardous alcohol use	−0.036 (−0.075 to −0.011)‡
Total indirect effect	−0.036 (−0.075 to −0.011)‡

*P <0.001.

^†P <0.01.

^‡P <0.05.

DISCUSSION

Intimate partner violence is a prevalent and varied social issue.² However, the IPV problems among marginalized groups are often overlooked.³⁰ Society's prevention and intervention measures for IPV among MSM are also limited.³⁰ We focus on factors influencing IPV perpetration behavior and wish to provide valuable insights for advancing IPV prevention efforts.

This study found that 18.6% of MSM reported one or more forms of IPV perpetration in the past year, with controlling behavior being the most prevalent type. The prevalence of IPV perpetration in our study was similar to that reported in a study conducted in the United Kingdom, where past-year IPV perpetration was reported at 16.3%.^31s Meanwhile, we assessed the threat of disclosing MSM/HIV status, which had a prevalence of 6.7% over the past year. One study in Chinese MSM showed that 12.4% had experienced the threat of being outed.^32s Therefore, it is essential to pay attention to the unique IPV type among the MSM population and further confirm the precise prevalence rate and impact of it. In the past year, 7.2% engaged in violence against the opposite sex partner. Because of traditional cultural influences, some MSM may enter into a formality marriage (MSM marriage with lesbians or heterosexual women).^33s As a result, their violent victims are not only same-sex partners but also opposite-sex partners. Therefore, it is necessary to recognize the factors that contribute to MSM IPV perpetration and implement appropriate intervention measures.

SEM analysis results showed that both enacted stigma and internalized stigma had direct effects on IPV perpetration, consistent with previous research findings.^34s MSM may learn and imitate violent behaviors through social learning process, particularly when they encounter additional discrimination based on their sexual orientation.^35s Furthermore, internalized stigma is associated with IPV perpetration and mediates the relationship between enacted stigma and IPV perpetration. When MSM face discrimination, the level of internalized homophobia tends to increase, thereby heightening the risk of IPV perpetration.¹³ Consequently, reducing internalized homophobia among MSM could potentially decrease the risk of IPV perpetration.

In our study, identity concealment had a direct negative correlation with IPV perpetration, and our finding was consistent with the study among lesbian women in Turkey.^36s However, contrary to previous research results, lower identity concealment was a protector for IPV perpetration.^1134s Cultural diversity may explain the different result. In China, disclosure of sexual minority status is treated as a rejection to have male heirs, which would be treat as scandal and lead shame for parents and families.^37s If MSM concealed their identity and entered into formality marriages, which could help to reduce pressure to get married and to obey filial piety.^38s,39s Also, our study also showed that identity concealment played a mitigating role in the relationship between enacted stigma and IPV perpetration (β = −0.058, P < 0.05). In a stigmatizing or unfriendly environment, MSM chose to conceal their identities as a protective strategy against discrimination, and this also reduces their connections with peers and decreases the risk of IPV perpetration.^40s,41s

Hazardous alcohol use has a direct impact on IPV perpetration, and reducing hazardous alcohol use may decrease the risk of IPV perpetration. A study conducted in New York City indicated a significant correlation between hazardous alcohol use and IPV perpetration,¹⁷ and our result supports this relationship. Hazardous alcohol use played a meditating role between enacted stigma and IPV perpetration, and also between identity concealment and IPV perpetration. MSM who experienced enacted stigma could deplete MSM's coping resources and increase hazardous alcohol use.^42s,43s Alcohol affects the brain's judgment, focusing attention on provocative stimuli, weakening the inhibition of IPV perpetration, and also increasing the risk of IPV perpetration.^1244s Moreover, identity concealment negatively associated with IPV perpetration via hazardous alcohol use (β = −0.036, P < 0.05). We speculated our findings based on previous studies.¹⁶ MSM individuals with a high degree of identity concealment have reduced contact with MSM communities, thereby experiencing less influence from the community's drinking norms, which could benefit to reducing IPV perpetration. In addition, the chain of relationships that enacted stigma indirectly influences IPV perpetration through identity concealment, and hazardous alcohol use indicates a complex mediation effect (β = −0.016, P < 0.05). However, the small negative effect size should be not be ignored, and replication studies are needed. Our findings suggest that identity concealment is seemingly a “protector” for hazardous alcohol use and IPV perpetration in Chinese MSM. However, concealment has exerted detrimental effects on various mental health outcomes.¹⁶ Moreover, it has also been strongly associated with internalized stigma in the study. Therefore, encouraging individuals to disclose their identity in isolation may not be appropriate. Instead, developing culturally adapted intervention strategies that reduce stigma and promote the reduction of hazardous alcohol use could be more effective in decreasing IPV perpetration.

Several limitations should be noted. First, a cross-sectional survey design, along with a snowball sampling method and the concealed nature of the minority group, restricts extrapolation and causal inferences. Second, IPV perpetration is a highly sensitive issue, and participants may opt to conceal or downplay their abusive behaviors due to feelings of shame, fear of legal consequence, or social stigma, which could lead to reporting bias. Third, in this study, IPV perpetration consisted of various types, but we combined it into a binary variable. It is necessary to conduct more targeted research on specific types of IPV perpetration.

CONCLUSIONS

This study has important implications because it is the first to examine the correlates of minority stress, hazardous alcohol use, and IPV perpetration among Chinese MSM. Enacted stigma, internalized stigma, identity concealment, and hazardous alcohol use are associated with IPV perpetration. Reducing stigma against MSM is a goal that our entire society needs to work toward together. In addition, reducing internalized stigma and increasing awareness of the negative effects of hazardous alcohol use are essential for prevention programs against IPV.

For further references, please see “Supplemental References,” http://links.lww.com/OLQ/B226.