Student Intervention Against Homophobic Name-Calling: The Role of Peers, Teachers, and Inclusive Curricula

Abstract

Encouraging bystander intervention is an effective strategy to prevent episodes of bullying victimization at school. Yet there remains a paucity of evidence on this behavior in situations of homophobic name-calling, a form of peer victimization aimed at mocking individuals based on their actual or perceived sexual orientation. The existing research has focused on intentionality rather than actual student intervention behaviors, and, of note, no previous studies have taken into consideration contextual factors at the classroom and school levels. The present study examined whether students’ observations of teacher and peer interventions against homophobic name-calling and perceptions of the representation of lesbian, gay, bisexual, and transgender (LGBT) issues in class are associated with student intervention behaviors against homophobic name-calling. A three-level multilevel approach was used to account for the nested nature of students’ experiences in classrooms and schools. The sample included 1,296 students (43.57% girls) recruited from 84 classrooms of 22 Italian public high schools. Preliminary analyses showed that the variability in students’ reports had more to do with which classrooms versus which schools students attend. Results from multilevel regressions indicated that students who observe teachers intervening during episodes of homophobic name-calling, and who perceive the representation of LGBT issues in class as positive, were more likely to intervene against homophobic name-calling and to observe other classmates intervene as well. Also, participants who observed other students intervening were more likely to intervene. These findings highlight the importance of the role of teachers in modeling classroom norms to encourage bystander interventions. Teachers can do so indirectly such as when providing a positive representation of LGBT issues in class, or directly when intervening to condemn episodes of homophobic name-calling. In addition, our results affirm the importance of peer influence in encouraging bystander interventions during episodes of homophobic name-calling.

Introduction

A growing body of evidence indicates that homophobic name-calling (HNC), a specific form of verbal victimization based on the victim’s actual or perceived sexual orientation, is linked to serious negative mental health outcomes including anxiety, depression, and personal distress (Collier et al., 2013; DeLay et al., 2017; Poteat & Espelage, 2007). Encouraging students to intercede in situations of HNC may represent an effective strategy to prevent and reduce such discriminatory behavior (Poteat & Vecho, 2016; Wernick et al., 2013, 2014). However, although student bystander intervention has received increased attention in the general bullying literature (Meter & Card, 2015), there remains a paucity of evidence on this behavior in situations of homophobic bullying victimization (Poteat & Vecho, 2016; Wernick et al., 2013, 2014). Prior studies have found that observing teachers or peers intervening against HNC (Wernick et al., 2013, 2014) and the presence of comprehensive sexuality education (Baams et al., 2017; Dessel et al., 2017) may play an important role in encouraging bystander interventions. However, existing knowledge has limitations. First, the existing research has focused on students’ willingness to intervene when HNC occurs rather than actual student intervention behaviors. Second, prior studies have rarely explored contextual factors associated with bystander interventions or accounted for the multilevel nature of students’ experiences in their classrooms and schools. Third, few studies take a comprehensive approach to include the multiple factors that have been found to be associated with bystander interventions in previous studies.

This study aimed to contribute to the existing literature by proposing a comprehensive model to test whether observing teacher and peer interventions against HNC and a positive representation of lesbian, gay, bisexual, and transgender (LGBT) issues in the curriculum are associated with student intervention behaviors against HNC. By using a three-level multilevel statistical approach to model the nested nature of students in classrooms and schools, we analyzed both participants’ intervention behaviors as well as their observations of peers intervening when HNC occurs.

Homophobic Name-Calling

HNC is a form of peer victimization that involves pejorative labels or denigrating phrases aimed at mocking individuals based on their actual or perceived sexual orientation (Horn, 2007; Poteat & Espelage, 2007). The use of HNC during adolescence could be driven by multiple factors. Some of those are related to the gender socialization process that is the mechanism through which youth learn about and come to adopt norms regarding gender-conforming behaviors (Martin & Ruble, 2010). During this process, youth learn that gender-conforming behaviors tend to be associated with peer rewards (e.g., peer acceptance) and gender nonconforming behaviors with peer punishments (e.g., bullying victimization). As a consequence of gender socialization, HNC may become a way to demonstrate one’s gender conformity, distance oneself from gender nonconformity, and elevate one’s status by regulating gendered behaviors and enforcing traditional gender norms (Ioverno et al., 2019, 2021b; Salvati et al., 2016, 2018). Thus, HNC targeted toward adolescents who are LGBT or gender nonconforming is likely used to punish or belittle them for having deviated from stereotypic masculine or feminine behaviors. Although LGBT students are more likely to be called homophobic names compared to heterosexual students, HNC can be also directed at heterosexual individuals, especially when their gender expressions are considered not conforming to the traditional notions of masculinity and femininity (Ioverno et al., 2019, 2021b; Salvati et al., 2016, 2018). In fact, gender-related attributes are generally used as cues for one’s sexual orientation and negative epithets related to gender nonconforming behaviors are persistent elements of HNC.

The use of HNC is commonplace among adolescents. In 2019, the Gay, Lesbian, and Straight Education Network (GLSEN) reported from a U.S. sample of LGBT middle and high school students that 76% had frequently or often heard homophobic remarks at school (Kosciw et al., 2020). Another survey investigating the experience of 93,000 LGBT adults aged 18 or over across Europe found that 8 in 10 of respondents had heard or seen negative comments or conduct against a schoolmate who was LGBT or perceived as such during their schooling experience (European Union Agency for Fundamental Rights, 2013). A more recent survey conducted on almost 140,000 LGBT and intersex people aged 15 years or older across 30 European countries showed that 91% of respondents heard or saw negative comments or conduct because a schoolmate or a peer was perceived to be LGBT or intersex (European Union Agency for Fundamental Rights, 2020).

HNC is a widespread form of peer victimization, yet the seriousness of its impact is often overlooked (Burn, 2000; Poteat, 2007). Many heterosexual students do not perceive the relation between HNC and homophobic beliefs and use homophobic terms as joking insults (Burn, 2000; Poteat, 2007). An Italian study suggested that in an environment where homophobic language is socially accepted, students would underestimate the seriousness of HNC (Hunt et al., 2016). However, even when HNC is not intentionally related to homophobic beliefs but is nevertheless used as an insult, it has the potential to harm students (Collier et al., 2013; DeLay et al., 2017). There is growing evidence that being a target of HNC significantly predicts anxiety (Poteat & Espelage, 2007), depression (DeLay et al., 2017; Poteat & Espelage, 2007), and personal distress (Collier et al., 2013; Poteat & Espelage, 2007). Moreover, youth who experience HNC are at a high risk of experiencing other forms of peer victimization (Evans & Chapman, 2014) and victimize other peers (Birkett & Espelage, 2015; Ioverno et al., 2021b). One study (Birkett & Espelage, 2015) showed that youth increase their perpetration of HNC in response to the HNC that they experience. A more recent study (Ioverno et al., 2021b) suggested that being the target of HNC could trigger adolescents’ drive to prove their gender conformity, and bullying gender nonconforming victims may be one way to accomplish this goal. Taken together, using HNC contributes to perpetuating homophobic prejudice in the school environment (Carnaghi & Maass, 2007).

Student Interventions Against Homophobic Name-Calling

Like other forms of victimization, HNC is a social phenomenon that often involves multiple roles: a victim, a bully, and different types of bystanders, such as assistants of the bullies, reinforcers who encourage bullying, outsiders who avoid the bullying situation, and the defenders who are supportive of the victims (Salmivalli et al., 1998). In the past decades, the roles of defenders have been acknowledged as crucial for reducing bullying (Salmivalli, 2014) and alleviating the psychological distress of the victims (Kärnä et al., 2010). However, defenders’ interventions are rare (Datta et al., 2016; Meter & Card, 2015) and they are even less frequent in situations of HNC. The last National School Climate Survey (Kosciw et al., 2020) showed that only 6.4% of LGBT students in the United States reported that their peers intervened always or most of the time when hearing homophobic remarks. Similarly, other studies showed that on average students rarely intervene or are willing to intervene to defend victims of homophobic behaviors (Baams et al., 2017; Poteat & Vecho, 2016) and never or rarely witness other students intervening when hearing homophobic or transphobic language (Baams et al., 2017; Wernick et al., 2013, 2014). The fear of being perceived as LGBT and ending up the next victim may explain part of the reason for the lower rates of student bystander interventions in situations of HNC compared to situations of nonbiased name-calling (António et al., 2020; Duhigg et al., 2010). Moreover, interceding in a situation of name-calling may depend on the ambiguity or clarity of the situation (Nickerson et al., 2014). For many bystanders, HNC does not constitute a clear situation of prejudice or emergency (Burn, 2000; Poteat, 2007) and, especially in school environments where the homophobic language is socially accepted, interceding in situations of HNC may be considered unnecessary (Hunt et al., 2016; Lindstrom Johnson et al., 2013). The perception of the victims’ identities as different may also contribute to a lack of peer interventions (Carrera-Fernández et al., 2022). Unresponsive behaviors are often justified with the devaluation or blame of the LGBT or gender nonconforming victims for bringing the discrimination and violence on themselves.

Taken together, minimizing the seriousness of HNC, avoiding, devaluating, or blaming the victims of HNC may represent forms of secondary victimization that contribute to aggravate the suffering of the victims (Carrera-Fernández et al., 2022). This phenomenon refers to the victimization that occurs not as a direct result of the bullying episode but through the negative or unresponsive behaviors toward the victim (Carrera-Fernández et al., 2022). This means that victims of HNC are often doubly victimized: first, by the negative consequences arising from being a target of HNC (i.e., primary victimization), and then by bystanders’ lack of interventions to defend the victims (i.e., secondary victimization).

Predictors of Student Interventions Against Homophobic Name-Calling

There is limited research on individual, interpersonal and contextual factors associated with bystander intervention during episodes of HNC. The existing studies consistently suggest that defenders are likely to be girls, LGBT, and have LGBT friends (Dessel et al., 2017; Poteat & Vecho, 2016). In addition to individual characteristics, peer influence has been found to have a critical role in determining the incidence of intervening behaviors. From a perspective of peer social learning, students intervening in situations of bullying victimization may set an example for other peers to follow (Brechwald & Prinstein, 2011; O’Connell et al., 1999). An experimental study (Paluck, 2011) demonstrated that students who identify peers in their social networks who confront prejudice are more likely to engage in anti-prejudice behaviors. Similarly, two studies conducted in four high schools in Michigan (Wernick et al., 2013, 2014) found that witnessing peers intervening to stop homophobic and transphobic behaviors was associated with students’ intention to intervene.

Characteristics of the school context is another aspect that warrants attention when examining student intervention. For instance, school practices that take place within classrooms, such as teacher responses to situations of HNC and the inclusion of LGBT topics in the school curriculum, may contribute to setting conventional norms regarding how to respond to situations of homophobic behaviors (Baams et al., 2017; Dessel et al., 2017; Wernick et al., 2013). One study conducted in four high schools in Michigan (Wernick et al., 2013) found that students who witnessed adults in their schools intervening against homophobic behaviors were more likely to be willing to intervene as well. Another study on six high schools in the Netherlands showed that students who had comprehensive sexuality education perceived other peers in their schools as more inclined to intervene when hearing HNC (Baams et al., 2017). Finally, a study using a sample of a U.S. Midwestern college students showed that the presence of LGBT content in courses positively correlated with intentions to intervene in situations of homophobic discrimination (Dessel et al., 2017).

The Italian School Context

There are some aspects of the Italian school culture that are worth highlighting to better understand the present study. First, contrary to schools in other cultures, Italian high school students are assigned to a specific class and stay with the same group of classmates for five years (Ioverno et al., 2016). Compared to settings in which there is significant movement and exchange among students across different classrooms, bullying episodes are likely to involve students within the same classroom. Moreover, bystanders are likely to personally know the perpetrator and/or the victim of the bullying episode. Second, secondary school students have the same team of teachers for multiple years. Thus, teacher–student interactions and classroom norms are likely to be more stable and consistent than in other countries where students choose elective subjects and move between classes (Ioverno et al., 2016). Third, in Italy, there is no antidiscrimination law that protects LGBT students or explicitly condemns homophobic behaviors in educational settings. School policies that enumerate protections for students based on sexual orientation and gender identity are not common in Italian schools and there are no available data about schools adopting such policies (IGLYO, 2018). Fourth, the inclusion of LGBT content in the curriculum varies from school to school and from classroom to classroom; addressing LGBT issues remains at the discretion of the teachers (IGLYO, 2018).

In terms of the prevalence of HNC in Italian schools, very little data are available. Research conducted in 23 Italian high schools showed that 48% of students reported that they heard homophobic remarks by schoolmates “often” or “very often” (Ioverno et al., 2016). Another study conducted in six Italian high schools showed alarming percentages of students who heard homophobic comments against gay men (96%) and lesbians (79%) (Prati et al., 2011). Finally, data from the European Union Agency for Fundamental Rights showed that only 9% of Italian LGBT and intersex people never heard or saw negative comments or conduct against a schoolmate who was perceived to be LGBT or intersex (European Union Agency for Fundamental Rights, 2020). To date, no known studies in Italy have examined factors associated with bystander interventions in situations of HNC.

Current Study

This study aimed to examine student-, school-, classroom-level factors associated with bystander interventions during episodes of HNC for 22 public high schools in Rome. The factors include observations of teacher and peer interventions during episodes of HNC and the quality of representation of LGBT issues in class. Given the uncommon nature of bystander interventions, we considered both self-reported participants’ intervention behaviors as well as their observations of interventions from peers during episodes of HNC as outcomes. Based on the literature reviewed, the following hypotheses were tested.

H1: Given that teachers have the potential to model constructive behaviors for their students (Yoon & Bauman, 2014), participants who observed teachers intervening in situations of HNC would be more likely to intervene and to observe other students intervening against HNC.

H2: In line with research showing a significant association between inclusive curricula and safer school climate for LGBT and all students (Baams et al., 2017), participants who perceived that the representations of LGBT issues provided during class are positive would be more likely to intervene during episodes of HNC and to observe other students intervening as well.

H3: Given that students within the same classrooms are likely to have similar school experiences, we expect that aggregated measures of teacher interventions and representation of LGBT issues at the classroom level would be associated with higher rates of both participants’ intervention behaviors and observed peer interventions during episodes of HNC.

H4: Finally, given the crucial role of peer influence in determining adolescent behaviors (Brechwald & Prinstein, 2011), participants who observed other students intervening during episodes of HNC would be more likely to intervene as well.

Methods

Participants

This study analyzed the experiences of 1,470 students (43.57% girls, M_age = 15.93, SD_age = 1.39) recruited from 84 classrooms of 22 public high schools in Rome, Italy. About 8% of the sample self-identified as LGBT and 51.63% had at least one LGBT friend. Student participation varied from 5 to 29 students across classrooms (M = 17.00, SD = 4.68) and from 7 to 193 students across schools (M = 83.90; SD = 52.05). Given that the focus of the study is on student intervention against episodes of HNC, the final analytical sample (N = 1,296) excluded participants who reported never hearing HNC in their classes (N = 174).

Procedure

This study was part of a city-wide project aimed at investigating and preventing homophobic bullying in high schools in Rome, Italy. It was the result of a collaboration between the Municipality of Rome, the university, and local LGBT organizations. All school administrators in Rome were officially invited to take part in this project. Participation was voluntary. In the first phase of the project, exploratory research was conducted in schools to measure and understand the phenomenon of homophobic bullying. Specifically, a survey was administered to students as the first essential step to develop effective intervention and prevention models. In the second phase, students and teachers across schools were presented with data collected during the first phase to stimulate a constructive discussion about homophobic bullying. Finally, students and teachers were invited to create videos, pictures, or stories to send a message against homophobic bullying. In this study, we used data from the first phase. After receiving the approval of the institutional review board, as well as principal approval, packets with study information and consent forms were sent home with students before data collection. Students who provided parental permission and their assent to participate were included in the study. Teachers nominated as coordinators of the project administered the survey during a 30-minute class period. Completion of the survey required 10–15 minutes. Nonparticipating students were given an alternative task to complete. The teacher coordinators were available to aid students in filling in the survey. No compensation was provided.

Measures

Students intervention and peer intervention against HNC.

If students reported that they witnessed episodes of HNC (from rarely to always), two follow-up questions were asked: (a) how often they intervened to stop the episodes by alerting the teachers (student intervention behavior) and (b) how often other students intervene to stop the episodes (peer intervention). Although the student intervention behavior question was specific about alerting teachers, the peer intervention question was worded in a more general way to catch observable peer behaviors and attitudes toward HNC. Response options ranged from 1 (never) to 5 (frequently). Given that student intervention behavior was highly skewed (87.24% of students never intervened during episodes of HNC) we dichotomized this variable (0 = non-intervention, 1 = intervention).

Teacher intervention against HNC.

If students reported that they witnessed episodes of HNC, another follow-up question was asked: how often teachers intervened to stop the episode of HNC. This item was measured on a 5-point Likert-type scale ranging from 1 (never) to 5 (frequently).

Representation of LGBT issues in class.

The representation of LGBT issues in class was assessed through two items. One item inquired whether, during the past academic year, participants attended lessons that included LGBT content (0 = no; 1 = yes). A follow-up question asked how positive or negative the representation of LGBT issues was answering the following prompt: “Do you think that in those lessons the representation of lesbian, gay, bisexual or transgender people was.” Response options ranged from 1 (very negative) to 5 (very positive). Given that only the respondents who answered “yes” to the first question could report on the quality of LGBT representation, these two items were coded in three dummy variables from the answers “very negative” or “negative” (negative LGBT representation), the answer “neutral” (neutral LGBT representation), and the answers “positive” and “very positive” (positive LGBT representation). In analyses, the absence of LGBT content in class served as a reference group.

Student-level covariates.

Students reported on their age and sex (0 = female, 1 = male). To assess the number of LGBT friends, the survey asked “how many lesbian, gay, bisexual, or transgender friends do you have?” Response options were 0 (none), 1 (one), 2 (from 2 to 4), 3 (from 5 to 10), and 4 (more than 10). Students’ sexual orientation was assessed through responses to a single item including the following categories: “Heterosexual”, “Gay/Lesbian”, “Bisexual”, “I’m not sure about my sexual orientation”, “Other”. From this item, a dichotomous variable was created (0 = heterosexual; 1 = LGBQ+). The frequency of hearing HNC was measured through the question “How often have you heard homophobic remarks (such for example, ‘faggot’) or the use of the word ‘gay’ or ‘lesbian’ in a derogatory manner from your classmates?” Response options were on a 5-point scale: 0 (never), 1 (rarely), 2 (sometimes), 3 (often), 4 (frequently). Students who reported that they had never heard HNC were excluded from the analyses. Thus, in the final analytical sample, the response options ranged from 1 to 4. A follow-up question asked how often teachers were present during episodes of HNC. Response options ranged from 1 (never or rarely) to 5 (frequently).

Classroom and school context.

School and classroom aggregated means were calculated on student-level variables. Specifically, aggregated scores were calculated on the frequency of hearing HNC, teacher intervention and presence during HNC, and representation of LGBT issues in class or in the school. As contextual variables, the three dummy variables reporting on the quality of the representation of LGBT issues represented the percentage of students in class or school perceiving the LGBT content as positive, neutral, or negative (the percentage of students reporting no LGBT content in class was used as a reference). Analyses were adjusted for classroom characteristics such as the number of students and the percentage of female and male students. The student-level sexual orientation variable was used to create a percentage of LGBQ+ students per class.

Plan of Analysis

First, Pearson’s bivariate correlations were performed to examine the associations among the student-level variables (Table 1). In the preliminary analyses, intraclass correlation coefficients (ICC) were used to examine the proportion of total variance in student-level variables across classrooms and schools (Table 1). Two series of multilevel regression analyses were tested to examine the variables associated with student intervention and observed peer intervention against HNC. Three-level multilevel analyses were employed to account for the nested nature of the data and to disentangle student-level effects from classroom-level and school-level effects. Given that parsimonious models are preferable, only significant aggregated scores were retained in the final models (Raudenbush & Byrk, 2002). Wald Chi-Square tests were used to cross-validate the different solutions across nested models. A significantly larger Wald Chi-Square suggests a better model fit. Continuous variables were classroom-mean centered. The complete case analyses resulted in a loss of a very small portion of the sample (from 1.16% to 4.17%), and missing values did not substantially limit the statistical power of the analyses. Thus, missing data were handled with listwise deletion following the recommendations of Allison (2010).

Table 1.

Student-level Correlations and Descriptive Analyses.

	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.
1. Student intervention	1.00
2. Peer intervention	0.14***	1.00
3. Teacher intervention	0.15***	0.28***	1.00
4. Teacher presence	0.13***	0.05	0.22***	1.00
5. Negative LGBT repr.a	−0.03	0.00	0.03	0.12***	1.00
6. Neutral LGBT repr.a	−0.02	−0.03	−0.01	0.00	0.20***	1.00
7. Positive LGBT repr.a	0.08**	0.13***	0.12***	−0.04	−0.15***	−0.35***	1.00
8. Hearing HNC	0.11***	0.01	0.07*	0.39***	0.10***	−0.01	−0.04	1.00
9. LGBQ+b	0.09**	0.05	−0.04	0.03	−0.03	−0.04	0.02	0.04	1.00
10. Malec	−0.03	−0.17***	0.03	0.11***	0.04	0.03	−0.10***	0.13***	−0.07*	1.00
11. Age	−0.01	−0.03	0.02	0.05	−0.02	0.03	−0.04	0.02	0.00	0.03	1.00
12. LGBT friends	0.08**	0.24***	−0.01	0.07*	−0.04	−0.01	0.10***	0.01	0.27***	−0.22***	0.09***	1.00
%/M(SD)	12.72%	2.11 (1.07)	2.19 (1.24)	1.95 (1.02)	7.95%	31.30%	21.11%	2.66 (1.09)	8.22%	56.43%	15.93 (1.39)	1.94 (1.08)
Classroom-level ICC	0.06	0.08***	0.08***	0.10				0.09**	0.1 1*	0.27**	0.70***	0 17***
School-level ICC	0.02	0.00	0.01	0.07**	0.10**	0.11		0.03	0.00	0.16*	0.35*	0.05

Note.

^*p < .05;

^**p < .01;

^***p < .001;

HNC = homophobic name-calling. ICC = Intraclass correlation. Reference groups are:

^aNo LGBT representation in class;

^bHeterosexual,

^cFemale. ICCs for Negative, Neutral and Positive LGBT representation in class were computed using multilevel multinomial logit models.

Results

Preliminary Analyses

Before testing the multilevel regression models, we examined the correlations among the study variables (Table 1). Student intervention and witnessing peer intervention were significantly and positively correlated with teacher intervention, a positive representation of LGBT issues in class, and having LGBT friends. In addition, student intervention behavior was significantly correlated with peer intervention, the presence of teachers during episodes of HNC, the frequency of hearing HNC in class, and being LGBT. Peer intervention was correlated with students’ sex: female students were more likely to observe their peers intervening than male students. No issues of multicollinearity were detected among the other study variables.

Then, we tested a series of unconditional multilevel models to establish the degree to which the student-level variables varied across classrooms and schools (Table 1). Results showed that the variance of the majority of the variables lied more at the classroom level than at the school level except for the presence of teachers during episodes of HNC, which significantly varied only across schools. The ICCs showed that the variation of student intervention across schools and classrooms was not significant. On the contrary, variations in class means accounted for about 8% of the total variability in peer intervention, whereas the variation across schools on this variable was not significant. Finally, there was a considerable variance in the composition of the classrooms in terms of sexual orientation, gender, age, students having LGBT friends. Age and gender composition also significantly varied across schools.

Witnessing Peer Interventions Against Homophobic Name-Calling

Two multilevel linear regression models were tested to examine the associations of teacher intervention against HNC and quality of the presentation of LGBT issues in class with observing peer intervention against HNC (Table 2). Because preliminary analyses showed that peer intervention significantly varied only across classrooms but not across schools, the models did not include school-level aggregated scores as predictors, and random slopes across schools were not tested.

Table 2.

Multilevel Linear Regression Models of Classroom and Individual Variables Associated With Witnessing Peer Intervention Against Episodes of Homophobic Name-Calling.

	Peer Intervention Against HNC
	Model 1 B(SE)	Model 2 B(SE)
Intercept	1.20(0.43)**	1.86(0.54)**
Classroom-level
• % boys	0.01(0.01)	0.01(0.01)
• N students	0.01(0.01)	0.01(0.01)
• % LGBT students	0.01(0.01)	0.01(0.01)
• Hearing HNC	0.08(0.13)
• Teacher intervention	0.23(0.10)*	0.18(0.08)*
• Teacher presence	−0.22(0.15)
• Negative LGBT repr.a	0.20(0.43)
• Neutral LGBT repr. a	0.12(0.26)
• Positive LGBT repr. a	0.46(0.26)
Student-level
• Hearing HNC		0.01(0.03)
• LGBQ+b		0.02(0.11)
• Malec		−0.24(0.06)***
• Age		−0.03(0.02)
• LGBT friends		0.19(0.03)***
• Teacher intervention		0.25(0.02)***
• Teacher presence		0.01(0.03)
• Negative LGBT repr.d		0.07(0.11)
• Neutral LGBT repr. d		0.01(0.07)
• Positive LGBT repr. d		0.16(0.08)*
Student-level variance	1.05(0.04)***	0.91(0.04)***
Classroom-level variance	0.06(0.02)**	0.05(0.02)**
School-level variance	0.01(0.01)	0.01(0.01)
Wald χ2 (df)	25.13(9)**	226.02(14)***
Δ Wald χ2(df)		182.97(5)***

Note.

^*p < .05;

^**p < .01;

^***p < .001;

HNC = homophobic name-calling. Reference groups are:

^apercentage of students reporting no LGBT representation in class;

^bHeterosexual,

^cFemale,

^dNo LGBT representation in class.

We first evaluated a model with classroom-level aggregated means of HNC, teacher presence and intervention, and representation of LGBT issues in class (see Model 1 in Table 2). Analyses were also adjusted for other classroom characteristics (the percentage of male students, the number of students, and the percentage of LGBQ+ students). Results showed that among all classroom-level variables, only teacher intervention was significantly associated with more peer intervention.

Then, all student-level variables were entered in Model 2. For the sake of parsimony, only classroom characteristics and the teacher intervention aggregated scores from Model 1 were retained. The chi-square difference test suggested that Model 2 fit the data significantly better than Model 1 (see Wald Chi-Square difference in Table 2). Results showed that female students with LGBT friends were more likely to observe peer interventions against HNC. Moreover, consistent with our hypotheses, students observing teachers intervening against HNC and perceiving the representation of LGBT issues in class as positive were more likely to observe peer interventions.

Finally, we tested a series of multilevel regression models allowing the slopes of the key predictors to vary across classrooms. Results indicated that the random slopes of teacher intervention, Δχ² = 1.19, df = 1, p = .275, and representation of LGBT issues, Δχ² = 1.57, df = 3, p = .667, did not significantly improve the models, supporting the use of a random intercept model over a random slope model.

Student Interventions Against Homophobic Name-Calling

Two similar multilevel regression models were tested to examine whether participant intervention behaviors were associated with observing peers and teachers intervening against HNC, and with the quality of the representation of LGBT issues in class. Specifically, multilevel logistic regression models were fitted to account for the dichotomous nature of the outcome variable. Results are presented in Table 3. Preliminary analyses showed no significant variations across classrooms and schools in student intervention, suggesting that most of the variance lied in student-level variables. Thus, aggregated scores were not entered in the models and random slopes were not tested.

Table 3.

Multilevel Logistic Regression Models of Individual Variables Associated With Student Intervention Against Episodes of Homophobic Name-Calling.

	Peer Intervention Against HNC
	Model 1 OR(SE)	Model 2 OR(SE)
Intercept	0.05(0.07)	0.04(0.06)
Classroom-level
• % boys	0.99(0.01)	1.01(0.01)
• N students	1.01(0.03)	1.01(0.03)
• % LGBT students	1.01(0.01)	1.01(0.01)
Student-level
• Hearing HNC	1.26(0.12)*	1.26(0.12)*
• LGBQ+a	1.98(0.59)*	1.97(0.59)*
• Maleb	0.87(0.18)	0.92(0.19)
• Age	1.04(0.08)	1.05(0.08)
• LGBT friends	1.18(0.10)	1.14(0.10)
• Teacher intervention	1.40(0.11)***	1.33(0.10)***
• Teacher presence	1.18(0.11)	1.19(0.11)
• Negative LGBT representationc	0.73(0.29)	0.72(0.29)
• Neutral LGBT representationc	1.07(0.24)	1.08(0.25)
• Positive LGBT representationc	1.69(0.40)*	1.64(0.40)*
• Peer intervention		1.25(0.11)*
Classroom-level variance	0.09(0.13)	0.10(0.14)
School-level variance	0.14(0.10)	0.14(0.10)
Wald χ2 (df)	57.74(13)***	62.50(14)***
Δ Wald χ2(df)		6.10(1)*

Note.

^*p < .05;

^**p < .01;

^***p < .001;

HNC = homophobic name-calling. Reference groups are:

^aHeterosexual,

^bFemale,

^cNo LGBT representation in class.

Model 1 included student-level variables controlling for classroom characteristics (percentage of boys, number of students, and percentage of LGBQ+ students). Results show that being LGBT and the frequency of times students hear HNC was significantly associated with student intervention. Consistent with our hypotheses, students who observed teachers intervening against HNC and who perceived the representation of LGBT issues in class as positive were more likely to intervene against HNC. In Model 2, observing peer interventions was entered as a predictor. The chi-square difference test suggested a better fit for Model 2 compared to Model 1. As expected, peer intervention against HNC was associated with participant student intervention against HNC.

Discussion

Our findings add to previous research in the homophobic bullying literature showing that students who observe teachers intervening during episodes of HNC, and who perceive the representation of LGBT issues in class as positive, were more likely to intervene against HNC and to observe other classmates intervene as well. In addition, participants who observed other students intervening were also more likely to intervene.

In the preliminary analyses, consistently significant ICCs for classrooms and nonsignificant ICCs for schools were observed in most of the measures. This is not surprising since students in Italian high schools are generally assigned to specific classes and tend to stay with the same group of classmates for all five years of school. Thus, variability in experiences has more to do with which classrooms versus which schools students attend. Notably, and contrary to the other measures, the ICC for teacher presence during episodes of HNC was significant at the school level but not at the classroom level. This variation could be an expression of the different levels of tolerance of homophobic language across schools. Consistent with this interpretation, previous studies conducted in Italy suggested that homophobic prejudice was widespread among teachers in several high schools (Baiocco et al., 2020) and homophobic remarks were often used also by school staff (Ioverno et al., 2016). Also, an unexpected positive correlation between teacher presence and the frequency of hearing HNC was found in the preliminary analyses. This result may suggest that HNC tends to occur especially during class time as students in the Italian context are rarely left alone without a teacher present during the school day. Thus, the more frequent HNC is, the more likely the teacher is to be present in class when HNC occurs.

Another interesting finding from preliminary analyses was that participant students’ reports of their intervention behaviors did not significantly vary across classrooms and schools. This result is probably due to the small number of participants reporting to intervene during episodes of HNC (12.72%). Some previous studies have documented that intervention against HNC is not a common behavior among students (Poteat & Vecho, 2016; Rinehart & Espelage, 2016). However, other studies found a higher incidence of student intention to intervene against HNC: In a sample of high school students from the Netherlands, Baams et al. (2017) found that 89% reported the intention to intervene in situations of HNC. In other two studies conducted in four high schools in Michigan, Wernick et al. (2013, 2014) found that, on average, students reported that they would intervene sometimes or often to stop episodes of homophobic and transphobic remarks. Our findings, together with those of previous studies reviewed, indicate that an important distinction should be made between behavioral intentions and actions with respect to intervening in episodes of HNC.

It is also worth noting that the preliminary correlations and the final regression model suggested that LGBQ+ students were more likely to alert teachers than heterosexual students during episodes of HNC. This finding may suggest that LGBQ+ students are more motivated to intervene because they are more likely to perceive the seriousness of HNC compared to their heterosexual peers. In fact, HNC is a form of victimization that directly targets their identities. Another interpretation is that many LGBQ+ students are victims of HNC and become aware of the negative impact of this form of victimization (European Union Agency for Fundamental Rights, 2013, 2020; Kosciw et al., 2020). As a consequence, they may feel a greater responsibility for providing help and be more aware of how to intervene.

In line with our first hypothesis, we found that more teacher intervention during episodes of HNC was significantly associated with both more participant intervention and observed peer intervention during episodes of HNC. In accordance with the present results, previous studies showed that students who observed teachers intervening when hearing homophobic or transphobic language reported more willingness to intervene as well (Wernick et al., 2013, 2014). The present study adds to the existing research by suggesting that teacher intervention may encourage both students’ intention and action of intervening against HNC. This also accords with findings from research on general bullying providing evidence that teachers influence prosocial classroom norms (Yoon & Bauman, 2014). Accordingly, when teachers repeatedly ignore HNC incidents they may indirectly communicate a lack of concern about these behaviors and the expectation that students will not be assisted by teachers. On the contrary, teacher interventions may communicate clear expectations that HNC behaviors are unacceptable. This is particularly important considering that, in many school cultures, homophobic language is socially accepted and the seriousness of HNC is minimized (Hunt et al., 2016). In such school cultures where responding to HNC is not socially expected, students may feel less motivated and supported to intervene.

The hypothesis that a positive representation of LGBT issues in class is associated with more student interventions against HNC was also confirmed at the individual level. This finding is in line with previous studies showing that the presence of comprehensive sexuality education at school is associated with more students’ willingness to intervene when hearing HNC (Baams et al., 2017; Dessel et al., 2017). However, our attention to the quality of the representation of LGBT issues in class and their associations with the actual student interventions rather than student intentions to intervene, add nuance to the existing literature. The observed higher likelihood of student interventions in the presence of positive representations of LGBT issues could be related to the documented impact that LGBT inclusive curricula have in creating a safe school climate for LGBT and all students (Baams et al., 2017; Snapp et al., 2015). A positive representation of LGBT issues may offer not only a sense of validation for LGBT students but an opportunity to understand the experiences related to the different sexual and gender identities (Snapp et al., 2015). This may contribute to reduce the tolerance for prejudicial attitudes and raise awareness about the seriousness of HNC (Baams et al., 2017), which may in turn help increase bystander interventions during episodes of HNC.

This study attempted to examine how student interventions were associated with teacher intervention and representation of LGBT issues in class at three distinct levels: individual, classroom, and school levels. However, the variation between schools and classrooms was not significant enough to warrant this approach. Only for peer intervention did significant variation between classrooms allow us to include classroom-level aggregated scores. Yet consistent with our expectations, aggregated teacher intervention was significantly associated with peer intervention. This result suggests that, in classrooms where many students observe teachers intervening against HNC, students are more likely to observe peers intervene against HNC compared to classrooms where teacher interventions are less frequent. Interestingly, the coefficient for the individual-level observation of teacher interventions was significantly stronger than the classroom-level coefficient. This difference may suggest that the experience of observing teachers or peers intervening during episodes of HNC is not always shared with all classmates. It may be that responses to HNC can also be observed outside of the classrooms (e.g., in the hall, in the gym, in front of the school). This finding corroborates the importance of teacher interventions also outside of class time (e.g., before or after classes or during recess) in encouraging student intervention against HNC.

Contrary to the student-level perception of LGBT representation in class, the classroom-level LGBT representation was not significantly associated with more student interventions. This finding may be explained by the subjectivity of the perception of the LGBT representation in class. Depending on the personal sensitivity to LGBT issues, students may perceive LGBT content in class in different ways. For instance, the preliminary correlations showed that female students and students with a great number of LGBT friends in our sample tend to perceive the LGBT content in class as more positive. It conceivable that these students, who tend to be supportive of LGBT issues compared to their peers (Herek & McLemore, 2013), may be more likely to perceive the importance of including LGBT content in their school curriculum and to interpret this inclusivity more positively. This may explain why a positive representation of LGBT issues in class was significantly associated with student interventions only when considering the individual-level perception of the representation quality.

Finally, our hypothesis that students would be likely to intervene when witnessing their peers intervene was also confirmed. There are similarities between these associations and those described by previous studies on students’ intention to intervene (Baams et al., 2017; Wernick et al., 2013, 2014). This finding may be explained by the role of peer influence as a hallmark of adolescent psychological functioning. Several theoretical frameworks argued that certain behaviors and attitudes are more likely to be expressed when they are recognized or valued by others and when they offer a means of socially connecting with the peer group (see Brechwald & Prinstein, 2011 for a review). Likewise, students may be more likely to engage in intervention behaviors when they are perceived as consistent with classroom norms, that is the students’ perceptions of one another’s values and behaviors. Thus, because HNC occurs within the context of a peer audience, how bystanders respond may model the classroom norms on behaviors to adopt during episodes of HNC. Moreover, from a perspective of peer social learning (Brechwald & Prinstein, 2011), by observing other students intervene, youth may be more likely to learn the means of intervention in situations of HNC and evaluate whether these means have the desired effect. Presumably, these evaluations may encourage students who are already willing to intervene to defend the victim of HNC.

Limitations

This study has a number of limitations. First, data were cross-sectional and unable to address causal associations. Longitudinal research would be necessary to determine directionality. Moreover, a quasi-experimental design comparing student intervention behaviors before and after a teacher training on how to intervene during episodes of HNC and how to talk about LGBT issues in class would provide stronger evidence on the role of teachers in student intervention behaviors. Second, given its uncommon nature, there was no significant variation in student intervention across classrooms and schools. Thus, we were unable to investigate school- or classroom-level associations on this outcome. A significant variation may have been detected with a larger number of schools and classrooms. However, a third-person perspective on student intervention obtained through our peer intervention item allowed us to detect classroom-level variations in this variable, suggesting the crucial role of shared classroom-level experiences in determining defending behaviors among students. Third, we only assessed whether LGBT issues were covered in class and the students’ perceptions of the LGBT representation. To identify effective didactic strategies aimed at increasing student interventions against HNC, it would be important to objectively assess the content and the timing of these representations, and the supportiveness of teachers in providing information about LGBT issues. Forth, we relied on single-item measures. The survey length precluded the inclusion of multi-item scales. Student intervention constructs used in this study represent observable behavior and therefore could be validly measurable by single items, yet future research could employ more robust measures of the constructs used in this study. Fifth, we used a sample of schools from a single Italian city. Generalization of the results to school contexts of other countries should be done with caution. Sixth, the survey did not include questions about transgender and nonbinary gender identity. Although there is a greater acceptance of the introduction of questions on sexual orientation in Italian surveys for adolescents, questions about gender identity are still considered too sensitive by many school administrators or parents in Italy. The acknowledged risk of adverse reactions to the introduction of such questions would have impacted the response rates or quality of data. Such risks limited our ability to fully examine the role of students’ gender identity. Although this topic was not the main focus of the present study, future studies should consider giving specific attention to participants’ transgender, gender nonbinary or cisgender identities when analyzing student interventions. As HNC targets youth who deviate from gender-conforming behaviors, there is reason to believe that transgender and gender nonbinary students would be more likely to alert teachers than their cisgender peers. Seventh, the project was voluntary for schools and required parental permission, and thus there may be some selection bias in that the schools that were willing to participate, and parents who consented to participation by their children, may have had more affirming attitudes about LGBT issues in school. Future research should examine whether the present findings hold across different levels of school inclusivity.

Implications

Our study points to a number of implications for educational practices and future research on bystander intervention during episodes of HNC. First, significant ICCs were detected mostly at the classroom level rather than at the school level. Although this finding is not surprising given that students in the Italian school system are assigned to, and spend the majority of their time in, specific classrooms, similar variations may be expected in contexts where students choose elective subject classes. In fact, specific classroom characteristics such as student composition and the quality of teacher practices may significantly affect the frequency of bullying and defending behaviors, and the classroom norms encouraging or inhibiting these behaviors. Thus, future studies or school interventions aimed at preventing the use of homophobic language should consider using a cluster design including classroom level as one of the organizational units.

Second, another implication refers to the different ways of measuring student intervention. Whereas some studies have focused on student intention to intervene (Baams et al., 2017; Wernick et al., 2013, 2014), other studies (Poteat & Vecho, 2016; Rinehart & Espelage, 2016) like the present one have focused on actual student intervention behavior. Taken together, these studies showed that the number of students reporting the intention to intervene in episodes of HNC was considerably larger than students reporting actual intervention behavior. Future research should investigate the connection between defending behavioral intention and behavior itself and identify strategies and practices to facilitate the shifts from intention to action.

Third, this study corroborates previous studies suggesting that teacher training on sexual orientation and gender identity would improve school climate for all students (Day et al., 2019; Ioverno et al., 2021a; Russell et al., 2016). Teacher training should stress the importance of intervening in situations of HNC and provide clear procedures on how teachers should respond in order to set an example for students to follow. Moreover, it may be important for teachers to realize that the way they represent LGBT issues in class may affect the way students think about using homophobic language and the seriousness of HNC.

Finally, our finding showed that students may be more likely to engage in intervention behaviors when they observe other students intervening. Providing clear reporting procedures in situations of HNC could help to increase the chances of student interventions and to model classroom norms on how to respond to HNC. Thus, discussing with students and clarifying the classroom norms concerning the use of homophobic language may represent a starting point for redirecting peer influence regarding intervention behaviors.

Conclusion

HNC is a widespread form of peer victimization linked to serious negative mental health for the victims (Collier et al., 2013; DeLay et al., 2017; Poteat & Espelage, 2007). Encouraging students to intervene when HNC occurs could represent an effective strategy to reduce the frequency of this form of peer victimization. Yet, very little attention has been paid to the factors associated with student intervention. The available research has focused on the willingness to intervene rather than on actual intervention behavior, and prior studies have been limited to analyses at the student-level without taking into consideration contextual factors at the classroom and school levels (Baams et al., 2017; Wernick et al., 2013, 2014). This study aimed at addressing these gaps by testing a comprehensive and multilevel model of how individual, classroom, and school-level factors are associated with student bystander interventions against HNC.

The findings reiterate the importance of the role of teachers in modeling classroom norms to encourage bystander interventions. Teachers can do so indirectly such as when providing a positive representation of LGBT issues in class, or directly when intervening to condemn episodes of HNC. Importantly, our results affirm the importance of peer influence in encouraging bystander interventions during episodes of HNC. Thus, empowering students by defining clear reporting procedures in situations of HNC or clearly defining the classroom norms on how to respond to homophobic language may contribute to improving the school climate and reducing episodes of homophobic victimization (Ioverno & Russell, 2021a, 2021b).

Biographies

Salvatore Ioverno, PhD, is a Postdoctoral Research Fellow in the Department of Sociology at Ghent University. His research interests include the identification of risk and protective factors for minority stress, mental health, and school achievement among lesbian, gay, bisexual, transgender, queer, and questioning students using cross-national methodologies.

Maria Rosaria Nappa, PhD, is a Postdoctoral Research Fellow in the Department of Law, Economics and Human Sciences (DIGIES), Mediterranea University of Reggio Calabria. Her research is focused on the role of teachers in preventing episodes of homophobic bullying at school.

Stephen T. Russell, PhD, is Priscilla Pond Flawn Regents Professor in Child Development and the Department Chair, Human Development and Family Sciences at The University of Texas at Austin. His research is focused on adolescent development with an emphasis on adolescent sexuality, LGBT youth, and parent-adolescent relationships.

Roberto Baiocco, PhD, is a Professor of Developmental Psychology at the Sapienza University of Rome. His research is focused on minority stress and the well-being of sexual and gender minorities.