Intersection of Race and Gender in Self-Reports of Violent Experiences and Polyvictimization by Young Girls in Brazil.

OBJECTIVE: To explore racial disparities in self-reports of violent victimization and polyvictimization among young girls in Brazil and to analyze the distribution of prevalence rates across race groups and the estimated odds of exposure.
DESIGN: Data from girls ages 15 and above (N=14,809) from the 2015 edition of PeNSE (National Adolescent School-based Health Survey) were analyzed. Survey weighted proportions and bivariate and multivariate logistic regressions were used to address the objectives.
RESULTS: Independent of the girls' age, socioeconomic status, and region of residence, black, indigenous, mixed, and Asian descendant girls (relative to Whites) were more likely to report past experiences of being bullied, suffering physical violence, forced sexual intercourse, and polyvictimization (reporting all three events). Blacks had the highest rates of reporting being bullied and polyvictimization. Asian descendants had the highest reports of physical violence. Indigenous girls had the highest reports of forced sexual intercourse.
CONCLUSIONS: This study documented disproportionate risks of violent victimization for young women of color among Brazilian students. The risks were significantly greater for those of darker skin tones and from ethnic minorities. IMPLICATIONS: These findings can inform the development of programs to protect young women from violence in Brazil and highlight the importance of including anti-racism strategies in such programs.

Introduction

Although interpersonal violence can affect all, considering the patriarchal aspects of power in most western societies, women are found at greater risk [1]. Violence against women and girls is a human rights violation, and it is also recognized as a public health issue [2-4]. It is a persistent social phenomena that involves psychological, physical, and moral aspects, encompassing use of real or symbolic force [5], and in various parts of the world, women are most victimized during adolescence and youth [2, 6] by their romantic partners, friends, and family members, a reality that is also verified in countries like Brazil [3, 6].

As coined by Brazilian researcher Maria Luiza Heilborn, “time is never neutral” (o tempo jamais é neutro), and this differential incidence of violence against women during adolescence and youth is mostly due to asymmetries in generation, power, and gender across relations in various spaces, such as families, romantic relationships, work environments, and specially schools [7, 8].

In general, violence operates alongside other social markers of difference, aggravating oppression systems against women according to race/skin color, generation, and socioeconomic class. This system of oppression based on gender-class-race puts young women under accumulated forms of violence that can worsen and magnify each other, as they are subject not only to gender violence but also racism and class discrimination [9].

Intersectionality

This connection between social markers such as race, gender, and class as they are applied to an individual or a group to create different forms of discrimination or privilege is what we refer to as intersectionality [10-12]. In studies of inequalities of race, class, and gender, intersectionality is an analytical and methodological tool that offers a lens through which we observe the interactions of different systems of oppression.

As a construct solidified within the scope of black feminism, intersectionality helps us account for the structural aspects of explicit and implicit forms of discrimination by decoding its measurable and unmeasurable components [10-12]. Therefore, we consider intersectionality as a fundamental theory for studies of structural inequalities, as it encourages us to understand the interaction between race and gender, therefore not as attributes of non-white women but as historically created hardships that result from the differential distribution of resources, privileges, and power [13].

Recent research on sexual and reproductive rights as well as gender studies in health have been focusing on intersectional approaches to highlight how oppression and violence against women is racialized since early ages, even through adolescence and youth [9, 12]. In epidemiology, however, there is still a gap in studies with such an approach. There is abundant evidence of higher prevalence of violent victimization against young women of color, especially black girls, a group that is already targeted by racism and therefore experiences intersectional victimization and the accumulation of exposures [9, 12, 14, 15].

Polyvictimization

We focus on polyvictimization, a concept that refers to the experience of multiple victimizations of various types, such as co-occurrences of sexual and physical abuse, bullying, and exposure to domestic violence, not just repeated episodes of exposure to the same sort [16]. This notion of exposure to multiple stressors that intersect and lead to more adverse and long-lasting outcomes is a common understanding in the developmental literature [17-19], and in the past decade, with the formal conceptualization of polyvictimization and the increased interest in the topic, researchers and clinicians are increasingly identifying subgroups of children and youth who are at a greater risk of being polyvictims [16, 20, 21].

It is known that the risk of polyvictimization peaks when children grow into adolescence and youth [19]. In the Brazilian studies of young samples, however, the common approach has been to focus on single violent events separately, and a recent systematic review of Brazilian literature has shown explorations of co-occurrence of violent exposures to be sparse [22].

To date, four studies have explored polyvictimization in youth among Brazilian samples. Barreira and colleagues [23] assessed the prevalence of physical and psychological violence among 302 dating adolescents (ages 15 to 19) from Recife, Brazil, to identify associated factors and the co-occurrence of both types of violence. Their findings reported a prevalence of 19.9% of physical violence, 82.8% of psychological violence, and 18.9% for the co-occurrence. When analyzing these reports by race, the authors presented a binary classification of whites and non-whites, for which no statistically significant difference was found when comparing these two groups.

Oliveira et al. [4] analyzed the evolution of the incidence of sexual violence and co-occurrences between 2001 and 2010 through a study of records of the Guardianship Councils in Feira de Santana, State of Bahia. The study found that 21.7% (N=308) of records involved co-occurrence of sexual and other forms of violence, with a higher prevalence among females (17.3%; N= 246) than males (4.4%; N=62) and most frequently during adolescence. No stratification or covariate adjustment by race was reported.

Faria (2015), in a study of adolescents (M=13.8; SD=1.4) from public schools in the State of Goiânia, found the prevalence of polyvictimization (based on items from the Juvenile Victimization Questionnaire - JVQ, [24]) to be of 19.3% (N=83) for the overall sample and of 24.7% (N=57) among girls against 13.3% (N=26) among boys. However, no stratification by race or ethnicity was presented.

Lastly, in 2017, a study on abuse and maltreatment among adolescent girls from public and private schools in São Paulo and Minas Gerais [7] found the prevalence of co-occurrence of different forms of violence to be higher among students in public schools, considering the type of school as a proxy of socioeconomic status. As in the two aforementioned studies, no data regarding race or ethnicity was incorporated in the analysis.

Although the international literature on polyvictimization has been making important advances in identifying subgroups with intersecting adversities, there is a considerable gap in the Brazilian literature regarding race. Given so, the present study is an exploratory effort focusing on the racial inequalities of exposure to multiple forms of victimization among young girls. To our knowledge, this is the first study on the topic of polyvictimization with a nationally representative sample of Brazilian youth. Furthermore, this work does not intend to delve into the mechanisms of the effect of racism on the risks of victimization; our aim is to analyze the distribution of prevalence rates across race groups and the estimated odds of exposure. Our key hypothesis is that polyvictimization follows the same racial pattern observed in past studies of exposure to single violent events in Brazilian youth, with whites experiencing it the least and brown/mixed, indigenous, and black youth, respectively, experiencing it at significantly higher rates, a skin color gradient in the risk of exposure to racism that follows what has been reported in previous studies [12, 25, 26].

Methods

Design and Survey Population

This study analyzed data from the 2015 PeNSE (Pesquisa Nacional de Saúde do Escolar), a school-based survey conducted by the Brazilian Ministry of Health in partnership with the Brazilian Institute of Geography and Statistics (IBGE – Instituto Brasileiro de Geografia e Estatística). It was designed to monitor health-related factors of children and adolescents enrolled in the 9th grade and in high school in public and private institutions. The survey sample was designed to be representative of the Brazilian population of 9th grade and high school students, separately. In this study, we used only data from the first sample, regarding 9th grade students. A more detailed description of the sampling process and survey design can be found elsewhere [27].

PeNSE data is openly available on the Brazilian Institute of Geography and Statistics (IBGE) website, and the survey was approved by the National Research Ethics Commission (CONEP) under protocol number 1.006.467.

For the purpose of the present study, we analyzed only data from girls ages 15 or above (N=14,809; 14.5%). While there are no universally accepted definitions of adolescence and youth, the UN and the WHO classify adolescents to include persons aged 10–19 years and youth as those between 15 and 24. Our choice to work only with the youth group is justified by the fact that this is the group with the highest rates of exposure to violence in Brazil [28].

Measures

Using a digital device (similar to a palmtop), the students answered a wide range of questions covering socioeconomic status, social support, bullying, nutritional habits, body image perception, asthma, hygiene habits, mental health, work activity, use of health services, oral health, physical activity, substance use (alcohol, drugs, and cigarettes), sexual activity, safety, accidents, and exposure to violence. For this paper, we focus on questions related to bullying, physical violence, and forced sexual intercourse.

Sociodemographic and Socioeconomic Characteristics

Age and geographical regions (North, Northeast, Center-West, South, and Southeast) were used as sociodemographic variables for model adjustment. Socioeconomic status (SES) was defined using the ABEP index criteria (ABEP – Associação Brasileira de Empresas de Pesquisa/Brazilian Association of Research Agencies), which determines a person’s socioeconomic status according to the educational level of the head of the household, possession of various types of household goods (e.g., television sets, desktop, or laptop computer), vehicle, and number of housekeepers. This scale sorts participants into standardized subgroups labeled from A to E (A1, A2, B1, B2, C1, C2, D, E; where A1 represents the highest economic level).

Self-Reported Bullying Victimization, Physical Violence, and Forced Sexual Intercourse

Student participants reported on their experience of different forms of victimization. Participants indicated whether or not they were bullied (“Have you ever been a victim of bullying?”), suffered physical violence (“How many times have you been physically assaulted in the last year?”—answers were dummy coded to “never” or “at least once”), and were forced to have sexual intercourse (“Have you ever been forced to have sex?”). To assess polyvictimization, we created an indicator variable taking the value of 1 if a participant responded positively to all the aforementioned questions and 0 otherwise.

Reasons related to experiences of bullying were verified through the question: “What was the reason your friends have belittled, mocked, scorned, intimidated, or scoffed at you in the last 30 days?” The answers to this question were analyzed according to the following options: (a) my race or skin color; (b) my religion; (c) the appearance of my face; (d) the appearance of my body; (e) my sexual orientation; (f) my region of origin; and (g) other reasons.

Analysis

First, the proportions of respondents who responded positively to each of the victimization questions were computed, with chi-square tests of the two-way tables for race/skin color × region, race/skin color × SES and race/skin color × each of the self-reported violent events (bullying, physical violence, forced sexual intercourse) and polyvictimization. Next, logistic regressions were used to analyze bivariate relationships between race/skin color, SES, age, and region with each of the self-reported violent events as well as polyvictimization. Lastly, a multivariate logistic regression model for race/skin color and the self-reports of violence and polyvictimization was adjusted by age, geographical region, and SES. All statistical analyses were completed in Stata (version 15.0; StataCorp LP, College Station, TX, USA) using the “svy” family of commands to adjust prevalence and variance estimates to account for the sample design and clustering.

Results

The sociodemographic characteristics of the sample are presented in Table 1. 47.28% of the students identified themselves as pardas (mixed/brown), 29.16% as white, 15.28% as black, 4.80% as yellow (of East Asian descent), and 3.29% as indigenous. The average age of the sample was 15.52 (SD= .19). The biggest proportion of participants was classified in the lowest socioeconomic strata (D-E) (50.87%) against a small percentage in the highest (A) (5.53%). Bullying was the most prevalent form of victimization for the total sample (49.88%), followed by physical aggression (18.54%) and forced sexual intercourse (7.13%). A total of 321 (2.49%) participants reported having suffered all three forms of victimization (polyvictimization).

Table 1

Sociodemographic characteristics of the sample and self-report of violent events, by race/skin color. PeNSE Survey 2015 (girls age 15 and above, N=14,809), Brazil

	Black		Indigenous		Parda (mixed/brown)		Yellow (East Asian descent)		White		Missing		Total	x2
	N (2023)	% (15.28)	N (586)	% (3.29)	N (7372)	% (47.28)	N (748)	% (4.80)	N (4065)	% (29.16)	N (16)	% (0.17)	% (100)
Region														124.86*
North	418	7.52	161	14.38	2284	17.03	140	9.33	709	7.47	1	0.55	12.32
Northeast	894	41.57	247	44.32	3077	40.10	357	41.74	1363	25.43	7	27.92	36.24
Southeast	376	37.75	75	23.80	979	30.33	125	33.60	629	32.66	3	53.53	32.12
South	161	8.53	34	8.22	288	5.95	36	7.76	904	28.02	1	7.03	12.94
Center-West	174	4.62	69	8.77	746	6.60	90	7.57	462	6.42	4	10.98	6.37
Socioeconomic position (ABEP criteria)														78.24*
A	76	3.15	43	7.84	395	4.14	67	7.77	443	8.17	3	42.85	5.53
B2	340	15.95	99	15.89	1268	16.29	157	22.38	903	21.07	2	16.65	17.91
C1	405	20.47	120	22.09	1499	21.6	158	20.28	992	26.68	2	8.79	23.42
C2	59	3.01	34	2.90	215	2.71	16	1.82	50	1.16	1	3.07	2.26
D-E	1143	57.42	290	51.27	3997	55.26	350	47.74	1679	40.93	8	28.64	50.87
Bullying	1062	55.02	271	47.63	3578	49.25	377	47.75	2006	48.96	7	25.53	49.88	5.22*
Physical aggression	411	21.43	116	16.36	1342	18.55	173	23.94	707	16.46	1	1.77	18.54	3.10*
Forced sexual intercourse	160	8.67	55	10.41	506	7.28	53	7.51	251	5.70	0	0	7.13	9.63*
Polyvictimization	52	4.27	13	2.07	170	2.50	17	1.97	69	1.68	0	0	2.49	7.72*

*p<0.05

Considering the distribution of race/skin color groups across socioeconomic strata, black girls were more concentrated in the lowest SES groups (C2 and D-E) than girls from other races/skin color (x216= 78.24, p<0.05). As for geographic regions, the stronger concentration of whites and Asian descendants in the South and Southeast regions, against a bigger concentration of blacks and indigenous in the North and Northeast regions (x216= 124.86, p<0.05), follows what is expected from the patterns of racial and ethnic distribution of the Brazilian population.

In the bivariate analysis, girls from the North (OR= 0.70; 95%CI 0.72–0.76), Northeast (OR= 0.64; 95%CI 0.63–0.65), and Center-West (OR= 0.90; 95%CI 0.87–0.93) regions were less likely to report bullying when compared to those from the South region. After the adjustment by race/skin color, age, and SES, lower odds were found for bullying among those from Northeast (OR= 0.61; 95%CI 0.60–0.62), North (OR= 0.72; 95%CI 0.70–0.74), Center-West (OR= 0.87; 95%CI 0.84–0.90), and Southeast (OR= 0.95; 95%CI 0.93–0.97) compared to those from the South region.

For forced sexual intercourse, lower odds in the bivariate analysis was observed for girls from the Northeast (OR: 0.56; 95%CI 0.54–0.58), Southeast (OR= 0.74; 95%CI 0.72–0.77), and Center-West (OR= 0.80; 95%CI 0.75–0.85) regions when compared to the South. When the model was adjusted for the aforementioned covariables, the results showed that those from Northeast (OR= 0.44; 95%CI 0.42-0.45), Southeast (OR= 0.65; 95%CI 0.63–0.68), Center-West (OR= 0.68; 95%CI 0.64–0.73), and North (OR= 0.78; 95%CI 0.74–0.82) were less likely to report forced sexual intercourse compared to those from the South region.

Considering physical aggression, girls from the Southeast (OR= 1.40; 95%CI 1.37–1.44) and Center-West (OR= 1.25; 95%CI: 1.20–1.30) regions reported more physical aggression compared to those from the South region (adjusted estimates). Also, we observed that girls from Northeast (OR= 0.31; 95%CI 0.29–0.33), Center-West (OR= 0.50; 95%CI 0.45–0.56), North (OR= 0.52; 95%CI 0.48–0.56), and Southeast (OR= 0.63; 95%CI 0.59–0.67) regions were less likely to report the occurrence of polyvictimization compared with those from South in adjusted models. For both outcomes, the same pattern was observed in bivariate analysis.

Regarding SES, the results for bullying were not statistically significant in the bivariate analysis. In the adjusted model, a small reduction in the odds of reporting bullying was found for girls in the B2 strata (OR= 0.94; 95%CI 0.91–0.97) and the C2 strata (OR= 0.74; 95%CI 0.70–0.78) when compared to the highest SES strata (A). Results followed a similar pattern for forced sexual intercourse and polyvictimization, with girls from the high SES reporting it at higher rates than girls for the other, lower strata. Lower odds were found in the bivariate analysis for self-reports of forced sexual intercourse in the C1 group (OR= 0.90; 95%CI 0.85–0.96), in the C2 group (OR= 0.79; 95%CI 0.71–0.88), and in D-E group (OR= 0.87; 95%CI 0.82–0.92) when compared to girls from the highest SES group. The comparison between the B2 × A strata only showed statistical significance in the covariate adjusted model (OR= 0.90; 95%CI 0.85–0.96), where all other findings were consistent with those from the bivariate analysis (C1 strata: [OR= 0.83; 95%CI 0.78–0.88]; C2 strata [OR= 0.72 ; 95%CI 0.65–0.80]; D-E strata: OR= 0.81; 95%CI 0.77–0.86]).

Similarly, for polyvictimization, the girls from high SES showed higher odds of reporting having suffered from it, in the bivariate analysis (B2 strata: [OR= 0.78; 95%CI 0.71–0.86]; C1 strata [OR= 0.82; 95%CI 0.75–0.90]; C2 strata [OR= 0.51; 95%CI 0.42–0.62]; D-E strata: OR= 0.80; 95%CI 0.74–0.87]), as well as in the adjusted model (B2 strata: [OR= 0.70; 95%CI 0.63–0.77]; C1 strata [OR= 0.72; 95%CI 0.65–0.79]; C2 strata [OR= 0.50; 95%CI 0.41–0.60]; D-E strata: OR= 0.72; 95%CI 0.66-0.78]).

For physical aggression, findings were statistically significant in the bivariate model for the comparison of strata B2 (OR= 1.27; 95%CI 1.21–1.32) and D-E (OR= 1.05; 95%CI 1.01–1.10) against those in the highest SES strata. These results were consistent for the model adjusted by race/skin color, region, and age, with girls all but the C1 strata reporting higher rates of physical aggression than those of high SES (B2 strata: [OR= 1.22; 95%CI 1.17–1.27]; C2 strata [OR= 1.20; 95%CI 1.12–1.28]; D-E strata: OR= 1.08; 95%CI 1.04–1.12]).

Among all race/skin color groups, black girls were the group with the highest prevalence of bullying (55.02%). When asked about what they believed was the cause for being bullied, 16.19% of black girls reported it to be their skin color/race. This percentage was of 9.79% among indigenous, 4.69% among Asian descendants, 4.23% among brown/mixed race girls, and 2.67% among whites.

Black girls had the highest prevalence of reports of polyvictimization (4.27%). Asian descendants had the highest reports of having suffered physical aggression (23.94%), and indigenous girls had the highest reports of forced sexual intercourse (8.67%).

Logistic regression was used to estimate the associated odds of each of the three forms of violent victimization and of polyvictimization by race/skin color, in the bivariate analysis as well as adjusted by the other measured socioeconomic and demographic variables. Figure 1 shows the odds ratios reported in our adjusted model, to aid in visualization of the findings presented in Table 2. Race/skin color was significantly associated with all forms of victimization and polyvictimization, with the estimated association being stronger for black girls on bullying and polyvictimization in both the crude estimates (Bullying: OR = 1.32; 95%CI 1.29–1.35. Polyvictimization: OR =2.60; 95%CI 2.45–2.77) and in the covariate-adjusted model (Bullying: OR = 1.46; 95%CI 1.42–1.49. Polyvictimization: OR =3.20; 95%CI 3.00–3.42) (Fig. 1).

Fig. 1

Odds ratios (95%CI) bullying, physical aggression, forced sexual intercourse, and polyvictimization by race/skin color. Logistic regression models adjusted by age, geographical region (North, Northeast, Center West, Southeast and South [reference category]), and socioeconomic position (ABEP criteria: levels A, B2, C1, C2, D-E). PeNSE Survey 2015 (girls age 15 and above, N=14,809), Brazil

Table 2

Results of the logistic regressions on self-reported violent events by race/skin color, crude, and adjusted estimates. PeNSE Survey 2015 (girls age 15 and above, N=14,809), Brazil

	Crude estimates				Adjusted model
	Bullying	Physical aggression	Forced sexual intercourse	Polyvictimization	Bullying	Physical aggression	Forced sexual intercourse	Polyvictimization
	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)
Race/skin color
Black	1.32 (1.29–1.35)	1.38 (1.35–1.42)	1.57 (1.51–1.63)	2.60 (2.45–2.77)	1.46 (1.42–1.49)	1.43 (1.40–1.47)	1.81 (1.74–1.88)	3.20 (3.00–3.42)
Indigenous	1.00 (0.96–1.04)	0.99 (0.94–1.05)	1.93 (1.81–2.06)	1.23 (1.08–1.41)	1.14 (1.09–1.18)	1.09 (1.03–1.15)	2.18 (2.04–2.33)	1.55 (1.35–1.78)
Parda (mixed/brown)	1.02 (1.01–1.04)	1.16 (0.96–1.18)	1.29 (1.25–1.34)	1.50 (1.42–1.58)	1.15 (1.13–1.17)	1.24 (1.22–1.27)	1.49 (1.44–1.54)	1.92 (1.81–2.04)
Asian descent	0.98 (0.95–1.01)	1.60 (1.54–1.66)	1.35 (1.27–1.44)	1.17 (1.04–1.32)	1.09 (1.06–1.13)	1.66 (1.60–1.73)	1.50 (1.41–1.60)	1.41 (1.25–1.59)
White	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Region
South	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
North	0.74 (0.72–0.76)	0.91 (0.88–0.94)	0.98 (0.94–1.03)	0.69 (0.64–0.75)	0.72 (0.70–0.74)	0.80 (0.78–0.83)	0.78 (0.74–0.82)	0.52 (0.48–0.56)
Northeast	0.64 (0.63–0.65)	0.79 (0.76–0.81)	0.56 (0.54–0.58)	0.43 (0.41–0.46)	0.61 (0.60–0.62)	0.69 (0.67–0.71)	0.44 (0.42–0.45)	0.31 (0.29–0.33)
Southeast	1.01 (0.98–1.03)	1.40 (1.37–1.44)	0.74 (0.72–0.77)	0.79 (0.75–0.84)	0.95 (0.93–0.97)	1.28 (1.25–1.32)	0.65 (0.63–0.68)	0..63 (0.59–0.67)
Center-West	0.90 (0.87–0.93)	1.25 (1.20–1.30)	0.80 (0.75–0.85)	0.61 (0.56–0.68)	0.87 (0.84–0.90)	1.15 (1.10–1.19)	0.68 (0.64–0.73)	0.50 (0.45–0.56)
Socioeconomic position (ABEP criteria)
A	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
B2	0.98 (0.76–1.26)	1.27 (1.21–1.32)	0.96 (0.91–1.02)	0.78 (0.71–0.86)	0.94 (0.91–0.97)	1.22 (1.17–1.27)	0.90 (0.85–0.96)	0.70 (0.63–0.77)
C1	1.02 (0.81–1.30)	1.04 (1.00–1.09)	0.90 (0.85–0.96)	0.82 (0.75–0.90)	0.99 (0.96–1.02)	1.04 (1.00–1.09)	0.83 (0.78–0.88)	0.72 (0.65–0.79)
C2	0.66 (0.42–1.05)	1.04 (0.97–1.12)	0.79 (0.71–0.88)	0.51 (0.42–0.62)	0.74 (0.70–0.78)	1.20 (1.12–1.28)	0.72 (0.65–0.80)	0.50 (0.41–0.60)
D-E	0.98 (0.77–1.24)	1.05 (1.01–1.10)	0.87 (0.82–0.92)	0.80 (0.74–0.87)	0.99 (0.96–1.02)	1.08 (1.04–1.12)	0.81 (0.77–0.86)	0.72 (0.66–0.78)
Age	0.90 (0.90–0.91)	1.04 (1.03–1.05)	1.28 (1.26–1.29)	1.20 (1.02–1.41)	0.93 (0.92–0.94)	1.08 (1.07–1.09)	1.31 (1.29–1.33)	1.26 (1.23–1.28)

In the bivariate analysis (crude estimates), Asian descendants were more likely to report having suffered physical aggression (OR = 1.60; 95%CI 1.54–.166), and indigenous were more likely to report forced sexual intercourse (OR= 1.93; 95%CI 1.81–2.06), with similar results shown in the models adjusted by age, region, and SES.

Discussion

Our findings showed that bullying was the most commonly reported form of violent victimization among the participants of the study, followed by physical aggression and forced sexual intercourse. The high prevalence of self-reports of bullying, present for nearly half our sample, is consistent with previous research showing that during middle and high school years, girls are under great risk of exposure to this form of aggression [29-31]. It is important to highlight, however, that the prevalence of bullying for the overall sample of girls in the PeNSE survey, as reported by Malta and colleagues [32], was much lower (7.2%) than what was found in our analysis of girls aged 15 and above (49.88%), reinforcing the relevance of studies focused on this subgroup of age and gender.

The socioeconomic inequalities in race/skin color groups reported by our findings, as well as their geographic distribution across regions, in which black and indigenous girls are in greatest proportion in the North and Northeast regions, are aligned with the literature on race and class in Brazil. Black and indigenous populations have been historically overrepresented in the lowest SES groups and in the poorest regions of the country, with stronger indicators of economic inequality and segregation for women [33, 34].

Self-reports of bullying victimization were not homogeneous among race/skin color groups, with a greater prevalence of blacks, followed by brown/mixed girls, indigenous, whites, and Asian descendants. Regarding self-reported reasons for being bullied, black girls were the most likely subgroup that reported to be racially bullied, followed by indigenous and Asian descendants, a finding that follows the trend in the general sample of the PeNSE survey [35], although in a smaller magnitude.

Previous research has already brought attention to school bullying as one of the first encounters with interpersonal racism for young people [36, 37], and literature also differentiates racism and bullying as distinct forms of violence with specific predictors, different consequences, and demanding different policies [38, 39]. In the present study, in fact, when it comes to bullying victimization, the story appears to be one of race, as blacks and indigenous girls, two groups that have been historically oppressed and victimized by racial discrimination in Brazil [40], were indeed more likely than other subgroups to report their race/skin color as a reason for being bullied. These historical processes of discrimination against black and indigenous people are deeply rooted in Brazil’s colonial history, as those were the two ethnic groups majorly victimized and exploited during the founding years of the country. The idea of whiteness as superior to native ethnicities, in case of indigenous groups, and also to black or mixed people is still vastly reproduced within Brazilian society in the form of interpersonal racism [36], manifested here as bullying against school age girls.

This finding is of particular relevance for public health, considering that in other studies of 2015 PeNSE, an association was found between being bullied and mental health-related outcomes such as insomnia, feelings of loneliness, and social isolation (not having friends) [32, 41], and there is abundant evidence of this association from international studies [42-44]. In our sample, the prevalence of bullying among young girls was considerably higher than what was found for the general PeNSE sample and even higher for blacks and indigenous. When developing anti-bullying interventions, our findings suggest that such programs also need to include an explicit anti-racism agenda and to go beyond the individual or classroom level. Aware that the structural aspects of racism are at the very basis of Brazil’s society and history, school staff, administrators, teachers, and students should engage in sustained, proactive efforts to be aware and debate its effects within their institutions, as well as be involved in policies at micro- and macro-levels against racism, microaggressions, and discrimination in the school environment.

Self-reports of physical violence had the highest prevalence among Asian descendants, followed by blacks. Studies of violence against Asian descendant girls in Brazil are scarce; however, episodes of xenophobia and discrimination against Asians in the country are present in media outlets and other documentation sources, [45, 46], have intensified after the coronavirus pandemic [47], and have been reproduced even by authorities within Brazilian government [48]. As an ethnic minority in Brazil, Asian girls face a reality of unbalance of power and lack of representation in school environments when compared to other ethnic groups, a factor that intersects with the oppressions related to gender and can add to their status of vulnerability in regards to violence in general.

Regarding recent exposure to sexual violence, national studies show young women to be the population at the highest risk [49, 50] than older women and than men in any age group. This was verified in this study, since in this subsample of young girls prevalences of self-reported forced sexual intercourse were higher than what was found for the general PeNSE sample and among girls of all age groups [51]. Our findings also showed black and indigenous girls to be at a highest risk than others for reporting sexual violence, more than twice as much when compared to whites, following evidence collected through health data or compulsory notifications [52, 53]. Understanding that, on top of gender-based violence, black and indigenous women have also been historically targeted by racist and colonial processes of hyper sexualization in Brazil [54]. Violence forms that intersect and intensify each other by putting black and indigenous women at a greater risk of suffering sexual violence since young ages, especially after the onset of puberty [55, 56] and throughout their adult life, in the context of intimate partner violence [57].

We hypothesized that polyvictimization would follow the same racial pattern observed in past studies of exposure to single events, with whites experiencing it the least and brown/mixed, indigenous, and black youth, respectively, experiencing it at higher rates. The findings were mostly in the direction of our hypothesis. Polyvictimization was indeed more prevalent among black girls; however mixed/brown reported it at higher rates than indigenous, followed by Asian descendants and lastly whites. Furthermore, in the adjusted model, the odds of polyvictimization was three times higher for black girls than for whites. With this particular finding, our study reinforces evidence that black and brown (mixed) young women in Brazil are experiencing what can be called intersectional violence, since they are at the junction of racial and gender oppressions and are differentially targeted by multiple forms of violence simultaneously [12, 14, 15].

As previously stated in the “Introduction”, no prior research on polyvictimization in Brazil has considered race-specific estimates [4, 7, 22, 23], so we are unable to contextualize such findings in comparison to other Brazilian studies. This highlights the relevance of the current study and of the adoption of intersectionality theory to understand these phenomena and how it impacts young women’s lives in Brazil. Our findings, however, are aligned with those of international studies showing higher prevalences of polyvictimization among black youth and adolescents [58-60].

Limitations

The present study is not without limitations. Data was cross-sectional in nature and thus did not allow for causal inferences. The inclusion of only adolescents who were attending school and were present in the classroom on the day the questionnaire was applied, a decision that may have caused some bias in the results, since absenteeism or school dropout may be related to exposure to violence. As a school-based survey, PeNSE is not representative of the entire Brazilian population of adolescents, only of those who are still in school. Therefore, their percentages of exposure to violence might be lower than those in the general adolescent population. Still regarding the survey characteristics, sample size for Asian and indigenous was relatively small, and this may have influenced the estimated magnitude of the observed association. Although a large sample was used for this study, the number of observations for polyvictimization was quite small for some subgroups.

Lastly, common biases in self-reported data were reported in previous editions of the PeNSE survey [61], and this could have also affected our findings. Considering that most questions used for this study touch on sensitive and traumatic events, relying on self-reported data brings some disadvantages, as subjects may not be able to assess themselves accurately, may give the more socially acceptable answer rather than being truthful, and are also susceptible to recall bias.

Conclusions

This study was an exploratory effort focusing on the racial disparities of multiple forms of violent experiences and polyvictimization among young girls in Brazil. To our knowledge, this is the first to approach the topic of racial disparities on polyvictimization in a Brazilian sample. We found that black and mixed girls were more likely to be bullied, and black and indigenous girls were more likely to report being bullied because of their race/skin color. Asian descendants reported victimization by physical violence at higher rates than other subgroups, and black and indigenous girls had the highest rates of reporting having suffered sexual violence. Lastly, polyvictimization was more prevalent among black girls than all other subgroups and when compared with white girls the odds for this outcome was increased threefold.

No previous research in Brazil has presented race-stratified findings regarding polyvictimization, which represents a novel contribution. We believe those findings not only could be used in the elaboration of protection policies focused on young girls in the country, but also to reinforce that racism is a factor that cannot be left out of the equation when developing such policies.

West (2004) attributes great relevance to the perspective of intersectionality in understanding the greater vulnerability of certain social groups to interactive and/or cumulative forms of victimization [62]. Since markers of oppression (e.g., gender, race and ethnicity, economic) are explanatory factors for the high rates of polyvictimization among historically vulnerable social groups, such as black, indigenous, and Hispanic women, we believe that such perspective is of great relevance, since empirical studies on polyvictimization in Brazil are still not sufficiently considering the role of gender and race.

The lack of such evidence in Brazilian literature can be interpreted in itself as evidence of the racist and patriarchal narratives that are still at the basis of our knowledge of women’s lives and of their encounters with violence. To further advance our understanding of the heterogeneity in the dynamics of violence against women, it is necessary to depart from color-blind approaches and to recognize the racialized nature of these phenomena, as well as its intersections with matters of class and gender.