Background

Race-based discrimination is a common occurrence in the lives of youth from historically marginalized groups and has a wide range of negative implications for their physical and psychological well-being (Trent et al., 2019). Particular interest has been paid to race-based discrimination’s association with exposure to and use of different forms of violence victimization and perpetration (Reed et al., 2010; Stueve & O’Donnell, 2008; Waltermaurer et al., 2006). However, many of these studies have been conducted among adult populations, and few explore how race-based discrimination exposure is associated with outcomes of violence among youth. In one relevant research study, among 550 Black (80.4%) and Latina (13.8%) 19-20-year-old women, the authors demonstrated that discrimination was associated with physical and emotional violence victimization and perpetration (Stueve & O’Donnell, 2008). Among this sample, discrimination (OR = 1.22, 95% CI 1.07–1.39) increased the odds of emotional victimization. Studies on how racism impacts violence among youth are needed to improve prevention efforts that address systemic root causes and mitigate the long-lasting effects of experiencing discrimination. Such research also requires careful attention to how experiences of race-based discrimination are measured.

Measures used to assess race-based discrimination among youth have not consistently been centered around their lived experiences. Numerous studies have instead adapted existing adult measures (Cheah et al., 2020; Goosby et al., 2018; Okamoto et al., 2009; Riina & McHale, 2012). For example, in a study examining relationships between discrimination and substance use among youth, (Okamoto et al., 2009) discrimination was assessed using an adapted version of the Everyday Discrimination Scale (EDS), (Williams, 1997) a measure originally developed to assess interpersonal discrimination experienced by adults. The findings of this study suggest that perceived discrimination measured through the EDS was associated with significant increased odds of lifetime and past 30-day tobacco smoking, alcohol, and marijuana use. Examples of measures developed specifically for youth include the Adolescent Discrimination Distress Index (Fisher et al., 2000) and the Index of Race-Related Stress for Adolescents, (Utsey et al., 2006) both of which have limitations because they validated scales on small samples from geographically homogenous areas and may not capture the nuances of more recent experiences of historically marginalized youth exposed to race-based discrimination in the United States. While applicable to the adolescent populations where the scale was regionally validated, applying these scales in diverse geographies may limit the ability of the scale to effectively measure exposure. Additionally, it may be important to evaluate the validity of a scale in the population among which will be newly applied prior to broad dissemination and to ensure accurate interpretation (Clark & Watson, 2016).

One of the most widely used scales developed specifically for youth is the Perceptions of Racism in Children and Youth Scale (PRaCY)(Pachter et al., 2010; Pachter, Szalacha, Pachter et al., 2010a, b). Pachter and colleagues created and validated this unidimensional scale among 277 Latinx, African American, West Indian or Caribbean and other multi-cultural 7-18-year-old youth from two New England cities. Validation of the PRaCY in the original studies was robust, examining face validity, construct validity through confirmatory factor analyses, differential item functioning by age, sex, and ethnicity, and convergent validity through the association between the scale and depression and anxiety. While the evidence from this rigorous analysis is strong, the psychometric properties of a scale should be evaluated for each new population with which it is used and additional validation analyses should be conducted to confirm the accuracy of the measurement tool assessing racialized experiences of youth (Arafat et al., 2016; Boateng et al., 2018; Harnois et al., 2020). By validating this scale in a new sample of historically marginalized youth, we can better compare results across populations and geographic regions and improve uptake of a scale that accurately describes the experiences of race-based discrimination among racially diverse youth. For example, in their examination of the construct validity of the Everyday Discrimination Scale among a sample of geographically and ethnically diverse US adults, Bastos and Harnois (2020) concluded that cross-group comparisons were not advisable, as the authors were unable to prove measurement equivalence across racial/ethnic groups. The authors suggest that these differences may stem from geographic or historical contexts that are related to specific experiences of the samples assessed. As such, the authors advised that the PRaCY scale should be further refined among each new population with which it is used.

Additionally, it is imperative to consider the multiple marginalized experiences of some Black youth, such as girls, sexual and gender minorities and cultural minority youth (foreign born or immigrant populations)(Cogburn et al., 2011; Cooper et al., 2013). These youth may experience more or compounded discrimination due to their existence at the margins, having identifies that elicit negative societal treatment from systems such as sexism, homophobia and xenophobia (Mallory & Russell, 2021; Seaton et al., 2010). To adequately capture the nuances of these experiences, it may be useful to examine measurement invariances across these groups.

While the original PRaCY studies have numerous strengths and have been cited in almost fifty studies, exposure to race-based discrimination is not limited to outcomes of impaired psychological functioning and may directly impact other health and wellness outcomes in youth, including violence exposure. Being safe and free from violence is a pivotal factor that contributes to positive youth development and thriving, particularly among historically marginalized youth, that has not been well explored (Benson et al., 2011; Search Institute, 2018; Smith et al., 2017). Specifically, we chose to examine sexual violence given its disproportionate prevalence among communities of color (Breiding et al., 2015). Furthermore, race-based discrimination and sexual violence are forms of structural violence that are a result of larger systems of oppression (e.g., racism and sexism) that intersect to exacerbate health inequities, as posited by well-known frameworks, including syndemics, intersectionality, and social dominance theory (Crenshaw et al., 1995; Farmer, 2004; Sidanius & Pratto, 2004; Singer et al., 2017). Finally, Fleckman et al. overview literature that demonstrates the association between rape myth acceptance (a significant predictor of sexual violence perpetration) and racism (Fleckman et al., 2023).

Therefore, the purpose of this study was to (1) validate the factor structure of the PRaCY in a sample of historically marginalized youth residing in low-resource neighborhoods in Pittsburgh, Pennsylvania, (2) evaluate measurement invariance by gender, and (3) examine convergent validity through the relationship of race-based discrimination on sexual violence.

Hypothesis

We hypothesize that a unidimensional factor structure will apply for the responses to PRaCY from these cohorts based on previous research (Pachter et al., 2010; Pachter, Szalacha, Pachter et al., 2010a, b). Additionally, experiences of race-based discrimination will be common amongst adolescent participants from these studies, and experiences of discrimination will associate with violence outcomes.

Methods

Participants and Procedures

The data for these analyses were from baseline data of two trials evaluating gender-transformative, community-based, sexual violence prevention programs for youth (Manhood 2.0 and Sisterhood 2.0). Participants were ages 13–19 years, recruited between 2015 and 2017 (Manhood 2.0) and 2017–2019 (Sisterhood 2.0) through community organizations from the same neighborhoods (Eugene et al., 2017). Briefly, Manhood 2.0 was a cluster randomized trial designed to reduce adolescent relationship abuse perpetration and victimization among male youth living in neighborhoods with concentrated social disadvantage(Abebe et al., 2018; Miller et al., 2020). Sisterhood 2.0 was a quasi-experimental study in which female adolescent received an adapted Manhood 2.0 curriculum (ClinicalTrials.gov, NCT04388696).

Ethical Considerations

Informed Consent was obtained for all participants prior to enrollment in the study. This study was reviewed and approved by the University of Pittsburgh IRB. All measures included in this study were taken prior to the randomization of participants to intervention or control sites and does not account for intervention effects.

Measures

At baseline, participants in both Manhood 2.0 and Sisterhood 2.0 completed in-person anonymous surveys about their demographic, school, and social factors. The following measures were included in this analysis.

Perceptions of Racism in Children and Youth

The PRaCY assesses race-based discrimination by asking youth whether they have ever experienced any of ten situations (i.e., “yes/no”) because of their skin color, language or accent, or culture (Pachter et al., 2010). Mean scores were calculated, and higher scores indicated more experiences with race-based discrimination.

Sexual Violence

Manhood 2.0 and Sisterhood 2.0 participants were asked if they had experienced (i.e., “yes/no”) sexual violence victimization (2 items) and perpetration (2 items) in their lifetime (Koss & Gidycz, 1985; Tancredi et al., 2015).

Analyses

To validate the factor structure of the PRaCY scale among the Manhood 2.0 and Sisterhood 2.0 populations, confirmatory factor analyses were conducted. Given prior research on the PRaCY (Pachter et al., 2010; Pachter, Szalacha, Pachter et al., 2010a, b), we tested its unidimensional structure using weighted least square mean and variance adjusted estimation in Mplus (Version 8) accounting for ordinal factor indicators. Commonly accepted goodness-of-fit measures were considered (CFI > 0.95, TLI > 0.95, RSMEA < 0.06, SRMR < 0.08) to determine model fit (Anderson et al., 1987). Measurement invariance was tested between male (Manhood 2.0) and female (Sisterhood 2.0) samples by comparing fit statistics and chi-square difference tests for increasingly constrained models (configural vs. weak factorial vs. strong factorial vs. strict factorial). Configural invariance shows that the constructs are measured by the same items across comparison groups; weak factorial invariance shows that the underlying factors are defined by the same items across each group; strong factorial invariance sets the item thresholds as equivalent; and, strict factorial invariance constrains item residuals as equivalent. For every increasing level of invariance, we can be increasingly confident in our ability to compare mean scores across groups, minimizing concern for measurement error (Bialosiewicz et al., 2013). Internal consistency reliability was assessed through Cronbach’s alpha using the full sample (Brown, 2002).

Once demonstrating construct validity through confirmatory factor analyses, we assessed convergent validity by evaluating the association between the mean PRaCY score and sexual violence victimization and perpetration using mixed-effects logistic regression to account for neighborhood clusters of participants. All models were run separately by gender given differences in clustering across neighborhoods and as participants were from two separate quasi-experimental studies. As an exploratory analysis, we compared experiences of each item of race-based discrimination between male and female participants using chi-square tests.

Results

Demographic characteristics are presented in Table 1. The mean age of the Manhood 2.0 (n = 560) and Sisterhood 2.0 (n = 189) samples were 15.5 (SD: 0.14) and 15.1 years (standard deviation: 0.14), respectively. The majority of both samples identified as Black/African American at 76.1% and 76.7%, respectively. The frequencies of each item of the PRaCY scale are reported for the entire sample (N = 749), as well as the Manhood 2.0 (n = 560) and Sisterhood 2.0 (n = 189) samples separately. Of note, female youth reported significantly higher rates of having “the feeling that someone was afraid of you,” being “called an insulting name,” and having “someone be rude to you” because of their skin color, language or accent, or culture compared to adolescent boys. In contrast, male youth reported significantly higher rates of being “treated unfairly by a police officer” compared to adolescent girls.

Table 1 Participant characteristics and experiences of sexual violence
Full size table

The confirmatory factor analysis results are presented in Table 2, indicating that each item strongly loaded (factor loading > 0.5) onto a single factor (i.e., the construct of race-based discrimination).

Table 2 Discrimination scale – manhood 2.0 and sisterhood 2.0a
Full size table

In Table 3, we report goodness-of-fit statistics for Manhood 2.0 and Sisterhood 2.0 samples, as well as measurement invariance results by male and female gender. Separately, the model had excellent fit among male youth (CFI = 0.961; TFI = 0.950; RMSEA = 0.067, 90% CI 0.055–0.081; SRMR = 0.068) and female youth (CFI = 0.995; TFI = 0.993; RMSEA = 0.040, 90%CI 0.000-0.070; SRMR = 0.058) as a unidimensional construct. When comparing fit statistics and chi-square difference testing among increasingly constrained models, we observed weak factorial invariance across gender, demonstrating that the factors were defined by the same items between male and female youth. Cronbach’s alpha was 0.83 for the full sample, 0.81 for the boys, and 0.87 for the girls, indicating strong internal consistency.

Table 3 Measurement invariance by gendera
Full size table

We examined the association between PRaCY scores for male and female experiences with sexual violence (Table 4) as a way to evaluate convergent validity. Higher PRaCY scores were not associated with higher odds of lifetime non-partner sexual violence perpetration among boys (OR = 1.69, 95%CI 0.60–4.79) or girls (OR = 0.30, 95%CI 0.02–2.91), but were associated with higher odds of lifetime non-partner sexual violence victimization among girls (OR = 10.80, 95%CI 2.23–52.33) and lifetime sexual violence victimization (by any perpetrator) among boys (OR = 3.03, 95%CI 1.43–6.42).

Table 4 Associations of discrimination and violence victimization and perpetration
Full size table

Discussion

Our findings provide additional psychometric evidence demonstrating that the PRaCY accurately measures a unidimensional construct of race-based discrimination. Further, we establish the utility of the scale among a large sample of predominantly Black male and female youth from historically marginalized, urban neighborhoods. Based on these findings, the PRaCY should continue to be considered as a scale that may have wide generalizability across different adolescent populations. As outlined in Critical Race Theory (Crenshaw et al., 1995) and Public Health Critical Race methodologies (Ford & Airhihenbuwa, 2010), critically examining the dimensionality of race as a social construct and specific experiences of discrimination can guide the evaluation of interventions with valid and reliable measures and broaden research on health inequities.

Our analyses also highlight some potential variability in race-based discrimination by gender. Female youth were more likely to report feeling that someone was afraid of them, being insulted, or having someone be rude to them because of the color of their skin, language or accent, or culture compared to male youth. In addition, we evaluated measurement invariance by gender, proving only weak factorial invariance. Our findings show that researchers should use caution when comparing mean scores between participants of different genders, as observed differences may be due to measurement discrepancies as opposed to true differences. Furthermore, these results demonstrate the importance of intersectional approaches in intervention design that address experiences of gendered racism among female youth (Harnois et al., 2020). More interventions that are grounded in Critical Race Theory, Public Health Critical Race praxis, and Black Feminist Theory (Crenshaw, 1989) are needed to mitigate the impact of racism and sexism in further perpetuating harmful health outcomes for women and girls of color (Hogan et al., 2018).

Among male youth in this study, we observed a significant increase in sexual violence victimization with higher mean PRaCY scores. Among female youth, we noted a similar pattern: a significant increase in non-partner sexual violence victimization with higher mean PRaCY scores. With regards to victimization among boys and men, some studies have examined links between minority stress and experiences of violence (Stephenson & Finneran, 2016) and the correlation of experiences of police harassment and intimate partner violence among Black men who have sex with men (MSM) (Feelemyer et al., 2021). However, to our knowledge, these links have not been explored in non-sexual minority Black men.

Experiences with sexual violence victimization are profoundly gendered. Our findings among girls confirm what has already been demonstrated in violence literature; Black adolescent girls and women are disproportionately impacted by sexual violence as a result of multiple societal factors that intersect to perpetuate health disparities. This includes the intersections of pervasive systems of oppression such as sexism, racism, and classism (Harnois et al., 2020). Beyond higher rates of sexual violence, young Black women are also more likely to experience the adverse health consequences associated with said violence victimization, including sexually transmitted infections, substance use, and unintended pregnancy(Holliday et al., 2017). As we continue to tailor violence prevention for specific populations of youth, we must address race-based discrimination to mitigate the deep and persistent inequities experienced by women and girls of color.

The findings from this study must be considered in light of several general limitations. Given sample size restrictions, we were unable to conduct separate exploratory factor analyses and therefore did not establish dimensional invariance across gender. Furthermore, small cell sizes in regression analysis require results be interpreted with caution. We were unable to examine within group variability among different racial identities. Our samples were comprised of youth from neighborhoods of concentrated social disadvantage from one geographic area; the relationship between race and socioeconomic status was therefore difficult to disentangle. While the participants were from two different studies, they were administered the same PRaCY questions and data were collected from the same neighborhoods among youth of the same age group. Measurement invariance by gender was tested to determine if and how race-based discrimination varied by gender. Furthermore, the sample size would not necessarily allow for multiple comparisons across groups. Another limitation is the different question format used to assess sexual violence victimization. For the study with male-identifying adolescents, victimization from any perpetrator was inquired about. For the study with female-identifying adolescents, victimization from non-partner perpetrators was inquired about. Therefore, comparisons across these two populations should be made with caution. Finally, our analyses examining the associations between the PRaCY and sexual violence were based on cross-sectional data limiting any causal inference between race-based discrimination and our examined outcomes. Furthermore, we did not examine include other forms of violence, including physical and psychological violence, given limitations within our datasets regarding sample size and similar metrics. Results from this analysis may be only applicable to youth from the geographic region from which they were recruited.

The 10-item PRaCY scale shows construct validity for assessing race-based discrimination in a large sample of historically marginalized male and female youth in one geographic area. Yet, caution is needed when comparing PRaCY scores between participants of different genders. This scale should be incorporated in studies of overall adolescent health and wellbeing, especially after confirming validity of the scale in additional populations of youth. Future research should consider how race-based discrimination is experienced among youth with intersectional and multiple historically marginalized identities. Finally, race-based discrimination was associated with higher rates of sexual violence victimization across both genders and should be confirmed in larger samples. Studies should continue to investigate how addressing race-based discrimination may mitigate violence victimization.