Violence Exposure and PTSD: The Role of English Language Fluency in Latino Youth
- First Online:
- Cite this article as:
- Kataoka, S., Langley, A., Stein, B. et al. J Child Fam Stud (2009) 18: 334. doi:10.1007/s10826-008-9235-9
- 264 Views
Although Latinos have been a rapidly growing population in the US, little is known about how mental health symptoms may present in Latino children especially in the context of those living in poverty and exposed to violence. We explored the level of violence exposure and trauma symptoms in Latino youth and the relationship of these factors with English language fluency. During 2000–2002, 1,601, Latino students from seven middle schools participated in a school-based screening to identify students with exposure to community violence and symptoms of Posttraumatic Stress Disorder (PTSD). The students completed a self-report instrument, in either Spanish or English, that combined a modified version of the Life Events Scale and the Child PTSD Symptom Scale (CPSS). Bivariate analyses and multivariate regression models showed that youth with higher English language fluency reported greater violence exposure and PTSD symptoms than those with lower fluency. No difference was found in functioning by English language fluency. English language fluency appears to be related to violence exposure and PTSD symptoms in these Latino youth. We discuss the importance of school-based programs especially designed to serve Latino students of varying English language fluency.
KeywordsTraumaViolence exposureLatino youthLanguage fluency
The President’s New Freedom Commission clearly identifies schools’ central role in improving access to child mental health screening, assessment, and intervention (US Department of Health and Human Services 2003). Schools have long been identified as ideal settings for improving access to mental health services for children (Allensworth et al. 1997); three-fourths of children who receive mental health services do so through schools (Farmer et al. 2003). Ethnic minority children may especially benefit from improvements in school mental health assessments and treatments, given that they are less likely to receive traditional mental health services (US Department of Health and Human Services 1999; Dryfoos 1994; Garrison et al. 1999; Weist 1997). On-campus services can eliminate key financial and structural barriers that often prevent low-income and ethnic minority children from receiving needed services (Garrison et al. 1999) while addressing national concerns about ethnic disparities in access to mental health care (US Public Health Service 2000).
Although violence affects all racial, ethnic, and socioeconomic groups, there is a disproportionate amount of community violence among urban, poor, and minority youth (Bureau of Justice Statistics 1997; Christoffel 1990; Schubiner et al. 1993; Stein et al. 2003a). Low-income minority children are at increased risk for violence exposure and mental health problems due to multiple risk factors, including poverty (Coulton et al. 1995; Garbarino 1995; Straussner and Straussner 1997) and school and community factors (Kataoka et al. 2002; Oswald et al. 1999). Data from the 2003 Youth Risk Behavior Surveillance Survey (Centers for Disease Control, Prevention 2004) indicate that Latino males are more likely than Caucasian males to fight at school (17% vs. 10%) and were more likely to miss school because of safety concerns (9% vs. 3%).
Youth exposed to violence, either as witnesses or victims, are at risk for developing posttraumatic stress disorder (PTSD), as well as depression, other anxiety disorders, and substance abuse (Brent and Perper 1995; Clarke et al. 1995; Saigh et al. 1997; Weine et al. 1995). In addition, exposure to violence is associated with impaired school functioning (Garbarino et al. 1992; Hurt et al. 2001; Saigh et al. 1997; Schwab-Stone et al. 1995), decreased IQ and reading ability, (Delaney-Black et al. 2002) lower grade-point average (GPA), increased school absenteeism (Hurt et al. 2001), and decreased graduation rates (Grogger 1997). However, little has been documented about Latino youth exposed to violence and its impact on their level of functioning.
Although Latino immigrant children generally perform better in school than their non-immigrant Latino peers, these children also generally experience less sense of self-control and self-efficacy (traits associated with lower self-esteem and greater mental health problems) than their non-immigrant Latino peers (Kao 1999). However, more acculturated Latino youth may engage more frequently in high-risk behaviors, such as drug use (Adam et al. 2005; Epstein et al. 1998), which may increase their likelihood of exposure to violence.
Elsewhere, we reported high rates of violence exposure and PTSD symptoms among recently immigrated youth in a US school-based sample (Jaycox et al. 2002). In this cross-sectional study, we explore the relationship between violence exposure and English language fluency in Latino youth, compare levels of post-traumatic symptomatology across students with varying degrees of English language fluency, and examine the effects of English language fluency on functional impairment in Latino youth. We examine these issues during early adolescence, a developmental period associated with greater emotional adjustment and risk for poor school performance (Eccles et al. 1993).
Use of English language has previously functioned well as a proxy for acculturation for Latinos of Mexican-American background (Cuellar and Gonzalez 1996; Marin and Gamba 1996), and other studies of health behaviors have used self-reported language usage as a measure of acculturation (Adam et al. 2005; Epstein et al. 1998; Yu et al. 2003). Based on previous research, we hypothesized that those youth with greater fluency in English would report greater violence exposure (Adam et al. 2005; Epstein et al. 1998) and that students with lower fluency would have greater expression of PTSD symptoms in the avoidance PTSD symptom cluster, which has been found in Mexican samples (Norris et al. 2001), and less functional impairment, especially in the academic domain (Kao 1999).
Our study was part of a larger program evaluation by the Los Angeles Unified School District (LAUSD), in collaboration with local academic institutions, to develop and assess the effectiveness of a school-based intervention for children exposed to community violence (Jaycox et al. 2002; Stein et al. 2003b). The data presented here represents the baseline data prior to students receiving any intervention.
Participants and Procedures
During 2000–2002, a convenience sample of 1,601 Latino students from seven public middle schools, in sixth through eighth grade, participated in a screening conducted by school district personnel to identify students with violence exposure and PTSD symptoms. The school district selected schools that were located in socio-economically disadvantaged areas in Los Angeles and had administrators who agreed to have their school clinicians deliver the screening and subsequent treatment program on campus. All of the participating schools had high enrollments of Latino students (>80%).
Bilingual school staff described the screening process to all eligible students. Staff explained that the screening results would determine who could participate in a school program aimed at helping students cope with violence-related stress. Bilingual passive consent forms with instructions on how to decline participation for their child were obtained from parents. Although refusals or the reasons for refusal were not systematically tracked (since this was a naturalistic study and screening was conducted by school staff as part of their school program), they represented less than 10% of those students eligible for screening. After complete description of the study to the participants, written informed consent was obtained.
Of those students participating in this screening, 47% (n = 754) were female, with an average age of 11.7 years (SD = 1.06). Places of birth included the US (42%; n = 681), Mexico, (38%; n = 607), and other Latin American countries (20%; n = 313).
The study was conducted in compliance with the LAUSD’s Research Review Committee and the Institutional Review Boards of RAND and UCLA.
Trained school clinical social workers administered child self-report instruments to groups of 20–30 children who completed individual answer sheets. Measures were available in Spanish and English and administered by bilingual bicultural clinicians (i.e. clinicians who were fluent in Spanish and English languages and knowledgeable of both US and Latino cultures). All measures were translated by the LAUSD Translation Unit, and prior to the study, the measures were pre-tested using cognitive interviewing with similar students.
Students were asked for sociodemographic information, including age, gender, grade, and country of birth. English language fluency was assessed by asking, “How well do you speak English?” (an item modified from the US Census). Students responded on a 4-point Likert type scale (“not at all”, “a little”, “good”, and “excellent”). We chose to examine language fluency rather than comparing youth by country of birth because of its relevancy in the school setting and since the use of English may vary between this adolescent population and foreign- or native-born Latinos. To test the reliability of this measure, we compared English language fluency to language of survey (English or Spanish) and nativity, and we found that self-reported English language fluency was highly associated with both language of the survey and nativity (F-value = 1.34, p < 0.0001 and F-value = 1.84, p < 0.0001, respectively).
We measured violence exposure using a modified version of the Life Events Scale, a 34-item measure that asks about several types of violence (threats, slapping/hitting/punching, beatings, knife attacks, and shootings) in multiple locations over the past year (Singer et al. 1995, 1999). Respondents used a four-point Likert scale (ranging from “never” to “almost every day”). For this study, all items were coded yes or no to indicate any occurrence in the past year. A total violence exposure score was calculated by adding the number of reported types of violence exposure (both victimization and witnessed), with scores ranging from 0 to 24. All other violence variables were coded yes or no to indicate any occurrence in the past year. Exposure to media or other indirect violence was specifically excluded. The LES has been pre-tested in English and Spanish with LAUSD elementary and middle school students and found to have good face validity with this population. It also has been shown to have acceptable reliability (Chronbach’s α > 0.70) in elementary, middle, and high schools in multicultural inner city populations (Singer et al. 1995, 1999).
The Child Posttraumatic Symptom Scale (CPSS) was used to assess symptoms of PTSD in the past month. This measure has been used in children as young as eight and has shown good convergent (r = 0.80, p < 0.001), and discriminant validity (Chronbach’s α = 0.70), high internal consistency (Chronbach’s α = 70–0.89) and moderate test–retest reliability of PTSD symptoms (K = 0.55) (Foa et al. 2001). In a similar immigrant Latino population (Jaycox et al. 2002), scale internal consistency was high (Chronbach’s α = 0.89). This 17-item measure, with a 4-point Likert response scale (ranging from “not at all” to “almost always”), included items from each of the three PTSD symptom clusters: avoidance, hyperarousal, and re-experiencing. In addition, the CPSS includes six yes/no questions about functioning difficulties specifically related to the traumatic events reported in the previous section (e.g., “Are you having problems with classmates or other people?” “Are you unable to go to school?”). Internal consistency of these functioning items (Chronbach’s α = 0.89) and test–retest reliability (r = 0.70, p < 0.001) were high.
We calculated descriptive statistics about base rates of exposure to violence, PTSD symptoms, and PTSD-related functioning difficulties. We used regression analysis to test for association English language fluency and exposure to violence, PTSD, and functioning difficulties in bivariate analyses.
To examine the association of language fluency by total violence exposure, we used multiple linear regression, with age, gender, and school as covariates. In studying the relationship between language fluency and types of violence exposure, we used multiple logistic regression, adjusting for age, gender, and school. To examine the association of language fluency by PTSD total score and PTSD symptom clusters, we used multiple linear regression, with age, gender, total violence exposure, and school as covariates. Finally, we examined English language fluency and its relationship to PTSD functioning using multiple logistic regression, adjusting for age, gender, total violence exposure, and school.
All analyses were performed using SAS 9.1 (SAS Institute Inc 2004).
Rate of Violence Exposure
Violence exposure, PTSD symptoms, and functioning by language fluency (N = 1,601)
Total violence exposure, mean (SD)
Victimization, no. (%)
Witnessed violence, no. (%)
Weapon exposure, no. (%)
Violence at school, no. (%)
Violence in the community, no. (%)
PTSD total score, mean (SD)
Avoidance score, mean (SD)
Hyperarousal score, mean (SD)
Re-experiencing score, mean (SD)
PTSD functioning difficulties
Upset, no. (%)
Problem with peers, no. (%)
Unable to go to school, no. (%)
Grades worse, no. (%)
Problem with parents, no. (%)
Problem with teachers, no. (%)
PTSD Symptoms and Functioning: Descriptive Comparisons
Participants’ mean PTSD total score was 9.8 (Table 1), with students reporting hyperarousal symptoms most frequently (mean score of 3.5), followed by avoidance and re-experiencing symptoms (mean scores of 3.3 and 3.1, respectively). Higher PTSD scores within the avoidance and re-experiencing symptom clusters, as well as total score, were more frequently reported by those youth with greater English language fluency. For the hyperarousal symptoms, there was no association with English language fluency.
Many youth reported functioning difficulties as a result of PTSD symptoms, with 24% (n = 383) feeling more upset than they used to be and 22% (n = 344) experiencing more peer problems. Seventeen percent (n = 266) of students reported that their grades had worsened, with 8% (n = 126) of students having more problems with their teachers. English language fluency was related to both problems with going to school and grades. Specifically, those students with higher fluency reported greater difficulty in going to school, while those with lower language fluency had more impairment in grades. No differences by English language fluency were found for other areas of PTSD-related functioning difficulties (feeling upset, peer problems, and problems with parents and teachers).
Violence Exposure, PTSD Symptoms, and PTSD Functioning Difficulties: Adjusted Analyses
Multiple regression results: violence exposure by language fluency, age and gender (n = 1,601)a
Violence at school
Violence in the community
Multiple regression results: PTSD symptoms by language fluency, age, gender and violence exposure (n = 1,601)a
PTSD total score
Multiple regression results: PTSD functioning difficulties by language fluency, age and gender (n = 1,601)a
Problem with peers
Unable to go to school
Problems with parents
Problems with teachers
We are the first to describe the relationship between English language fluency in Latino youth and three areas related to trauma: violence exposure, PTSD symptoms, and functioning difficulties. In light of the President’s New Freedom Commission Report (US Department of Health and Human Services 2003), our study’s findings are particularly relevant to educators and school mental health providers. Over three-quarters of students reported exposure to violence at school (78%, n = 1,250) compared to half of the students (55%, n = 881) who reported exposure to violence in their community. In addition, we found significant problems in school functioning: 9% (n = 139) of students reported being unable to go to school and 17% (n = 266) had worsening grades as a result of PTSD symptoms.
Our first hypothesis that youth who are more fluent in English would report greater exposure to violence than students with lower English fluency was confirmed by our findings. The overall high levels of violence exposure among these Latino students are similar to those documented in other urban populations (Finkelhor et al. 2005; Schwab-Stone et al. 1995). These results are consistent with findings from previous studies that document greater health risk behaviors by more acculturated Latino youth (Balcazar et al. 1996; Harris 1999). Future research should explore to what extent factors such as familism and family cohesion play a role in protecting immigrant Latino children compared to their US-born counterparts (Rumbaut 1999) and how this difference may affect their exposure to violence.
Overall more fluent Latino students reported greater PTSD symptoms than less fluent students. This was not surprising, given the greater violence exposure of the more fluent students and the well-documented association between violence exposure and development of PTSD (Fitzpatrick and Boldizar 1993; Pynoos and Nader 1993). Even after controlling for total violence exposure, we continued to find greater symptoms in the more fluent students. These findings suggest that perhaps the expression of trauma symptoms may vary in Latino youth by level of acculturation. We also found that both avoidance and re-experiencing symptoms were more common in those students with greater English language fluency, which appears counter to what others have found in adults, where avoidance symptoms were more common in Mexico than in the United States (Norris et al. 2001). More research is needed to better understand how symptoms may be expressed in adolescents and what role acculturation plays in the expression of these symptoms.
Finally, we predicted a positive relationship between English language fluency and PTSD-related functioning difficulties that was not supported by our findings. We found that students with lower English language fluency appeared to have disproportionate impairment in the area of difficulties with grades, and higher fluency students had greater difficulties with going to school. However, after controlling for other factors such as age, gender, level of violence exposure, and school, no difference by English language fluency was found. Further research is needed to examine whether other characteristics of functional impairment may be more relevant for non-English speaking Latino students.
There are several limitations of our study worth noting. Given that this was a school-partnered project with concerns about participant burden, full acculturation measures were not administered. Instead, we chose to measure a self-report of English language fluency due to both its relevance to schools and that self-reported language has been shown in past studies to serve as a proxy for acculturation. Future research should broaden this examination of acculturation by using a multidimensional model of acculturation and examine such issues as the role of acculturative stress and parental acculturation on PTSD and functioning. Because only students’ place of birth was collected, we do not have ethnicity data for the US-born students and cannot determine its effect on violence, PTSD symptoms, and functional impairment. However, when we examined the data for Mexican immigrants compared to other non-Mexican immigrant students, we found no effect. Finally although our study has a large sample size, it is primarily comprised of poor, urban, Mexican-American and Central-American middle school students from one public school district, and the results of our study may not generalize to other Latino student populations.
Nevertheless, our study has particularly salient findings for mental health clinicians who are serving Latino youth. We document that many of the Latino youth whom we surveyed reported exposure to violence and symptoms of PTSD. Our findings also highlight the importance of culturally sensitive care for these youth that takes into account their language fluency and possible differences that Latino youth of varying language fluency may have in their level of violence exposure and types of symptoms that they may develop. Others have noted how essential it can be to integrate cultural issues into mental health treatment for youth, including the youth and family’s expressions of distress, communication styles, migration experiences, family values, and sociopolitical history (Bernal 2006; Hwang 2006; Pumariega et al. 2005; Szapocznik et al. 1986). Working with cultural liaisons, such as Promotoras (lay health promoters) who have cultural knowledge and clinical expertise, can be one way to improve mental health services to Latino communities (Ngo et al. 2008).
The fact that students in our study also reported such high levels of violence exposure occurring on school campus should alert school mental health providers and educators of the need for improved violence prevention efforts on campuses and greater early interventions for treating traumatized youth. School-based mental health services may be especially important for Latino youth, who have been found to have lower access to specialty mental health services than their non-Latino peers (Kataoka et al. 2002). Programs for traumatized Latino youth have been successfully delivered in schools (Kataoka et al. 2003; Stein et al. 2003b) and tailored to meet the needs of youth of varying acculturative levels. Further dissemination of such programs in community settings such as schools that can address the high level of mental health care needs of traumatized Latino students is warranted.