The author(s) shown below used Federal funds provided by the U.S. Department of Justice and prepared the following final report: Document Title: Experience of Intimate Partner Violence Among U.S. Born, Immigrant and Migrant Latinas Author(s): Andrea Hazen, Fernando I. Soriano Document No.: 211509 Date Received: September 2005 Award Number: 2000-WT-VX-0017 This report has not been published by the U.S. Department of Justice. To provide better customer service, NCJRS has made this Federally- funded grant final report available electronically in addition to traditional paper copies. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Running Head: Latinas and Intimate Partner Violence Experiences of Intimate Partner Violence Among U.S. Born, Immigrant and Migrant Latinas June 2005 By Andrea Hazen, Ph.D. Child and Adolescent Services Research Center Children’s Hospital San Diego, CA and Fernando I. Soriano, Ph.D. Department of Human Development California State University San Marcos San Marcos, California Child and Adolescent Services Research Center Children’s Hospital San Diego, CA Effort was funded by the National Institute of Justice (Grant No: 2000-WT-VX-0017). The views herein are those of the authors and do not purport to reflect the position of the National Institute of Justice or the U.S. Department of Justice. This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Table of Contents Abstract………………………………………………………………………………..…ii Executive Summary……………………………………………………………………...iv Project Goals and Objectives…………………………………………………….iv Research Design and Methodology……………………………………………...v Research Results and Conclusions……………………………………………….v Recommendations………………………………………………………………..x Introduction……………………………………………………………………………....1 Terms Used in the Report………………………………………………………..1 Methods…………………………………………………………………………………...3 Sampling………………………………………………………………………….3 Pilot Study………………………………………………………………………..6 Interviewing……………………………………………………………………....7 Human Subject Protocols………………………………………………………...8 Translation………………………………………………………………………..9 Data Entry and Management……………………………………………………..9 Overview of Study Measures…………………………………………………….10 Description of Measures…………………………………………………………11 Results…………………………………………………………………………………....18 Sample Characteristics………………………………………………………….. 18 Lifetime Rates of Intimate Partner Violence…………………………………….19 Past Year Rates of Intimate Partner Violence……………………………………22 Lifetime Rates of Injury and Receipt of Medical Care By Victims of Physical Assault……………………………………………………………………………26 Past Year Rates of Injury and Receipt of Medical Care By Victims of Physical Assault……………………………………………………………………………27 Background and Sociodemographic Factors Associated with Intimate Partner Violence………………………………………………………………………….28 Childhood Risk Factors Associated With Intimate Partner Violence……………30 Intimate Partner Violence, Mental Health, Physical Health, and Substance Use Problems…………………………………………………………………………31 Discussion………………………………………………………………………………..41 References………………………………………………………………………………..50 Appendix Section……………………………………………………………….………..60 i This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Abstract The three aims of this stratified random sample study were: (1) To assess the prevalence and patterns of intimate partner violence, including physical, sexual, and emotional/psychological abuse, in three stratified groups of Latina women ((a) Migrants: Migrant or seasonal workers (includes immigrants and non-immigrants), (b) Immigrants: Immigrants, but not migrant or seasonal workers, and (c) U.S. born Latinas); (2) to identify differential risk and protective factors associated with this violence, which included examining the role of cultural (e.g., acculturation, acculturative stress, ethnic identity, and bicultural self-efficacy), socioeconomic, psychosocial and social support factors; and based on these findings, (3) to outline specific implications of the findings for intimate partner violence prevention and intervention programming. The study used a quasi-experimental approach. The sample of 291predominantly Mexican American Latinas revealed a high rate of lifetime and past year experience with IPV. Thirty-three percent (33.9%) experienced some form of physical violence, 20.9% experienced sexual coercion, and 82.5% experienced psychological aggression by an intimate partner at some time in their life. Rates of victimization in the preceding year were high with 18.5% of the women reporting physical assault, 14.4% reporting sexual coercion, and 72.6% reporting psychological aggression. Having a partner with a substance use problem was associated with victimization in the present study. The study found an association between adult partner violence victimization and reports of intimate partner violence in women’s families of origin after controlling for other potential childhood risk factors and demographic variables. Experiencing childhood sexual abuse was strongly related to reports of adult intimate partner violence victimization. Rates of IPV differed ii This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence between groups with US born populations experiencing the highest level of violence followed by migrants. Higher levels of acculturation were also found to be associated with IPV, as well. Recommendations are made regarding the need for early screening and intervention among Latina populations, but also for children who are exposed to IPV. A call is made for continued research on Latina populations, and specifically on the stratified groups of focus in this study. More research is also needed on the role of culture in IPV among Latina populations, as well. iii This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Executive Summary Project Goals and Objectives Statement of Purpose. Violence against women is receiving increased national attention due to recognition of its alarming prevalence, and because of the need to expand services available to victims. While the past decade has seen significant growth in the amount of empirical research on this problem, what we know about its occurrence in the Latino population is extremely limited, even while it is one of the fastest growing ethnic minority groups in the U.S. (Morash 2000 p. 67). In order to address this gap in knowledge, we investigated the prevalence of various types of intimate partner violence among diverse Latina populations, as well as examined factors that increase and protect against victimization. More specifically, the three aims of the study were: (1) to assess the prevalence and patterns of intimate partner violence, including physical, sexual, and emotional/psychological abuse, in three groups of Latina women (migrant workers, immigrants who are not migrant workers, and U.S. born Latinas who are not migrant workers); (2) to identify differential risk and protective factors associated with this violence, which included examining the role of cultural (e.g., acculturation, acculturative stress, ethnic identity, and bicultural self-efficacy), socioeconomic, psychosocial and social support factors; and based on these findings, (3) to outline specific implications of the findings for intimate partner violence prevention and intervention programming. Research Subjects. A study involving 291 predominantly Mexican American Latinas examined differential patterns of intimate partner violence among three groups: (1) Migrants: Migrant or seasonal workers (includes immigrants and non-immigrants), (2) iv This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Immigrants: Immigrants, but not migrant or seasonal workers, and (3) U.S. born Latinas: Non-immigrants and not migrant or seasonal workers. Research Design and Methodology Methods. A quasi-experimental approach was used with random selection of the sample from a list of clients of a large primary health care organization serving over 43,000 predominantly low-income Latinos residing in one of the largest counties in California. The focus was on Latina women between the ages of 18 to 45. A survey interview gathered information on intimate partner violence experiences, as well as cultural, socioeconomic, psychosocial, family functioning, social problem, and social support network factors. Data Analysis. This study added to our understanding of the experience of intimate partner violence among Latina women. The research acknowledged the tremendous diversity of this population and examined the role of culture in light of other risk and protective factors. The findings provided important information to organizations interested in addressing violence issues among Latina women. Research Results and Conclusions Intimate partner violence victimization was prevalent among the Latina women in the present study. Lifetime Experiences Thirty-three percent (33.9%) experienced some form of physical violence, 20.9% experienced sexual coercion, and 82.5% experienced psychological aggression by an intimate partner at some time in their life. Experiences in the Preceding Year Rates of victimization in the preceding year were high with 18.5% of the women reporting physical assault, 14.4% reporting sexual v This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence coercion, and 72.6% reporting psychological aggression. The rates of physical assault were at least as high, and in some cases higher, than those generated by recent studies involving nationally representative samples. For instance, the rates of less severe (18.2%) and severe (7.5%) physical assault experienced in the preceding year were nearly equivalent to the rates reported by the National Comorbidity Survey (i.e., 17.4% for less severe, 6.5% for severe; Kessler et al., 2001). In contrast, the rate of any past year physical violence in the current sample was substantially higher than that obtained by the National Violence Against Women Survey (i.e., 18.5% vs. 1.3%; Tjaden & Thoennes, 2000). The rate of 25.4% for past year physical assault experienced by the U.S. born women in our sample was higher than the 16-17% prevalence rates found in other recent studies (Aldarondo et al., 2002; Lown & Vega, 2001). Meanwhile, the rate of physical assault among the immigrant women (i.e., 12.8%) was approximately the same (i.e., 13.4% reported by Aldarondo et al., 2002) or higher (i.e., 7.1% reported by Lown & Vega, 2001) than reported by these other studies. For the migrant women, our results on recent physical assault (i.e., 14.3%) are relatively comparable to the findings of Hightower and colleagues, who reported a 19.0% rate of physical or sexual assault among women employed as migrant or seasonal workers. Subgroup Comparisons Lifetime rates of intimate partner violence experienced by the U.S. born, immigrant, and migrant women in the current sample revealed differences. US born women more likely to have experienced physical assault and psychological aggression compared to the immigrant and migrant women. Differences were also noted vi This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence in the reports of sexual coercion with the US born women more likely to report these experiences than immigrant women. Reports of physical assault and psychological aggression in the preceding year were highest among the US born women and they were more likely to experience these types of victimization compared to immigrant women. US born women were also more likely to report sexual coercion by a partner than immigrant women. Women in the migrant group were also more likely to report recent sexual victimization compared to immigrant women. However, the study did not find a relationship between past year intimate partner violence victimization and women’s status as U.S born, migrant/seasonal, or immigrant Latinas when sociodemographic and other background characteristics of the women and their partners were taken into account. Socioeconomic Status and IPV Women’s age, marital status, poverty status of household, number of children in the household, and partner education were also not associated with partner violence. The lack of association between women’s age and risk of victimization is not consistent with previous findings that have identified younger age a risk factor in non-Latina and Latina women (e.g., Kessler et al., 2001; Lown & Vega, 2001; Suitor et al., 1990). The results do confirm previous findings with Latinos that have also not identified a relationship between intimate partner violence and marital status (Cunradi et al., 2002). Other research has similarly not found an association between income and partner violence among Latinos (Aldarondo et al., 2002; Kaufman Kantor et al., 1994; Lown & Vega, 2001) although this has not universally been the case (Cunradi et al., 2002). vii This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Substance Use Having a partner with a substance use problem was associated with victimization in the present study. Thirty percent of the women who experienced intimate partner violence in the preceding year reported partner substance use problems compared to 13.9% of women who did not experience any violence. Women who reported that their partner had alcohol and/or drug problems had more than two times the odds of reporting experiences with intimate partner violence in the preceding year. Past Family Experiences The study found an association between adult partner violence victimization and reports of intimate partner violence in women’s families of origin after controlling for other potential childhood risk factors and demographic variables. Women reporting violence between their parents or parental caregivers had approximately twice the odds of experiencing recent partner violence victimization compared to women who did not report parental violence. Past Childhood Abuse Experiencing childhood sexual abuse was strongly related to reports of adult intimate partner violence victimization. After controlling for other childhood risk factors and demographic variables, sexual abuse was associated with more than a three-fold increase in odds of reporting recent partner violence. Women who reported childhood physical abuse were approximately nine times more likely to report adult victimization. Cultural Factors A look at differences between each of the stratified groups of Latinas revealed significant differences between the three groups of Latinas with U.S. born and migrant Latinas reflecting the highest proportions of experience with any form of intimate partner violence within the past year (40% and 39%, respectively), compared to immigrant Latinas at 23 percent who showed the least proportion (X2 = 8.25, df = 2, viii This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence p<.02). Regardless of group or differences between groups, even at 23 percent, the proportion of Latinas experiencing any form of IPV is high—both numerically, and in comparison with proportions in the general population. Acculturation proved to be a significant cultural variable both as a unidimensional variable and as a multidimensional variable denoting groups or typologies varying in acculturation levels. The findings showed that the greater the acculturation of Latinas the more likely they were victims of IPV. There were few who could be typed as American oriented based on their acculturation scores to see if this acculturation type is tied to higher or lower experiences with IPV. However, the findings did show statistically significant differences in experiences with IPV between culture of origin oriented, bicultural and marginalized Latinas where those with higher IPV scores where marginalized and bicultural individuals (47% and 38%, respectively). This compared with 27% of culture of origin-oriented individuals. These findings suggest that acculturation is an important cultural variable that is tied to generalized IPV. The generalized IPV outcome variable is a macro measure and does not lend itself to a closer scrutiny of the relationship between the other cultural variables and AIPV. Further analysis is called for that uses an IPV measure with more variability, which can more carefully be examined in its relationship with the various cultural measures used in this study. Subsequent manuscripts will be prepared that will report on such analyses. In general, the present findings of this study suggest the need for screening for intimate partner violence in health care and other settings that serve Latina women. ix This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Recommendations • Our findings on childhood risk factors suggest that early identification and intervention efforts should target children who are exposed to intimate partner violence and who are victims of maltreatment, especially sexual abuse, in order to reduce risks for future victimization in intimate relationships. • Longitudinal studies that systematically investigate the interconnections between childhood and adult victimization are needed in order to identify optimal strategies for prevention and intervention programming. The findings underscore the importance of effective early intervention and prevention programs to mitigate short and long term risk of experiencing adverse outcomes. Such programming should be considered an essential component of the community response to intimate partner violence and childhood victimization. • In addition to assessing experiences with psychological, sexual, and physical violence, risk factors, including partner substance use, childhood history of maltreatment and intimate partner violence in the family of origin, should be considered. • In screening for intimate partner violence, service providers should also assess women’s mental health functioning to ensure that the often disabling sequelae associated with victimization are not overlooked. • The findings on cultural factors also suggest that cultural context is important, and in particular acculturation. The findings are in keeping with other health findings suggesting that those most at-risk are Latinos who are more acculturated (c.f., Gordon-Larsen, Harris, Ward, & Popkin, 2003). x This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence • More research is needed which uses a larger sample size to more adequately examine the importance of various cultural factors, as well as other risk factors of focus in this study. • Additional research is needed to investigate the applicability of these models to Latino populations and to test elaborated models that incorporate important cultural variables such as acculturative stress. • There is a clear need for further investigation of the pathways between childhood exposure to intimate partner violence and adult victimization and of intervening individual, family, and community factors. • Latino populations represent a large segment of the U.S. population with documented negative experiences with IPV. Their diversity in terms of culture and poor socioeconomic experiences status needs to continue to be acknowledged, incorporated and studied within future research on IPV. xi This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Introduction In recent years, there has been growing awareness about the pervasiveness and significant short and long term consequences of intimate partner violence. However, what we know about the occurrence of this social problem in the Latino population continues to be limited. In order to address this gap in knowledge, we investigated the prevalence of various types of intimate partner violence as well as the factors that may increase and protect against this victimization among Latina women who were migrant workers, immigrants, and U.S. born non-immigrants. The ultimate goal of this work was to obtain sufficient knowledge to inform the development of effective intimate partner violence prevention and intervention programs for Latina women. The specific aims: 1. To determine the lifetime and past year rates of various types of intimate partner violence (physical, sexual, emotional/psychological) in three groups of Latina women (migrant workers, immigrants who are not migrant workers, and U.S. born Latinas who are not migrant workers). Rates and patterns of intimate partner violence were compared for the three groups of Latina women. 2. To examine, among these same groups, cultural, socioeconomic, and psychosocial factors that may serve as risk and protective factors for intimate partner violence. 3. To draw implications from the study findings that will inform future research and program development focusing on migrant and non-migrant Latina populations. Terms Used in the Report For the sake of clarification, definitions of the terms used in this study are in order. We use the term "Latino" to refer to the population of Latin American descent, 1 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence which includes those who trace their ethnic ancestry to the former Spanish colonies of the Caribbean, Mexico, Central America, and South America. "Latina" refers to the female population who trace their ancestry to the aforementioned regions. "Immigrant" refers to those who were born outside the United States. "U.S. born or non-immigrant Latinas" refers to Latinas who were born in the United States and trace their ancestry to the aforementioned regions. "Migrants" or "seasonal workers" refer to Latinas, both U.S. born and immigrants, who are employed in the seasonal/temporary labor force and/or must move locations to find work. This category also included women who had a household member who contributed financially and who worked as a migrant or seasonal worker. The individual classified as a migrant or seasonal worker had to be employed in the agricultural industry (with qualifying jobs based on federal migrant education guidelines) in the preceding 24 months and a) that individual had moved in the preceding 24 months in order to perform his or her work or the individual was required to be absent overnight from his or her permanent place of residence (migrant worker) or b) the individual was employed in agricultural work of a seasonal or temporary nature (less than 12 months out of the year or had a specific end date or timeframe that work would be completed) (seasonal worker). 2 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Methods Sampling The sample is composed of three sub-groups of Latina women: SS1 = Latina migrant or seasonal workers, SS2 = Latina immigrants, and SS3 = U.S. born Latinas. Both the Latina immigrant (SS2) and U.S. born Latina (SS3) subsamples include Latinas who were not employed as migrant or seasonal workers. Inclusion criteria for the study were: (1) female, (2) 18-45 years of age; (3) of Latino ethnic background; (4) English or Spanish as primary language; and (5) had an intimate partner or contact with a former intimate partner within the 12 months preceding the survey. To study intimate partner violence in the Latino population, the research team secured access to the targeted population with the assistance of a large health care system made up of eight clinics in North San Diego County referred to in this report as HCS. Every patient or client of HCS provides information on ethnicity and occupation. For federal reporting purposes, HCS also categorizes clients as migrant or non-migrant. Using this information along with language and age, HCS provided us with a pool of 3,928 Latinas residing in North San Diego County who received health services between January 1, 2002 and March 31, 2002. It was unknown whether the women were immigrants or born in the United States and HCS’s definition of migrant differed from our own. To compensate we used “proxy” cells to stratify the sample. We placed all women designated as migrant by the health clinics into the proxy migrant group (n=681); Spanish-speaking, non-migrant women were put into the proxy immigrant group (n=2,615); and English-speaking, non-migrant women (n=631) were put into the proxy U.S. born group. From the pilot study we learned that the majority of the 3 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence “misclassifications” would come from the migrant and U.S. born proxy groups (i.e., proxy U.S. born and proxy migrants would actually meet our criteria for immigrant). A system was developed for reintegrating the misclassified names back into their correctly classified “proxy” group to improve the accuracy of future recruitment draws. Approximately each week, a proportion of names was selected for screening and recruitment from a randomly ordered list of the names in each of the three proxy groups. We drew a higher percentage of names from the U.S. born and migrant proxy groups because we expected to have the most misclassifications in these groups. Telephone contact was attempted with potential participants; if we were unable to reach individuals by phone, in person contact was attempted. During the recruitment call or in-person contact, a comprehensive screening was conducted to determine the group classification (i.e., migrant, immigrant, U.S. born) for the potential respondent and to ensure other study criteria were met. If the proxy classification matched the actual classification, the individual was asked to participate in the study and an interview was scheduled. If the status did not match the proxy classification, the individual was told that we would not be interviewing her at that time, but might be calling her in the near future for an interview. At the end of each week the names of the women who had been “misclassified” were placed in the correct “proxy” group. The list of names in each proxy group was then re- randomly ordered and the next proportion of names were selected for recruitment. If an individual appeared on the list who had already been screened and classified, the recruiters called the woman back and scheduled an interview. During the fieldwork, a total of 1,528 names were released. All 631 proxy U.S. born and 681 proxy migrants were released and recruitment attempted. Two hundred and 4 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence sixteen names were released from the proxy immigrant group. We confirmed the group status of 1,312 (85.9 %) women. Of those whose classification we were able to confirm 58.2% (n=763) were misclassified. As expected, the majority of misclassifications came from the U.S. born and migrant proxy groups. Only 7.0% (n=41) of the 584 confirmed classifications from the migrant group were true migrants, 88.2% (n=515) fit the immigrant category, and 4.8% (n=28) the U.S. born category. From the U.S. born proxy group, 87.8% (n=554) classifications were confirmed. Of these, 62.8% (n=348) were U.S. born; 35.2% (n=195) were immigrants; and 2.0% (n=11) were migrants. From the immigrant proxy group, 174 (80.6%) women were confirmed. The majority, 92.0% (n=160), were immigrants; 6.3% (n=11) were migrants; and 2.0% (n=3) were U.S. born. Of the 1,528 names released to recruiters, 627 (41.0%) cases were misclassifications that were subsequently not released. This left a potential of 901 women to interview. Overall the project completed 295 interviews: 118 immigrants, 50 migrants (5 U.S. born, 45 immigrants), and 127 U.S. born. Eighty-seven (9.7%) women could not be located; 172 (19.1%) women refused; 186 (22.9%) women did not meet study criteria, and 161 (19.8%) were eliminated as potential respondents because the U.S. born cell was full. If the 87 women who could not be located, the 161 women who were not eligible because the U.S. born cell was full, and the 186 who did not meet study criteria are excluded from the number of potential respondents, 295 (63.2%) of the remaining 467 cases were completed. Within each category the final status was as follows: Table 1. Final Sample Status by Category Final Immigrants Final Migrant Final US Born Final Unknown Totals Complete 118 50 127 N/A 295 5 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Did not meet criteria 74 6 40 66 186 Refusal 52 7 49 64 172 Unable to locate 1 0 0 86 87 Cell full 0 0 161 N/A 161 Totals 245 63 377 216 901 The 172 refusals can be categorized into four main reasons for refusal: not interested in participating (n=57), too busy to participate (n=56), partner influenced non- participation (n=25), and those with whom we had 3 or more contacts without successful completion of the interview by the end of the project (n=34). Pilot Study A pilot study was conducted to test our sampling methodology and translation of measures. Twenty-eight women were interviewed: 15 immigrants, 7 migrants, and 6 U.S. born. The pilot testing enabled us to make necessary changes to our sampling methodology, test the length of the interview, gauge the effectiveness of the measures, and refine the Spanish translations of measures and supporting materials. Pilot study sampling methodology. We randomly selected 180 names for recruitment from a list of 2,887 non-duplicated names of Latina women ages 18-45 that utilized services at HCS during November or December 2001. During the pilot, we used the migrant/non-migrant status the HCS assigned to approximate which category the women would belong to. We selected two-thirds (n=120) of the names from the non- migrant category assuming we had the best chance of identifying U.S. born and immigrant women from this category. The other third (n=60) were selected from those 6 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence labeled migrant. The list of names was randomly ordered and a proportion of names was selected weekly for recruitment. A total of 113 (63%) of the 180 names were drawn for recruitment for the pilot study. Seventy-two (63.7%) of these names came from the non-migrant category and 41 (36.3%) from the migrant category. Recruitment was attempted with all of the women classified as English-speaking by the HCS (n=28, three of which were from the migrant category) and with 85 women classified as Spanish-speaking. We were able to locate and confirm migrant classification of 81 (71.7%) women: 28 (68.3%) from the migrant category and 53 (73.6%) from the non-migrant category. There were 12 confirmed migrants, 9 confirmed U.S. born and 60 confirmed immigrants. Of the 81 classified, 28 (34.6%) interviews were completed: seven migrant interviews, 15 immigrant interviews and 6 US. born interviews. Nine (11.1%) women refused to participate; four (4.9%) did not meet study criteria and 40 (49.4%) women were not pursued because the immigrant cell was full. Of those we were unable to classify (n=32), four were out of the area, one was in an institution, and 27 were not pursued because of time limitations on the pilot study. Interviewing Surveys were administered in a face-to-face format. Respondents chose the language they were most comfortable conversing in. Interviews were conducted in a location that was most convenient for the participant and that assured protection of privacy and safety. Usually the interview occurred in the participant’s home and occasionally at one of the health clinics. Childcare was provided when necessary. 7 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence All interviewers were bicultural/bilingual and were trained by experienced research staff. Interviewers underwent an intensive initial training that included background information on intimate partner violence, methods of approaching respondents, procedures for ensuring protection of safety and confidentiality, standardized interviewing methods, crisis protocols and role-playing. Practice interviews were conducted with other interviewers and with research staff. Challenging interview situations were presented along with protocol for response and interviewers were assessed for their judgment and reaction. Before new interviewers became independent, they were required to conduct three successful “mock” interviews with experienced staff and their first three field interviews were “shadowed” by an experienced interviewer. Human Subject Protocols Procedures for the protection of human subjects were reviewed by the Institutional Review Board of California State University San Marcos. Prior to the initiation of any survey data collection, each research subject was informed in detail about the nature of the study, its sponsors, objectives and goals, probable duration of the study, and extent of respondent participation. Before beginning the interview, written informed consent was obtained with consent forms available in English and Spanish. Respondents were told that they could refuse to answer any questions, and could decline to participate or withdraw at any time without any adverse effect on their relationship with the HCS or other agencies. Interviewers and other personnel were trained to be sensitive to the conditions faced by victims of intimate partner violence, particularly with regard to closely guarding the information and protecting women’s safety. The research team was experienced in 8 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence taking appropriate precautions to minimize risk when recruiting and interviewing women who are victims of intimate partner violence. Respondents were assured that data would only be reported in aggregate form and would be stored in locked file cabinets with all identifiers removed. All computer files and databases were password protected, and only authorized research staff had access to project data. Translation All survey materials including informed consent forms were available in both Spanish and English. Interviews and supporting information were provided to respondents in the language of their choice. Fifty six percent chose to be interviewed in Spanish and 44% in English. Spanish translations of survey measures were taken from existing sources or were developed using established forward and back translation procedures when new translation was necessary. Data Entry and Management Data entry and management were overseen by the project data manager and included several components for ensuring the accountability, accuracy, and confidentiality of the data. The first was the tracking database, which served multiple functions. One was to maintain the three separate data tables based on migrant status and facilitate the integration of misclassified names, random ordering and release of new names for recruitment. It was also the main database used by the field coordinator to manage all day-to-day recruiting and interviewing operations. For data accountability, an initial check-off allowed the field coordinator to document that each interview measure and other paperwork were completed and submitted by the interviewer. A paper checklist was also initialed by the field coordinator and forwarded with the interview packet to the 9 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence data manager. Upon receiving the packet, the data manager verified all pieces of the interview were accounted for and that no responses were missing or unclear. The data manager then input basic data (respondent ID, migrant status, language, date of interview, interviewer ID#, and any comments pertinent to the case) into a Master SPSS file. On a weekly basis, the Master file was crosschecked with the Tracking database to verify that the data manager had received data for all completed interviews. Weekly, each completed interview packet was separated into groups by measure for data entry into individual SPSS files. The majority of measures were double data entered for quality control. Fifteen percent from each of the remaining measures were randomly selected for double data entry. The individual SPSS measures files were crosschecked regularly with the Master file to verify all interview sections were data entered for each respondent. Data were backed up nightly on a secure server and also downloaded on a zip disk weekly. To ensure confidentiality, access to all computer files were password-protected and all paper forms were placed in locked filing cabinets. Files that included identifying information were kept separate from data files, which were identifiable only by number. All back-up disks were stored in a locked filing cabinet. Overview of Study Measures The survey examined Latina women’s experiences with various types of intimate partner violence and examined the relationship of intimate partner violence to cultural, socioeconomic, psychological, social problem, and childhood risk factors. The specific measures used in the study are listed in Table 2 below. Table 2. Study Measures 10 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Domain Measure Intimate Partner Violence Revised Conflict Tactics Scales (CTS2; Straus, Hamby, Boney- McCoy, & Sugarman, 1995) Cultural Factors Ethnic Identity Multi-Group Ethnic Identity Measure (MEIM; Phinney, 1992) Acculturation Short Acculturation Scale (SAS; Marin et al., 1987) Pan-Acculturation Scale (PAN; Soriano, 1999) Acculturative Stress Acculturative Stress Scale (AS: Mena et al., 1987) Bicultural Self- Efficacy Bicultural Self-Efficacy Scale (BISES; Soriano & Bandura, 1994) Childhood Risk Factors Parental Intimate Partner Violence Conflict Tactics Scales – Adult Recall Version (CTS2-CA; Straus, Hamby, Finkelhor, Boney-McCoy, & Sugarman, 1995) Child Maltreatment Parent-Child Conflict Tactics Scales – Adult Recall Version (CTSPC-CA; Straus, Hamby, Finkelhor, Moore, & Runyan, 1995) Parental Substance Use Shuckit Family History Interview (Shuckit, 1984) Mental Health Functioning Brief Symptom Inventory (BSI; Derogatis, 1993) Short-Form Health Survey – Mental Health Scale (SF-12; Ware et al., 1996) Beck Scale for Suicide Ideation (BSS; Beck & Steer, 1991) Physical Health Functioning Short-Form Health Survey – Physical Health Scale (SF-12; Ware et al., 1996) Substance Use Alcohol use Respondent Partner Alcohol Use Disorders Identification Test (AUDIT; Saunders et al., 1993) Drug use Respondent Partner Drug Abuse Screening Test (DAST; Skinner, 1994) Demographic Characteristics Demographic questionnaire - Age, race/ethnicity, language, socioeconomic status variables (education, occupation, income), household characteristics 11 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Description of Measures Intimate partner violence In keeping with recent recommendations by the Centers for Disease Control and Prevention (1999), we assessed the following categories of intimate partner violence: physical violence, sexual violence, psychological/emotional abuse, and injury related to violence. Women’s experiences with these types of intimate partner violence were assessed with the Revised Conflict Tactics Scales (CTS2; Straus et al., 1995). The CTS2 contains the following scales: psychological aggression, physical assault, sexual coercion, and injury. Each scale has minor and severe subscales, reflecting the severity of the abusive or violent acts. For example, on the physical assault scale, the less severe subscale includes items inquiring about being pushed, shoved, grabbed, and slapped, whereas the severe subscale includes questions about being choked, punched, burned, beaten up, and threatened with a knife or gun. Response categories range from 0 (never) to 6 (more than 20 times), indicating the frequency of occurrence of the abusive or violent acts in the preceding 12 months. For events that did not occur in the previous 12 months, the respondent was asked to indicate if they ever happened in an intimate relationship. For each scale and subscale, prevalence scores reflecting lifetime and past year experiences can be derived. The prevalence variables are dichotomous, with a score of 1 assigned if one or more of the acts in the scale (or subscale) occurred and a score of 0 assigned if none of the acts occurred. Good internal consistency has been reported for the 12 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence CTS2 scales, with reliability coefficients ranging from .79 to .95 (Straus, Hamby, Boney- McCoy, & Sugarman, 1996). Cultural factors Ethnic identity. Ethnic identity was measured using the Multigroup Ethnic Identity Measure (MEIM; Phinney, 1992), a widely used 14-item scale that covers self- identification (ethnic label) and ethnicity, ethnic behaviors and practices, affirmation and belonging, and ethnic identity achievement. In our study, the alpha coefficient for the MEIM was .73. Acculturation. Acculturation was measured with the Short Acculturation Scale (SAS; Marin et al., 1987) and the Pan-Acculturation Scale (PAN; Soriano, 1999). The Short Acculturation Scale is a 12 item self report acculturation measure developed for Latino populations. It includes items inquiring about language use, preferences regarding media, and social relationships. The scale has been found to correlate highly with respondents’ generation, length of residence in the United States, age at arrival, and ethnic self-identification. Acceptable reliability and validity have been reported (Marin et al., 1987). In the current study, alpha reliability was .95. The PAN is comprised of six subscales measuring language use, identity, social support, cultural practices, generational status, and cultural values and beliefs. Responses to the 22 items of the scale are used to measure affinity to one’s culture of origin and to American culture. It can also be used to classify respondents as bicultural, American dominant, culture of origin dominant, and marginalized (affinity to neither). In the present sample, alpha reliability was .91. The scale has been validated using a Mexican 13 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence American subsample, which reflected a correlation of .65 with Mendoza's Cultural Lifestyles Inventory. Acculturative stress. Acculturative stress has been discussed extensively in the area of Latino health (Vega & Rumbaut, 1991; Zimmerman, Vega, & Gil, 1994). In the present study, Mena et al.’s (1987) 24-item acculturative stress scale was used (alpha = .87 in the current sample). This instrument measures strains due to contact with a host society in the areas of language problems, perceived discrimination, and perceived cultural incompatibilities. Bicultural self-efficacy. Bicultural self-efficacy was measured using the Bicultural Self-Efficacy Scale (BISES; Soriano & Bandura, 1994). The BISES is a 13 item scale that measures the respondent’s efficacy in responding to personal and social challenges related to their affinity or involvement with members of the dominant (host) society. Essentially, it measures efficacious responses to challenges posed by members of their culture of origin and dominant culture within the life domains of school or work, peers, family, community, and society. The alpha reliability coefficient was .82 in this study. Childhood risk factors Parent/caregiver intimate partner violence. Information on study participants’ childhood exposure to violence between parents or caregivers was obtained with the Conflict Tactics Scales – Adult Recall Version (CTS2-CA; Straus, Hamby, Finkelhor, Boney-McCoy, & Sugarman, 1995). For the analyses described in this report, the physical assault and injury scales of the CTS2-CA were used. Thus, parent/caregiver intimate partner violence was considered present if the respondent reported that sometime 14 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence during childhood either parent (or other significant caregiver such as a stepparent) had engaged in one or more acts of physical assault against the other parent (e.g, pushing, shoving, grabbing, beating up, burning, using a weapon), had experienced a physical injury, or used medical services related to a violent incident. Child maltreatment. The Parent-Child Conflict Tactics Scales – Adult Recall Version (CTSPC-CA; Straus, Hamby, Finkelhor, Moore, & Runyan, 1995) was used to assess childhood history of maltreatment. The CTSPC-CA includes scales that assess experiences with physical assault, sexual abuse, and neglect. Separate items inquire about the commission of each of the abusive acts by the respondent’s mother (or maternal caregiver) and by the father (or paternal caregiver). (Note: The sexual abuse questions are an exception in that no reference is made to the identity of the person responsible for the abuse.) For the analyses described herein, separate dichotomous scores were derived for presence of physical abuse, sexual abuse, and neglect with a score of 0 indicating that none of the abusive acts were committed during the respondent’s childhood by either her mother (or maternal caregiver) or father (or paternal caregiver) and a score of 1 indicating that one or more acts were committed by either the mother (or maternal caregiver) or father (or paternal caregiver). The physical assault scale of the CTSPC-CA contains subscales organized according to the severity of the abusive acts, specifically minor, severe, and very severe. For this study, childhood physical abuse was considered present if the respondent reported she had experienced one or more of the abusive acts contained in the severe and very severe assault subscales. These items include being thrown to the ground, hit with a fist, beat up, burned or scalded, and threatened with a knife or gun. The sexual abuse 15 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence scale inquires about unwanted sexual touch and forced sexual contact with an adult (i.e., family member or non-family member) or older child experienced prior to the age of 18. Childhood sexual abuse was considered present if the respondent reported that any abusive acts occurred on one or more occasions during childhood. The neglect scale consists of items measuring parent or caregiver failure to meet the respondent’s basic needs during childhood such as not providing adequate food or supervision. Childhood neglect was considered present if the respondent indicated she had experienced any of the neglectful acts assessed by the scale. Parental/caregiver substance use. The Shuckit Family History Interview (Shuckit, 1984) was used to obtain information on respondents’ family history of alcohol and drug problems. In this study, a substance use problem was defined as the occurrence of one or more of the following significant alcohol or drug related problems: marital separation or divorce; job loss or layoff; two or more arrests; harm to physical health; inability to carry out daily responsibilities; receipt of treatment for substance use problems; and intravenous drug use. Parental/caregiver substance use problems were considered present if the respondent reported that either a biological parent or other significant caregiver (e.g., step parent, adoptive parent) experienced one or more of these problems during the respondent’s childhood. Mental health functioning Participants’ mental health functioning was assessed with the Global Severity Index of the Brief Symptom Inventory (Derogatis, 1993) and the mental health scale of the Short-Form Health Survey (SF-12; Ware, Phillips, Yody, & Adamczyk, 1996). The Global Severity Index reflects overall psychological distress with higher scores indicative 16 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence of greater distress (alpha = .96 in the current sample). The SF-12 mental health scale assesses general mental health functioning and limitations due to mental health problems, with higher standard scores reflective of better functioning. Prior research has supported the reliability and validity of both the Brief Symptom Inventory (Derogatis, 1993) and SF-12 (Ware et al., 1996). An item from the Beck Scale for Suicide Ideation (BSS; Beck & Steer, 1991) was used to assess history of suicide attempts. Respondents reported whether they had ever attempted suicide at some time in their lives. Physical health functioning Physical functioning was assessed with the physical health scale of the Short- Form Health Survey (SF-12; Ware, Phillips, Yody, & Adamczyk, 1996). This scale includes questions about physical functioning, general health, experiences with bodily pain, and limitations due to physical health problems. Acceptable test-retest reliability and validity have been reported for the SF-12 physical health scale (Ware et al., 1996). Substance use – Respondent and partner Respondent alcohol use problems were assessed with the Alcohol Use Disorders Identification Test (AUDIT; Saunders et al., 1993). The AUDIT is a 10 item screening questionnaire which obtains information on the quantity and frequency of alcohol use, major life problems caused by alcohol use (e.g., occurrence of injuries), and alcohol dependence symptoms (e.g, need for alcohol after awakening to avoid or relieve withdrawal symptoms) experienced in the preceding 12 months. The total AUDIT score ranges from 0 to 40 (alpha = .88 in the current sample); a score of 8 or greater has been derived as a cutpoint indicating problems with alcohol use. 17 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Respondent drug use was assessed with the Drug Abuse Screening Test (DAST; Skinner, 1994). The DAST consists of 10 questions that assess the use of drugs, physical and medical complications, and emotional and personal problems resulting from drug use in the preceding 12 months. The total score, which has a range of 0 to 10, is obtained by summing across all item scores (alpha = .71 in the current sample). Acceptable psychometric properties have been reported for both the AUDIT (Saunders et al., 1993) and DAST (Gavin et al., 1989; Skinner, 1994). Modified versions of the AUDIT and DAST were used to obtain participant reports on their current or most recent partners’ alcohol and drug use in the preceding 12 months. The questions in both measures were reworded in order to ask women about their partners’ use of substances and the occurrence of substance related problems. The partner versions of the AUDIT and DAST followed the same scoring rules as the original measures. For the purposes of the analyses contained in this report, partner substance use problems were considered present if a total score of 8 or greater was obtained on the AUDIT and/or if a total score of 6 or greater was obtained on the DAST. In the present study, alpha reliability was .97 for the partner version of the AUDIT and .87 for the partner version of the DAST. Results Sample Characteristics Sample characteristics are presented in Table 3. Demographic information is provided for the overall sample and for the three groups of participants (i.e., US born, immigrant, migrant/seasonal). All 295 study participants were of Latino or Hispanic origin, with the majority (92.2%) Mexican or Mexican American. Fifty three percent of 18 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence the women were born in Mexico, 44.7% were born in the United States, and 2.0% were born in other countries. The mean age of the participants was 27.8 years (sd = 7.13). Slightly more than half (i.e., 54.9%) had less than a high school diploma or equivalent, and only 30.8% were employed full-time. Thirty two percent were living at or below the federal poverty threshold. Demographic characteristics of the intimate partners of the study participants are presented in Table 4. This information is provided for the participants’ current partners or for their most recent partners if they were not in an intimate relationship at the time of the interview. For 51.9% of the women, the partner they reported on was their spouse, for 29.2%, it was a cohabitant, for 12.5%, it was someone they were seriously dating, and for 6.4%, it was someone they were casually dating. The mean age of the partners was 30.15 years (sd = 7.45), and the majority (i.e., 84.6%) were Latino. While seventy nine percent were employed full-time, more than half (i.e., 60.3%) had less than a high school diploma or equivalent. Lifetime Rates of Intimate Partner Violence Lifetime rates of psychological, physical, and sexual intimate partner violence are presented in Table 5, and rates of victimization by specific acts of abuse and violence are provided in Tables 7-9. Lifetime intimate partner violence was considered present if violence was reported in the past year or any prior period. Rates are reported for the overall sample and for the three groups of study participants (US born, immigrant, migrant/seasonal). To examine differences between groups, 3 (US born, immigrant, migrant) X 2 (presence vs. absence of type of intimate partner violence) chi-square analyses were conducted. Where there were significant differences (p < 0.05), follow-up 19 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence pairwise chi-square tests were carried out with a Bonferroni correction applied to each set of pairwise analyses (i.e., the acceptable alpha level was .05/3 = .017). 20 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Lifetime physical assault – overall sample. As shown in the first column of Table 5, 33.9% of the women in the sample reported that they experienced physical violence perpetrated by an intimate partner at some time in their life, with 32.9% reporting at least one incident of less severe assault (such as being pushed, grabbed, or slapped) and 17.8% reporting at least one incident of severe physical assault (such as being punched, kicked, or beat up). Lifetime rates of specific types of less severe physical assault ranged from 12.0% of the sample being slapped to 21.6% being grabbed, pushed, or shoved. Rates of experiencing severe acts of physical violence ranged from 0.7% being burned or scalded to 12.7% being slammed against a wall. Lifetime physical assault – subgroups. In the U.S. born group, 48.4% of the women reported victimization by physical violence, including 47.6% of the women reporting experiences with less severe physical assault and 20.6% reporting experiences with severe physical assault. Among the immigrant women, 22.2% reported physical assault in their lifetime, with 22.2% reporting less severe physical assault and 15.4% reporting severe physical assault. Some 24.5% of the women in the migrant/seasonal worker group experienced physical violence, including 20.4% reporting less severe assault and 16.3% reporting severe physical assault. There were significant group differences for any physical assault and for less severe physical assault but not for severe physical assault. Pairwise chi-square tests indicated that the US born group experienced a significantly higher rate of less severe physical assault and any physical assault compared to the immigrant and migrant groups. Lifetime sexual coercion – overall sample. The study found that 20.9% of the sample experienced sexual coercion by an intimate partner in their lifetime, with 19.9% 21 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence reporting a least one incident of less severe sexual coercion (such as a partner insisting on sex without a condom) and 6.5% reporting at least one incident of severe sexual coercion (such as a partner using physical force or a weapon to coerce sexual intercourse). Rates of experiencing specific acts of less severe sexual coercion ranged from 6.8% of the sample reporting that a partner had insisted on oral or anal sex to 16.4% indicating that a partner had insisted on sex without using physical force. Rates of experiencing more severe acts of sexual coercion ranged from 2.7% of women reporting that a partner had used threats to coerce oral or anal sex to 5.5% indicating that a partner had used physical force or a weapon to coerce sexual intercourse. Lifetime sexual coercion – subgroups. Lifetime rates of any sexual coercion were highest in the U.S. born group. Some 29.4% of the women in this group reported sexual coercion by an intimate partner whereas 20.4% of the migrant women and 12.0% of the immigrant women reported similar experiences. There was a significant difference in the rates of any sexual coercion, with U.S. born women more likely to report these experiences than immigrant women. Rates of experiencing less severe acts of sexual coercion were also highest in the U.S. born group; 27.8% of the women in the U.S. born group, 18.4% of the women in the migrant group, and 12.0% of the women in the immigrant group reported experiences with less severe sexual coercion. The women in the U.S. born group were significantly more likely to report these experience than women in the immigrant group. Rates of more severe sexual coercion were 8.2% in the migrant group, 6.3% in the U.S. born group, and 6.0% in the immigrant group. These rates were not significantly different. 22 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Lifetime psychological aggression – overall sample. The study results indicated that the majority of women in the sample experienced psychological aggression in an intimate relationship at some time in their life. In the overall sample, 82.5% reported experiencing some form of psychological aggression, with 82.2% experiencing less severe and 33.9% experiencing severe psychological aggression by an intimate partner. The most frequently reported acts of less severe psychological aggression were being shouted or yelled at, followed by being insulted or sworn at. With regard to severe psychological aggression, being called “fat” or “ugly” and having a partner destroy one’s possessions were reported most frequently. Lifetime psychological aggression – subgroups. The lifetime rates of any psychological aggression experienced in an intimate relationship were highest in the U.S. born group (92.9%), followed by the migrant (79.6%) and immigrant (72.6%) groups. The women in the U.S. born group were significantly more likely to report any psychological aggression than the women in immigrant and migrant groups. Similarly, the women in the U.S. born group had a higher rate of experiencing less severe psychological aggression and were significantly more likely to report these experiences compared to immigrant women. The women in the U.S. born group were also significantly more likely to report experiences with severe forms of psychological aggression than women in the immigrant group. However, they did not differ from the migrant women in their reports of severe psychological aggression, nor did the immigrant and migrant groups differ. Past Year Rates of Intimate Partner Violence 23 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Rates of psychological, physical, and sexual intimate partner violence experienced in the year preceding the survey are presented in Table 6, and rates of victimization by specific acts of violence are provided in Tables 7-9. Rates are reported for the overall sample and for the three groups of study participants (US born, immigrant, migrant/seasonal). To examine differences between groups, 3 (US born, immigrant, migrant) X 2 (presence vs. absence of type of intimate partner violence) chi-square analyses were conducted. Where there were significant differences (p < 0.05), follow-up pairwise chi-square tests were carried out with a Bonferroni correction applied to each set of pairwise analyses (i.e., the acceptable alpha level was .05/3 = .017). Past year physical assault – overall sample. Some 18.5% of the sample reported that they were victims of physical assault perpetrated by an intimate partner in the year prior to the interview, with 18.2% reporting at least one incident of less severe assault (e.g., being pushed, grabbed, or slapped) and 7.5% reporting at least one incident of severe physical assault (e.g., being punched, kicked, or beat up). Rates of victimization by less severe types of physical assault ranged from 3.4% of the women in the sample being slapped to 10.3% being grabbed. Reports of severe acts of physical violence ranged from 0.3% being burned or scaled or experiencing a knife or gun being used in a violent incident, to 4.8% being slammed against a wall. Past year physical assault – subgroups. Past year rates of any physical assault by an intimate partner were highest among the U.S born women. The women in this group had a significantly higher rate of any past year physical assault (25.4%) compared to women in the immigrant group (12.8%). However, the U.S. born group did not differ from migrant group, nor was there a significant difference between the immigrant and 24 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence migrant groups. Comparisons among the three groups did not reveal significant differences in reports of less severe physical assault (e.g., being pushed, shoved, slapped). Similarly, the groups did not differ in victimization by acts of severe physical violence (e.g., being punched, choked, beat up). 25 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Past year sexual coercion – overall sample. As shown in Table 6, 14.4% of the sample reported experiences with sexual coercion by an intimate partner in the preceding year. The rate of less severe sexual coercion was 13.4% (e.g., partner insisting on sex without a condom) and the rate of severe sexual coercion was 1.7% (e.g., partner using physical force or a weapon to coerce sexual intercourse). Rates of experiencing specific types of less severe sexual coercion ranged from 3.4% reporting that a partner insisted on oral or anal sex without using physical force to 10.3% reporting that a partner insisted on sexual intercourse without using physical force. Reports of experiencing more severe acts of sexual coercion ranged from 0.0% indicating that a partner had used threats to force oral or anal sex to 1.4% reporting that a partner had used physical force or a weapon to coerce sexual intercourse. Past year sexual coercion – subgroups. The study found that 21.4% of the women in the U.S. born group, 18.4% of the women in the migrant group, and 5.1% of the women in the immigrant group experienced some type of sexual coercion in the preceding year. The rate of sexual coercion was significantly higher in the U.S. born group compared to the immigrant group as well as in the migrant group compared to the immigrant group. The U.S. born and migrant groups did not differ. Rates of experiencing less severe types of sexual coercion were highest in the U.S. born group (20.6%), followed by the migrant group (14.3%) and immigrant group (5.1%). The rate of less severe sexual coercion was significantly higher among the U.S. born women compared to the immigrant women, but was not significantly higher relative to the women in the migrant group. Further, there was no significant difference between the immigrant and migrant groups. 26 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence The past year rate of more severe sexual coercion was 6.1% in the migrant group and 1.6% in the U.S born group. None of the immigrant women indicated that they had experienced any of the severe forms of sexual coercion. Comparisons among the groups revealed a significant difference between the migrant and immigrant women in their reports of severe sexual coercion; however, this finding should be interpreted with caution given the small sample size. Past year psychological aggression – overall sample. In the overall sample, 72.6% reported some form of psychological aggression by an intimate partner in the preceding year, with 71.6% experiencing at least one incident of less severe psychological aggression and 21.6% experiencing at least one incident involving more severe psychological abuse. Rates of exposure to less severe psychological aggression ranged from 16.8% indicating that a partner said something to spite them to 53.8% reporting that a partner shouted or yelled at them. Rates of experiencing more severe psychological aggression ranged from 3.1% reporting that a partner accused them of “being a lousy lover” to 14.7% reporting that a partner had called them “fat” or “ugly”. Past year psychological aggression – subgroups. In the U.S. born group, 84.1% of the women reported psychological aggression by an intimate partner in the preceding year, with 81.7% reporting experiences with less severe psychological aggression and 24.6% indicating they had experienced severe psychological aggression. Among the immigrant women, 60.7% reported psychological aggression, with 60.7% reporting less severe and 15.4% reporting experiences with severe psychological aggression. In the migrant group, 71.4% reported exposure to psychological aggression including 71.4% indicating they had experienced less severe psychological aggression and 28.6% 27 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence reporting experiences with severe psychological aggression. There were significant group differences for any psychological aggression and for less severe but not for severe psychological aggression. Pairwise chi-square tests indicated that the U.S. born group experienced a significantly higher rate of less severe psychological aggression and any psychological aggression compared to the immigrant group. There were no significant differences between the U.S born and migrant groups, and between the immigrant and migrant groups in the rates of any psychological aggression and less severe psychological aggression. Lifetime Rates of Injury and Receipt of Medical Care By Victims of Physical Assault Lifetime injury - overall sample. The lifetime rates of physical injury and receipt of medical care reported by women victimized by physical assault are presented in Table 10. As shown in the first column, 50.5% of the women in the overall sample who experienced physical assault at some time in their life also experienced a physical injury, reported need for medical care, or reported receipt of medical care. Most frequently reported were relatively minor injuries such as sprains, bruises, and cuts (41.4%). Reports of more severe injuries, such as broken bones (5.1%) and passing out from being hit on the head (6.1%), were less common. Approximately one in five women reported that they needed medical care at some point but did not receive it, and only 14.1% reported the use of medical care on at least one occasion because of a violent incident with their partner. Lifetime injury – subgroups. Some 39.3% of women in the U.S. born group who were victims of physical assault in their lifetime reported injury, or indicated the need for or use of medical care. In this group, 26.2% sustained sprains, bruises, or cuts inflicted 28 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence by an intimate partner and 32.8% had complaints of pain following a violent incident with their partner. As with the overall sample, broken bones (3.3%) and passing out from a blow to the head (6.6%) were less frequently reported. With regard to use of medical services, 16.4% perceived that they needed medical care on at least one occasion but had not received it, and 14.8% indicated they had received medical care at least once. Among the victims of physical assault in the immigrant group, it is noteworthy that the majority (73.1%) reported some type of injury or indicated the need for or use of medical care because of partner inflicted physical injury. Almost 70% reported that they sustained sprains, bruises, or minor cuts and 42.3% had complaints of pain. Relatively few reported more serious injuries such as broken bones (7.7%). Slightly less than one- quarter felt they required medical care following a violent incident but did not receive assistance, and 15.4% reported receipt of medical care on at least one occasion. Of the victims of physical assault in the migrant group, 58.3% indicated that they had experienced an injury, needed medical care, or used medical care. Slightly more than 50% of these women reported minor injuries, such as sprains, bruises, and cuts, but reports of more serious injuries were infrequent. While almost half of the women perceived the need for medical assistance in relation to a violent incident with their partner, the receipt of medical care under these circumstances was rare. Past Year Rates of Injury and Receipt of Medical Care By Victims of Physical Assault Past year injury - overall sample. The past year rates of physical injury and use of medical care by women who experienced physical assault are presented in Table 11. In the overall sample, 35.2% of the women who were victims of physical assault experienced some type of injury or reported either the need for or use of medical care 29 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence related to partner violence. The most frequently reported types of injuries were sprains, bruises, and minor cuts (27.8%), followed by broken bones (3.7%), and passing out due to a blow to the head (1.9%). With regard to use of medical care, 11.1% perceived that they needed services at some time in the preceding year as a result of intimate partner violence but did not receive it, while only 1.9% reported they had received medical care at least once in the past year related to their victimization. Past year injury – subgroups. In the U.S. born group, 28.1% of the victims of physical assault in the preceding year reported that they experienced an injury, had need for medical care, or received medical care due to violence perpetrated by an intimate partner. In the immigrant group, the rate was 33.3%, and in the migrant group, 71.4%. Across the three groups, minor injuries, such as sprains, bruises, and cuts, were most commonly reported, and few women experienced more severe injuries. Use of medical services related to victimization was also rare. Background and Sociodemographic Factors Associated with Intimate Partner Violence The associations of intimate partner violence with background and sociodemographic characteristics of the respondent and her intimate partner were examined with multiple logistic regression (see Table 12). In this analysis, the outcome variable was past year intimate partner violence victimization, which was considered present if a respondent reported any physical assault, sexual coercion or severe psychological aggression by an intimate partner in the year preceding the survey. In the overall sample, 97(33.2%) women were classified as experiencing intimate partner violence based on these criteria. 30 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Predictors included in the logistic regression model were: respondent age (18-24, 25-31, 32-38, 39-45), respondent marital status (never married, divorced or separated, living with partner, married), poverty status of household relative to the federal poverty threshold (at or below poverty level, above poverty level), number of children in household (0, 1-2, 3 or more), partner education (less than high school, high school or GED, at least some post-secondary education), partner substance use problem (yes/no), and group (U.S. born, immigrant, migrant). Partner substance use problems were coded as present if the recommended cutpoint was met or exceeded on either the partner version of the Alcohol Use Disorders Identification Test or the Drug Abuse Screening Test. The variables included in the logistic regression model were selected a priori based on theoretical considerations and previous research. Diagnostic procedures did not detect problems with collinearity among the predictor variables. Prior to conducting the final logistic regression analysis, we checked for interactions between the dummy variables for group (i.e., U.S. born vs. immigrant, migrant vs. immigrant) and each of the other predictor variables. Separate logistic regression analyses were run in which the group dummy variables, other predictor variable, and group dummy variables X predictor interactions were entered with the outcome variable past year intimate partner violence. There were no significant group X predictor interactions; therefore, no interaction terms were included in the final logistic regression model. All variables were entered simultaneously in the final model. The Hosmer and Lemeshow statistic suggested that the model had a reasonable fit to the data. Table 12 presents the odds ratios and 95% confidence intervals for the sociodemographic and background variables. Only having a partner with substance use 31 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence problems was associated with intimate partner violence. The odds of reporting intimate partner violence were 2.395 times higher (95% confidence interval = 1.205, 4.757) among women whose partners had substance use problems compared to women whose partners did not have such problems. No associations were found for respondent age, marital status, poverty status of household, number of children in household, partner education, and group. Childhood Risk Factors Associated With Intimate Partner Violence Multiple logistic regression was used to examine the associations of past year intimate partner violence with childhood risk factors (see Table 13). As in the preceding analysis, the outcome variable was intimate partner violence victimization, which was considered present if a respondent reported any physical assault, sexual coercion, or severe psychological aggression by an intimate partner in the preceding year. Predictor variables included: parent/caregiver substance use problem (yes/no), parental/caregiver intimate partner violence (yes/no), physical abuse as a minor (yes/no), sexual abuse as a minor (yes/no), and neglect as a minor (yes/no). Control variables, which were selected a priori, were respondent age (18-24, 25-31, 32-38, 39-45), education (less than high school, high school or GED, at least some post-secondary education), and group (U.S. born, immigrant, migrant). Diagnostic procedures did not indicate problems with collinearity among the predictor and control variables. As described above, prior to conducting the final logistic regression analysis, we checked for interactions between each of the predictor variables and the group dummy variables (U.S. born vs. immigrant, migrant vs. immigrant). Separate logistic regression analyses were run in which the group dummy variables, predictor variables, and group X 32 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence predictor interactions were entered with the outcome variable past year intimate partner violence. There were no significant group X predictor interactions; therefore, no interaction terms were included in the final logistic regression model. All variables were entered simultaneously. The Hosmer and Lemeshow statistic suggested that the model had a reasonable fit to the data. Table 13 presents the odds ratios and 95% confidence intervals for the childhood risk and control variables. The results indicated that exposure to parental/caregiver intimate partner violence was associated with adult victimization by an intimate partner. The odds of reporting intimate partner violence were 2.194 times higher (95% confidence interval = 1.094, 4.397) among women who reported violence between their parents or caregivers compared to women who did not report such experiences. Being a victim of sexual abuse as a minor was also associated with victimization by a partner. The odds of reporting intimate partner violence were 3.537 times higher (95% confidence interval = 1.612, 7.762) among women who reported sexual abuse compared to women who did not report such victimization. No associations were found for parent/caregiver substance use problems, physical abuse as a minor, neglect as a minor, respondent age, education, or group (U.S. born, immigrant, migrant). Intimate Partner Violence, Mental Health, Physical Health, and Substance Use Problems The relationships between intimate partner violence victimization in the past year, mental health, and physical health functioning were assessed with multiple linear and logistic regression. In these analyses, intimate partner violence was coded as being present if a respondent reported any physical assault, sexual coercion or severe psychological aggression by an intimate partner in the year preceding the survey. 33 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence Mental health Mental health functioning was assessed with the Global Severity Index of the Brief Symptom Inventory and the mental health scale of the Short-Form Health Survey. The Global Severity Index reflects overall psychological distress with higher scores indicative of greater distress. The SF-12 mental health scale assesses general mental health functioning and limitations due to mental health problems, with higher standard scores reflective of better functioning. Multiple linear regression analyses predicting Global Severity Index and SF-12 mental health scores were conducted. The following variables were simultaneously entered in each respective model: age (years), education (less than high school, high school or GED, at least some post-secondary education), marital status (living with partner, divorced or separated, never married, married), group (U.S. born vs. immigrant, migrant vs. immigrant), and past year intimate partner violence. Interaction terms for intimate partner violence victimization and group status (U.S. born, migrant, immigrant) were initially entered but were not significant, and therefore not retained in the final models. Diagnostic procedures did not detect problems with multi-collinearity. The study also assessed women’s history of suicide attempts with an item from the Beck Scale for Suicide Ideation. History of suicide attempts was coded as present or absent. Multiple logistic regression predicting women’s history of suicide attempts was conducted. The predictor variable in the logistic regression model was intimate partner violence victimization (physical, sexual, or severe psychological) in the past year (yes/no). Control variables, which were selected a priori, were respondent age (18-24, 25- 31, 32-38, 39-45), education (less than high school, high school or GED, at least some 34 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence post-secondary education), marital status (living with partner, divorced or separated, never married, married) and group (U.S. born, migrant, immigrant). All variables were simultaneously entered in the model. BSI Global Severity Index. The means and standard deviations for the Global Severity Index for the overall sample and for women who did and did not report past year intimate partner violence are presented in Table 14. In multiple regression analysis (see Table 15), education was a significant predictor of Global Severity Index scores, with women who had less than high school education more likely to have higher GSI scores (indicative of greater psychological distress) compared to individuals with at least some post-high school education (standardized β = .160, p < 0.05). Marital status was also a significant predictor with women who were divorced or separated (standardized β = .133, p < 0.05) and women who were never married (standardized β = .199, p < 0.01) more likely to have higher GSI scores compared to women who were married. Age and group status were not significant predictors of Global Severity Index scores. Intimate partner violence victimization was significantly related to Global Severity Index scores (standardized β = .272, p < 0.0001), after controlling for demographic and group status variables, with women who experienced intimate partner violence more likely to have higher scores (indicating greater psychological distress). 35 SF-12 mental health scale. The means and standard deviations for the SF-12 mental health standard scores are provided in Table 14, and the multiple linear regression results are presented in Table 16. Similar to the results for the Global Severity Index, marital status was a significant predictor of SF-12 mental health standard scores in the multiple regression analysis, with women who were divorced or separated (standardized This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence β = -.211, p < 0.0001) and women who were never married (standardized β = -.218, p = 0.001) more likely to have lower SF-12 mental health scores (indicating poorer functioning) than women who were married. Age, education, and group status were not significantly associated with SF-12 mental health scores. Past year intimate partner violence was a significant predictor (standardized β = -.241, p < 0.0001). Women who reported victimization were more likely to have lower SF-12 mental health standard scores reflecting poorer functioning. History of suicide attempts. In the overall sample, 15.6% of the women reported one or more suicide attempts (see Table 14). Almost 17% of women who experienced intimate partner violence in the past year and 15% of the women who did not experience violence had a history of suicide attempts. Table 17 presents the odds ratios and 95% confidence intervals for the logistic regression analysis predicting history of suicide attempts. Age, education, and group status were not significantly associated with women’s suicide attempts. The odds of reporting one or more suicide attempts was 3.401 times higher (95% confidence interval = 1.427, 8.107) among women who were never married compared to married women. Being a victim of partner violence in the past year was not associated with a history of attempting suicide. Physical health Physical health functioning was assessed with the physical health scale of the Short-Form Health Survey. The SF-12 physical health scale assesses physical health functioning and limitations due to physical health problems, with higher standard scores indicative of better functioning. 36 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence A multiple linear regression analysis predicting SF-12 physical health standard scores was conducted. The following variables were simultaneously entered in the model: age (years), education (less than high school, high school or GED, at least some post-secondary education), marital status (living with partner, divorced or separated, never married, married), group (U.S. born, migrant, immigrant), and past year intimate partner violence. Interaction terms for intimate partner violence victimization and group (U.S. born vs immigrant, migrant vs. immigrant) were initially entered but were not significant and therefore not retained in the final model. Diagnostic procedures did not suggest problems with multi-collinearity. As reported in Table 18, results of the regression analysis showed that age was a significant predictor of SF-12 physical health standard scores with increasing age associated with lower scores (standardized β = -.167, p < 0.05). Education, marital status, group and partner violence victimization were not significantly associated with physical health scores. Substance use Women’s alcohol use was assessed with the Alcohol Use Disorders Identification Test and drug use with the Drug Abuse Screening Test. Relatively few women in the sample were classified as exhibiting problematic drinking. As shown in Table 14, only 3.4% of the sample had an AUDIT score that met or exceed the recommended cutpoint. The survey also obtained information on the quantity and frequency of alcohol use, and this information was used to categorize levels of use based on a slightly modified version of the classification scheme developed by Kaufman Kantor and Straus (1987). About 52% of the sample were classified as “abstainers”, 30.8% were classified as exhibiting 37 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice. Latinas and Violence “low” alcohol use, 11.8% as exhibiting “moderate” use, and 5.9% as exhibiting “high/binge” use. Similarly, very few women reported problematic drug use. Based on categories developed for the DAST, 92.1% of the sample reported no problems with drug use, 4.5% were classified as having “low” drug use, 2.7% were classified as having “moderate” use, 0.7% as having ‘substantial” use, and none were classified as having “severe” use. It should be noted that regression models predicting women’s alcohol and drug use could not be fit. Estimates were unstable due to the small number of cases reporting problematic alcohol and drug use. Cultural Factors An important objective of this study was to assess the relationship between IPV and various cultural factors. Several cultural measures were used to assess the importance of culture. These were: (1) acculturation (two measures of acculturation (PAN and the SAS), (2) acculturative stress (AS), (3) ethnic identity (MEIM), and (4) bicultural self- efficacy (BISES). Acculturation The Short Acculturation Scale--SAS (Marin et al., 1987) was used, along with another more recently established measure called the Pan Acculturation Scale (Soriano & Hough, 2000). The SAS is an established twelve item scale measuring acculturation on a single dimension ranging from high acculturation (high participation in American culture) to low acculturation (high participation in one’s culture of origin). Like most other acculturation scales (e.g., ARSMA (Cuellar, et al., 1994)), the SAS relies primarily on assessing acculturation using language use as a proxy for cultural participation in one of two cultures--American and in one’s culture of origin. The Pan 38 This document is a research report submitted to the U.S. Department of Justice. This report has not been published by the Department. Opinions or points of view expressed are those of the author(s) and do not necessarily reflect the official position or policies of the U.S. Department of Justice.