Prevalence and Factors Associated With Severe Physical Intimate Partner Violence Among U.S. Black Women: A Comparison of African American and Caribbean Blacks

Abstract

This study explored prevalence rates and factors associated with lifetime severe physical intimate partner violence among U.S. Black women. Data from the National Survey of American Life were examined. Rates of severe physical intimate partner violence were higher among African American women compared with U.S. Caribbean Black women. Risk factors associated with reported abuse were similar to those found in earlier studies but differed by ethnic backgrounds. Demographic, resource, and situational factors were associated with severe physical intimate partner violence among U.S. Black women in general but made unique contributions by ethnic group. Implications and suggestions for future studies were discussed.

Introduction

After decades of research, scholars have documented intimate partner violence (IPV) among women across a range of ethnic/racial groups, ages, socioeconomic statuses, educational levels, and relationship statuses (e.g., married, cohabitating, dating; Renner & Whitney, 2010). Yet it is imperative that researchers continue to investigate the differences in incidence and prevalence of IPV based on race and ethnicity, particularly among Black women in the United States. Based on a systematic review of the published literature in the past 40 years, Black Americans, whether as individuals or couples, consistently report higher rates of overall severe, mutual, and recurrent past-year and lifetime IPV victimization and perpetration when compared with their White and Hispanic counterparts (for a review, see West, 2012).

It is also important to assess ethnic differences in relation to the level of violence experienced by victims. The common practice of “ethnic lumping” has obscured differences within groups. This has occurred among Black Americans who have often been perceived as a monolithic racial group of African descent whose historical roots can be traced to the trans-Atlantic slave trade. However, there are considerable demographic and cultural variations within the Black population. Currently, Black immigrants account for approximately 10% of the total Black population, with just over half originating from Caribbean countries such as Jamaica, Haiti, Trinidad and Tobago, and the Dominican Republic (Kent, 2007; Thomas, 2012). But with a few exceptions (e.g., Sabri et al., 2013), there is a dearth of IPV research on women within this population. Studies have provided evidence of ethnic differences in prevalence rates of IPV among Hispanics (e.g., Mexican American, Puerto Rican; Kantor, Jasinski, & Aldarondo, 1994; West, Kaufman Kantor, & Jasinski, 1998) and Asians (Vietnamese, Filipino, Chinese; Chang, Shen, & Takeuchi, 2009), and a similar pattern likely exists within the U.S. Black population. This study used data from the National Survey of American Life (NSAL) to examine prevalence and associated factors of severe physical IPV among U.S. Black women. The NSAL is the most comprehensive and detailed study of non-institutionalized adult Americans of African descent ever conducted, with the first representative sample of Caribbean Blacks.

Prevalence and Risk Factors of IPV Among U.S. Black Women

Researchers have discovered substantial rates of both overall and severe IPV, defined as acts of beating, choking, threatening, and assaulting with a weapon among Black women (Straus & Gelles, 1990; West, 2012). According to the first National Family Violence Survey (NFVS) conducted in 1975, the rate of severe violence toward wives in Black families was 113 per 1,000. A decade later, the 1985 National Family Violence Resurvey (NFVR) revealed that 17% of Black wives reported at least one act of aggression during the survey year, and 7% suffered a severe act of aggression (Hampton & Gelles, 1994). In the 2000 National Longitudinal Couple Survey (NLCS), 4% of Black married/cohabiting couples reported severe violence that was perpetrated by the male partner during the year prior to the survey (Caetano, Field, Ramisetty-Mikler, & Lipsky, 2009). More recently, 40.9% of Black women in the National Intimate Partner and Sexual Violence Survey (NISVS) reported physical victimization that was committed by an intimate partner during their lifetime (Black et al., 2011). The varying estimates in rates of IPV over time along with limited abuse research on Caribbean women provide a need for additional studies using large national Black samples that may provide more accurate estimates as well greater understanding of sociocultural factors that may be associated with IPV.

The prevalence of IPV cannot be explained by a single factor. Rather, there have been several influences associated with IPV, including demographic, resource, and situational factors. Across racial groups, higher rates of reported IPV were more frequent among younger couples. In particular, women between the ages of 18 and 24 years experienced the highest per capita rates of intimate violence (19.6 per 1,000; Rennison & Welchans, 2000). Similarly, the NISVS found that nearly one-half (47.5%) of women who experienced some form of IPV (rape, physical violence, stalking) were first victimized between 18 and 24 years of age (Black et al., 2011). This pattern was also found among Black Americans where rates of severe male-to-female partner violence were more than 3 times greater for Black couples under the age of 30, compared with Black couples older than 40 years of age (Hampton & Gelles, 1994).

Residence in certain geographic locations has also been associated with increased risk of IPV. Around 10 per 1,000 urban women were victims of IPV compared with 8 per 1,000 women in suburban and rural areas (Rennison & Welchans, 2000). Furthermore, residence in impoverished neighborhoods has been associated with higher rates of IPV. Black women who live in impoverished areas are at a threefold risk for male-to-female partner violence compared with women who do not reside in poor areas (Cunradi, Caetano, Clark, & Schafer, 2000). Black women who reside in southern states (South Carolina, Louisiana, Tennessee) also appear to be overrepresented among victims of femicide (Violence Policy Center, 2013). However, less is known about regional and race/ethnic differences in rates of severe IPV.

The association between relationship status and IPV is more complex. Although Frieze (2005) discovered that married, divorced, and never-married couples reported comparable rates of IPV, this was not fully supported in other studies. For example, among Black college students, women who were married or in domestic partnerships were more likely to report any past-year IPV (Barrick, Krebs, & Lindquist, 2013). In contrast, divorced and separated women experienced the highest rates of intimate partner victimization, followed by never-married women. When severity of IPV was considered, cohabitating couples had more than 6 times the rate of IPV compared with their dating and married counterparts (Stets & Straus, 1989). Nonetheless, researchers could not ascertain whether the victims were separated or divorced at the time of the abuse or if relationship termination followed the violence (Rennison & Welchans, 2000).

A consistent risk marker of partner violence has been the woman’s socioeconomic status, measured by education, occupation, and income. Generally, women with access to fewer economic resources report higher rates of IPV (Field & Caetano, 2004; Hampton & Gelles, 1994). To illustrate, among welfare recipients, recently abused Black women reported less stable employment and worked less consistently than Black women who had not experienced abuse (Staggs & Riger, 2005). Black couples who also reported male-to-female partner violence had significantly lower mean annual incomes than couples who did not report male perpetrated IPV (Cunradi, Caetano, & Schafer, 2002). However, National Crime Victimization Survey (NCVS) data revealed that when compared with Black women with less than a high school education, Black women who graduated from college or had additional education beyond a 4-year degree reported more sexual and other forms of abuse (Christy-McMullin, 2005). Results from national data further revealed that among Black couples, there were more reported incidents of male-to-female IPV among employed women than among those who were retired (Cunradi et al. 2000). There was even greater frequency of recent IPV among college-educated abused Black women in Baltimore than non-abused women (Stockman et al., 2013).

The association between home ownership (a proxy for socioeconomic status) and IPV has been less clear. Data from the NCVS revealed that women who lived in rental housing were victimized by an intimate partner at more than 3 times the rate of women who lived in owned homes (Rennison & Welchans, 2000). Among Black women specifically, renters were more likely to report emotional abuse than women who lived in households where the homes were owned or in the process of being purchased. Yet the link between renting and emotional abuse was no longer significant when age and marital status were considered (Christy-McMullin, 2005).

Finally, studies have provided compelling evidence of the relationship between IPV and alcohol use and drinking patterns (Devries et al., 2014; Lacey, McPherson, Samuel, Sears, & Head, 2013). When compared with White and Hispanic couples, drinking alcohol during the violent episode was more frequently found among Black couples (Caetano, Schafer, & Cunradi, 2001). Black couples who reported female alcohol-related problems had a threefold risk of male-to-female partner violence compared with Black couples not reporting female alcohol-related problems. At the same time, Black couples who reported male alcohol-related problems had a sevenfold risk of male-perpetrated IPV compared with Black couples not reporting male alcohol-related problems (Cunradi et al., 2000). Alcohol’s role in IPV may be explained by alcohol’s physiological disinhibiting effects and because heavy drinking may contribute to impulsive behaviors (Caetano et al., 2001; Field & Caetano, 2004).

Prevalence and Risk Factors of IPV Among Caribbean Women

A growing body of IPV research has been documented in various Latin American and Caribbean countries with estimates ranging from 17–52.3% (Bott, Guedes, Goodwin, & Mendoza, 2012). For example, approximately half of the women in Barbados (50%), Jamaica (45.3%) and Trinidad and Tobago (45.2%) reported victimization by an intimate partner (Le Franc, Samms-Vaughan, Hambleton, Fox, & Brown, 2008). In Haiti, 29% of women experienced some form of IPV in the past 12 months, with 13% having experienced at least two different forms (physical, emotional, sexual) (Gage, 2005). In the first-ever IPV module within the Behavioral Risk Factor Surveillance System (BRFSS) conducted by the Centers for Disease Control (CDC), it was found that 22.5% of women of African descent in the U.S. Virgin Islands have been victims of IPV (Breiding, Black, & Ryan, 2008).

There are similarities in the profiles of battered Caribbean women to those found in general studies. Specifically, rates of severe victimization declined with age (Le Franc et al., 2008); younger women between the ages of 20 and 49 years who visited walk-in clinics in Trinidad were more likely to report abuse compared with women between the ages of 51 and 61 years (Maharaj et al., 2010).

Regarding relationship status, partnered Caribbean women reported higher rates of IPV compared with those who were not involved with an intimate partner (e.g., widowed, divorced; Maharaj et al., 2010). In another study on abused Trinidadian women, high rates of IPV were found among women who were married (55.6%), in common-law relationships (40.7%), or those who were single (66.7%) (Maharaj et al., 2010).

The association between various economic resources and IPV for women of African Caribbean descent is sparse and contradictory. In particular, when Haitian women had access to money that they could use in any way they wanted, their risk for physical victimization was reduced by 44% (Gage, 2005). In addition, rates of IPV were greater for higher income women who used walk-in clinics in Trinidad compared with lower income women (Maharaj et al., 2010). However, when households were stratified into different socioeconomic classes in Central Trinidad, the greatest prevalence of IPV was found within the working-class and lower-middle class groups when compared with the middle- and upper-class categories (Nagassar, Sampson, Chankadyal, Ramasir, & Boodram, 2010).

As expected, Trinidadian women with less education reported higher rates of intimate partner victimization than their more educated counterparts (Maharaj et al., 2010). Lower odds for sexual violence were found for higher educated Haitian women relative to their male partners, whereas failure to complete primary school was a risk factor for all forms of IPV (Gage, 2005).

The research concerning economic resources as measured by employment status was less clear. Employed Trinidadian women were more likely to report abuse than their unemployed counterparts (Maharaj et al., 2010). Conversely, lower-class women in Central Trinidad identified unemployment and financial constraints as the reasons for IPV in their homes (Nagassar et al., 2010).

Alcohol use was also associated with higher rates of IPV among Caribbean women. Haitian women who reported that their partner had a history of drunkenness were over 3 times more likely to be physically assaulted than those with partners who had never been inebriated (Gage, 2005). Furthermore, females in Central Trinidad most commonly cited drug and alcohol as the cause of physical abuse (Nagassar et al., 2010).

Goals of the Study

This study examined prevalence rates and factors associated with severe lifetime physical partner violence of U.S. Black women. Another goal of the study was to address the contributions of demographic, resource, and situational factors in accounting for severe physical IPV among three cohorts of women: U.S. Black women in general, African American women, and Caribbean Black women residing in the United States.

Method

Sample

The study was based on cross-sectional data from the NSAL. As part of the Collaborative Psychiatry Epidemiological Study (CPES), the NSAL is the most comprehensive and detailed study of non-institutionalized adult Americans of African descent conducted, and the first-ever representative sample of Caribbean Blacks in the United States (see Jackson et al., 2004). The NSAL was designed to explore racial and ethnic differences in mental disorders, psychological distress, and informal service use from within the context of a variety of presumed risk and protective factors in the African American and Afro-Caribbean populations of the United States (Jackson, 2007; Jackson et al., 2004). Collection of the data began in February 2001 and was completed in March 2003. Multi-stage probability sampling procedures were used to generate the sample. In-person interviewing was the primary method of data collection, with approximately 10% of the interviews conducted entirely or partially by phone due to scheduling conflicts, privacy issues, or safety concerns. Interviews typically lasted 1 hr 30 min–2 hr 45 min but differed slightly between ethnic groups. Interviews with African Americans on average lasted approximately 2 hr 20 min in length compared with an average of 2 hr 43 min for Caribbean Blacks. In total, the sample consisted of 6,082 participants: 3,570 African Americans, 1,623 Caribbean Blacks, and 891 non-Hispanic Whites. An overall response rate of 72.3% was obtained: African American 70.7%, Caribbean Blacks 77.7%, and non-Hispanic Whites 69.7%. For this study, the analyses focused on women of African descent.

Measures

IPV

To determine whether participants had experienced severe IPV, they were asked the question, “Were you ever badly beaten up by a spouse or romantic partner?” In the survey, this question was measured on a binary scale (e.g., yes or no). For the current analyses, a code of “0” was given to participants who did not report experiencing severe violence, and “1” was assigned to those participants who responded in the affirmative. The IPV measure used was compared with the National Comorbidity Survey Replication (NCS-R) dichotomously defined severe partner violence measure from the severe physical violence subscales of the Conflict Tactics Scale (CTS) (Kessler, Molnar, Feurer, & Applebaum, 2001; Straus, 1979) within the CPES, and was found to have fair agreement across two different estimates (odds ratio [OR] = 4.5, p < .001; area under the receiver operating characteristic [ROC] curve [AUC] > 0.65). Respondents who affirmed having experienced severe partner violence within the NSAL were highly likely to have affirmed having experienced severe partner violence within the CPES, and the AUC was greater than or equal to 0.65 indicating fair ability of the current IPV measure to correctly classify those who have experienced severe partner violence and those who have not when compared with the CTS measure in the CPES (Lalkhen & McCluskey, 2008).

Demographic variables

The demographic variables included participants’ race/ethnicity, age, and geographical region. The racial/ethnic backgrounds of participants in the study were U.S. Blacks, African Americans, and U.S. Caribbean Blacks. U.S. Blacks refers to the entire sample as a whole. African Americans were persons who self-identified as Black but did not report Caribbean ancestry. U.S. Caribbean Blacks were persons who self-identified as Black and answered in the affirmative to inclusion criteria such as whether (a) they were of West Indian or Caribbean descent, (b) they were from a Caribbean-area country, or (c) they had parents or grandparents who were born in a Caribbean-area country (Williams et al., 2007). The sample includes individuals of English, French, and Spanish-speaking Caribbean regions, with a few from the Dutch Caribbean (Jackson et al., 2004). Age was separated into five categories: 18–24, 25–34, 35–49, 50–64, and 65 and above. Geographic regions included Northeast, Midwest, South, and West.

Resource variables

Participants’ relationship status, education, household income, occupational status, and home ownership were considered resource factors. The relationship status variable included categories of married, partnered, separated or divorced, widowed, and never married. Level of education was categorized into four: less than high school, high school graduate, some college, and college educated. The income level of participants was separated into the following categories: less than US$25,000; US$25,000–US$34,999; US$35,000–US$49,999; US$50,000–US$74,999; and US$75,000 and above. Occupational status consisted of participants who were employed, unemployed, and not in the labor force. The home ownership measure consisted of participants who owned homes versus participants who did not own or rented.

Situational factors

We used alcohol abuse as an indicator of an important contextually confounding factor. Criteria for alcohol abuse was determined by a slightly modified version of the World Health Organization Composite International Diagnostics Interview (WHO CIDI) defined by the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994).

Analytic Strategy

Specialized statistical techniques were used to account for the multistage sample design of the NSAL. All analyses were conducted using STATA 11, which uses the Taylor expansion approximation technique for calculating the complex designed-based estimates of variance (StataCorp, 2009). Cross-tabulations were conducted to assess ethnic difference on rates of severe IPV and differences by demographic and resource variables. We used the Rao–Scott chi-square statistics representing a complex design-corrected measure to assess associations. Nested hierarchical logistic regression analysis was also conducted to explore the significant contributions of demographic, resource, and situational factors in predicting severe physical intimate partner abuse. Negelkerke’s R² was computed and reported. A cutoff level of .05 was set to assess statistical significance.

Weighting

Design weights were included in analysis to adjust for differential sample selection and non-response. Standard errors were corrected for weighting, clustering, and stratification.

Sample

Regarding the demographic profile of the sample, a third of women across cohorts were between the ages of 35 and 49 years (see Table 1). Around one third of the sample of women had never been married. The level of educational attainment of women differed significantly by ethnic groups. U.S. Caribbean Black women tended to have significantly higher levels of educational attainment than African American women (p = .012). That is, 19.8% of Caribbean Black women had attained a college degree versus 13.9% of African American women. Although around one half (50.6%) of women in the sample had household income less than US$25,000; significantly more African American women had incomes within this category compared with U.S. Caribbean Black women (51.3% vs. 38.8%; p = .019). Most participants across groups were employed. Although rates were not significantly different by ethnic groups, a higher proportion of employed women were of Caribbean descent in comparison with African Americans (63.2% vs. 71.1%; p = .060).

Table 1.

Characteristics of Female Participants.

Characteristicsb	U.S. Blacks % (No.)	African Americans % (No.)	Caribbean Blacks % (No.)	Ethnic differencea
Age of participants
18–24	15.9 (440)	15.8 (300)	17.8 (140)	p = .613
25–34	20.5 (755)	20.5 (528)	21.0 (227)
35–49	33.1 (1,124)	33.0 (762)	35.4 (362)
50–64	18.2 (575)	18.3 (414)	15.6 (161)
>65	12.3 (383)	12.5 (2,299)	10.2 (88)
Relationship status (%)
Married	27.4 (802)	27.3(517)	28.5 (285)	p = .117
Partnered	8.5 (223)	8.2(152)	11.6 (71)
Separated or divorced	20.3 (759)	20.2(540)	21.4 (219)
Widowed	11.5 (360)	11.8(300)	6.4 (60)
Never married	32.8 (1,121)	32.4(779)	32.1 (342)
Education (%)
Less than high school	24.7 (772)	25.0 (600)	21.0 (172)	p = .012
High school graduate	36.2 (1,152)	36.6 (865)	30.2 (287)
Some college	24.9 (799)	24.6 (525)	29.1 (274)
College	14.2 (554)	13.9 (309)	19.8 (245)
Household income (%)
<US$25,000	50.6 (1,706)	51.3 (1,314)	38.8 (392)	p = .019
US$25,000–US$34,999	14.4 (502)	14.2 (326)	16.5 (176)
US$35,000–US$49,999	14.4 (461)	14.3 (304)	16.9 (157)
US$50,000–US$74,999	11.9 (360)	11.6 (217)	15.7 (143)
>US$75,000	8.8 (248)	8.5 (138)	12.1 (110)
Occupation status (%)
Employed	63.7 (2,140)	63.2 (1,447)	71.1 (693)	p = .060
Unemployed	11.1 (361)	11.1 (262)	10.3 (99)
Not in the labor force	25.2 (771)	25.7 (585)	18.7 (186)
Region (%)
Northeast	18.8 (977)	15.8 (261)	63.8 (716)	p = .000
Midwest	18.9 (400)	19.9 (393)	3.7 (7)
South	54.0 (1,756)	55.8 (1,508)	26.9 (248)
West	8.3 (144)	8.5 (137)	5.6 (7)
Home ownership
Own home	47.2 (1,423)	47.4 (1,060)	43.4 (363)	p = .227
Rental/does not own home	52.9 (1,854)	52.6 (1,239)	56.6 (615)
Ever badly beaten by partner	17.5 (505)	17.9 (392)	12.0 (113)	p = .032
Number of times victimized
1	21.7 (98)	21.7 (73)	21.7 (25)	p = .917
2–9	39.1 (197)	39.0 (151)	41.7 (46)
10 or more	39.3 (180)	39.8 (145)	36.9 (35)
N	3,277	2,299	978

Note. The figures are based on within group percentages.

^aP-values are based on chi-square tests obtained from bivariate analysis between the African Americans and Caribbean Blacks.

^bDescriptives are based on weighted percentages.

The sample was almost evenly divided between homeowners and renters (47.2% vs. 52.9%) with no significant difference by ethnic groups. Both African American and U.S. Caribbean Black women had slightly higher rental rates (52.6% and 56.6%, respectively). Overall, the majority (54.0%) of women resided in the South. U.S. Caribbean Black women were typically concentrated in the Northeast regions compared with African American women who were more likely to reside in the South (63.8% vs. 55.8%; p = .000).

Results

Prevalence and Number of Times Victimized

Overall, lifetime severe physical IPV for U.S. Black women was 17.5% (see Table 1), but prevalence of intimate victimization between Black ethnic groups differed significantly. Higher rates of severe physical partner violence were found for African Americans when compared with U.S. Caribbean Black women (17.9% vs. 12.0%; p = .032). However, the number of times women in the sample experienced severe physical abuse was not significantly different by ethnic designations. Women across the cohorts experienced severe physical IPV at least 10 or more times on average over the course of their lives.

Associated Risk Factors of Severe Physical IPV

U.S. Black women

Women’s age was associated with severe physical IPV (p = .000; see Table 2). Women between the ages of 50 and 64 years generally experienced the highest rate of severe physical abuse (22.1%). An association was also found between relationship status and severe physical IPV (p = .000). Notably, separated or divorced women experienced higher proportions of severe physical partner abuse (28.8%). The results further revealed that socioeconomic status was a risk factor for severe physical partner violence. Rates of severe physical partner violence were particularly high for participants with less than a high school education (23.9%, p = .001), women who earned less than US$25,000 (21.7%, p = .001), and unemployed women (25.0%, p = .002). Additionally, reported severe physical partner violence was higher for women who did not own a home and were renting (21.5%, p = .000). No association was found between geographic region and severe physical IPV among women in the sample.

Table 2.

Prevalence of Lifetime Severe Physical Intimate Partner Abuse.

Characteristics	U.S. Black women % (SE)	African Americans % (SE)	Caribbean Blacks % (SE)
Age of respondents
18–24	9.6 (1.99)	9.7 (2.11)	8.3 (3.42)
25–34	17.5 (1.50)	17.6 (1.56)	15.1 (4.53)
35–49	21.9 (2.20)	22.5 (2.37)	13.4 (2.94)
50–64	22.1 (2.07)	22.7 (2.18)	10.6 (3.46)
>65	9.1 (1.99)	9.1 (2.11)	9.3 (3.39)
χ2	100.8	78.1	10.0
p value	.000	.000	.544
Relationship status (%)
Married	13.2 (1.82)	13.5 (1.95)	9.1 (3.21)
Partnered	20.7 (3.52)	21.1 (3.78)	16.1 (6.59)
Separated or divorced	28.8 (2.26)	29.6 (2.41)	17.2 (4.27)
Widowed	14.0 (2.65)	14.2 (2.75)	9.8 (3.84)
Never married	14.5 (1.34)	14.8 (1.42)	9.9 (2.73)
χ2	127.0	96.2	19.3
p value	.000	.000	.332
Education (%)
Less than high school	23.9 (1.77)	24.4 (1.86)	14.8 (4.20)
High School graduate	15.4 (1.47)	15.3 (1.56)	17.4 (3.92)
Some college	16.8 (1.98)	17.3 (2.15)	9.9 (2.94)
College	13.1 (2.44)	13.9 (2.69)	3.9 (0.93)
χ2	51.8	38.4	39.8
p value	.001	.003	.010
Household income (%)
<US$25,000	21.7 (1.50)	21.9 (1.58)	17.9 (4.43)
US$25,000–US$34,999	15.1 (2.12)	14.8 (2.23)	18.9 (6.36)
US$35,000–US$49,999	13.2 (2.26)	13.9 (2.46)	4.0 (1.36)
US$50,000–US$74,999	14.3 (2.90)	15.1 (3.17)	5.2 (1.12)
>US$75,000	9.0 (2.73)	9.5 (2.95)	2.9 (1.47)
χ2	70.6	46.7	79.9
p value	.001	.004	.002
Occupation status (%)
Employed	16.1 (0.99)	16.6 (1.06)	10.2 (1.71)
Unemployed	25.0 (2.94)	25.4 (3.06)	19.2 (10.28)
Not in the labor force	17.7 (1.75)	17.9 (1.81)	14.4 (4.87)
χ2	26.6	18.5	14.2
p value	.002	.004	.376
Region (%)
Northeast	19.0 (2.39)	20.9 (3.02)	11.5 (1.79)
Midwest	22.9 (2.08)	22.9 (2.10)	21.6 (14.66)
South	14.8 (1.28)	15.0 (1.32)	11.0 (3.52)
West	20.2 (6.22)	20.4 (6.47)	15.8 (11.84)
χ2	35.5	28.9	7.8
p value	.110	.110	.655
Home ownership
Own home	13.0 (1.03)	13.5 (1.10)	4.7 (1.25)
Rental/does not own home	21.5 (1.49)	21.8 (1.60)	17.5 (3.54)
χ2	63.9	42.9	65.5
p value	.000	.015	.001

Note. The IPV measure used was compared with the NCS-R dichotomously defined severe partner violence measure derived from the Severe Physical Violence subscales of the Conflict Tactics Scale (CTS; Kessler, Molnar, Feurer, & Applebaum, 2001; Straus, 1979), and was found to have fair agreement across two different estimates (OR = 4.5, p < 0.001; AUC > 0.65). IPV = intimate partner violence; OR = odds ratio; AUC = area under the receiver operating characteristic curve.

African American women

Within ethnic group analysis produced similar results for African Americans and the U.S. Black female population in general. For example, increases in severe physical partner violence were observed for women between the ages of 35 and 49 (22.5%) and 50 and 64 (22.7%; p = .000). Separated or divorced African American women also had increased levels of severe abuse (29.6%, p = .000). Also, severely battered African American women were found to have lower socioeconomic backgrounds. Higher victimization rates were observed for African American women who earned less than a high school education (25.4%, p = .003), had household incomes lower than US$25,000 (21.9%, p = .004), and were unemployed (25%, p = .004). Rates of severe physical partner violence were also higher for women who did not own a home (21.8%, p = .015).

U.S. Caribbean Black women

Among U.S. Caribbean Black women, risk for lifetime severe physical IPV was limited to educational background, household income level, and home ownership. Higher victimization rates were found among high school graduates (17.4%, p = .010). Women with income levels ranging from US$25,000–US$34,999 also experienced higher levels of severe physical partner violence (18.9%, p = .002). Similar to other women in the sample, Caribbean Black women who did not own a home reported higher severe victimization rates (17.5%, p = .001). Associations were not found between severe physical intimate victimization and other risk factors (e.g., age, region, occupational status, region) for women within this population.

Multivariate Analyses Predicting Severe Partner Violence

Table 3 illustrates the contributions of demographic, resource, and situational factors in predicting severe physical partner violence among the three cohorts of women. The sets of variables were included at each successive step of the three models. R² change was examined at each step when other factors were controlled. Ethnic backgrounds were excluded in models that focused on African American and Caribbean women separately.

Table 3.

Summary of Hierarchical Logistic Regression Analysis Examining the Contributions of Demographic, Resource, and Contextual Factors in Predicting Lifetime Severe Physical Abuse of U.S. Black Women.

Characteristics	US Black women	AOR	African Americans	AOR	Caribbean Blacks	AOR
Step 1: Demographic	Ethnicity		Ethnicity		Ethnicity
	Caribbean Blacks	0.645	—	—	—	—
	Age		Age		Age
	18–24	1.44	18–24	1.46	18–24	1.37
	25–34	2.87***	25–34	2.89**	25–34	2.97
	35–49	3.33***	35–49	3.40***	35–49	2.72
	50–64	3.43***	50–64	3.51***	50–64	2.66
	Region		Region		Region
	Midwest	1.25	Midwest	1.23	Midwest	1.60
	South	0.705	South	0.692	South	1.02
	West	0.856	West	0.837	West	0.791
	R2	3.5***	R2	3.4**	R2	2.0
	ΔR2	3.5***	ΔR2	3.4**	ΔR2	2.0
Step 2: Resource	Education		Education		Education
	Less than HS	1.55	Less than HS	1.54	Less than HS	2.86*
	HS graduate	1.00	HS graduate	0.959	HS graduate	2.78**
	Some college	1.07	Some college	1.05	Some college	1.78
	Income		Income		Income
	<US$25,000	1.94	<US$25,000	1.87	<US$25,000	2.88
	US$25,000–US$34,999	1.19	US$25,000–US$34,999	1.10	US$25,000–US$34,999	3.12
	US$35,000–US$49,999	1.21	US$35,000–US$49,999	1.22	US$35,000–US$49,999	0.723
	US$50,000–US$74,999	1.58	US$50,000–US$74,999	1.61	US$50,000–US$74,999	1.32
	Occupation status		Occupation status		Occupation status
	Unemployed	1.39	Unemployed	1.40	Unemployed	0.969
	Not in labor force	1.06	Not in labor force	1.04	Not in labor force	1.52
	Home owner		Home owner		Home owner
	Owned home	0.719**	Owned home	0.746*	Owned home	0.201*
	Relationship status		Relationship status		Relationship status
	Partnered	1.21	Partnered	1.21	Partnered	1.42
	Sep-Div	1.93**	Sep-Div	2.01**	Sep-Div	1.23
	Widowed	1.20	Widowed	1.22	Widowed	0.764
	Never married	0.859	Never married	0.881	Never married	0.671
	R2	9.6**	R2	11.2***	R2	14.7**
	ΔR2	6.1**	ΔR2	7.8***	ΔR2	12.7**
Step 3: Situational	Alcohol use		Alcohol use		Alcohol use
	Used alcohol	3.63***	Used alcohol	3.74***	Used alcohol	2.09
	R2	9.8***	R2	11.4***	R2	14.8
	ΔR2	0.2***	ΔR2	0.2***	ΔR2	0.1

Note. Adjusted odds ratios (AOR) are reported for the final step of the analysis. Reference categories are not included but can be found in previous tables.

HS = high school; Sep-Div = separated or divorced.

^*p < .05.

^**p < .01.

^***p < .001.

Among Black women in general (Model 1), all sets of variables contributed significantly to the model and explained 9.8% of the variance in severe physical partner violence: demographic variables, 3.5%; resource variables, 6.1%; and situational variables, 0.2%. Although these variables combined contributed to the model, a stronger association was found between the resource variables and severe physical partner violence.

In Model 2 that focused on African American women, demographic, resource, and situational variables also contributed significantly to severe physical IPV. Altogether, these variables combined represented 11.4% of the variance in reports of severe physical violence. Individually, demographic factors represented 3.4%, resource 7.8%, and situational 0.2% of the variance in severe physical partner violence. Again, the resource variables contributed the most to the model and were more strongly associated with severe physical IPV, though other factors contributed uniquely.

In the analysis that focused on U.S. Caribbean women (Model 3), demographic, resource, and situational factors explained 14.8% of the variance in severe physical IPV. Noticeably in this model, as compared with others, a larger percentage of the variance was explained by the resource variables (12.7%). Demographic (2.0%) and situational (0.1%) variables did not contribute significantly and provided very little additional explanation of severe physical partner violence among this group of women.

Discussion

The study addressed prevalence and factors associated with lifetime severe physical abuse using a national representative sample consisting of U.S. Blacks. According to our study findings, rates of severe physical IPV for Black women were lower in comparison with those found in recent national studies (see Black et al., 2011; Caetano et al., 2001; Tjaden & Thoennes, 2000). Rates of victimization also differed significantly by ethnic groups. African American women experienced higher rates of severe victimization by a spouse or romantic partner compared with Caribbean Black women.

Similar to the general studies, risk factors for severe IPV included age, relationship status, education, household income, occupational status, and household ownership for Black women (Hampton & Gelles, 1994; Rennison & Welchans, 2000; Tjaden & Thoennes, 2000). However, there were slight differences in risk markers for severe partner violence in relation to previous general studies. Contrary to studies that reported higher victimization rates among younger women, Black women in this study between the ages of 50 and 64 years reported higher rates of severe physical intimate victimization; this may be a more accurate representation of the victimization rate for women within the U.S. Black population. The findings further revealed that separated or divorced women experienced high levels of severe physical intimate partner victimization. This finding may reflect some women’s continued victimization by their spouse or partner, even after separating from them. Abused women within the sample could have also had other visiting relations that turned out to be abusive, while separated or divorced. The study also revealed home ownership (or a lack thereof) as a risk factor for severe physical IPV among U.S. Black women. Higher rates of severe physical IPV were found among women who did not own or were renting.

The study findings further highlight differences in risk factors for severe physical partner violence between ethnic groups within the U.S. Black population. Although many of the proposed risk factors of severe physical IPV among African American women mirrored that of the U.S. Black population, they differ for U.S. Caribbean Black women. Risk markers for severe physical IPV for Caribbean women were typically limited to women’s level of education, household income, and whether the woman had been renting a home.

The multivariate findings showed that although demographic, resource, and situational variables contributed to severe physical partner violence in general, a stronger association was found for the resource variables. This was especially true for Caribbean women. Women’s alcohol use also had a negligible effect on severe partner violence among women within the Caribbean population, as compared with women from the other cohorts. This has been supported by studies that found lower risk of substance disorder among Caribbean women compared with African American women (Lacey, Powell Sears, Matusko & Jackson, 2015; Williams et al., 2007).

The study was limited to severe physical victimization and did not account for other types of violence (e.g., psychological, emotional, and stalking). Severe physical abuse arguably can have more immediate detrimental effects on the health and well-being of women, and may be readily identified by some ethnic women as the major form of abuse (Torres et al., 2000; Yick & Agbayani-Siewert, 1997). Another potential limitation stems from the use of a single dichotomous measure to estimate IPV, although single-item measures have been used in prior IPV research (see Flake & Forste, 2006; Lacey et al., 2015). The measure used in this study was compared with a standard severe IPV measure and was found to have fair association. The measure used made it possible to estimate rates and predictive factors of abuse across ethnic groups among U.S. Blacks, including Caribbean Black women using a national representative sample, which had never been possible before. Another possible limitation of the study was the reliance of women’s own accounts of violence over their life course. Lapses in memory on specific events could play a role in the validity of findings, particularly if the experience was traumatic and the woman sought to suppress the memory. Finally, the study only assessed unidirectional victimization; we lack information on female perpetrators.

Although there are a number of limitations to the study, it should be noted that this is the first study to use a predominantly Black national sample to estimate prevalence and associated factors of severe physical IPV. Not only was the use of this large national sample of individuals of African descent useful in clarifying risk markers of severe physical partner violence among U.S. Black women, but it also made it possible to examine ethnic variations among abused women, while providing estimates on rates of violence among the Caribbean Black population, one of the fastest growing subgroups in the United States (Manuel, Taylor, Zhang, & Jackson, 2012).

This study is foundational and has many policy and intervention implications. For instance, the finding provides validation that U.S. Black women are not a monolithic group and are affected differently by severe physical IPV. By aggregating groups in a single combined category, we run the risk of making broad assumptions that may not be factual of all ethnic groups, which has been supported by this study. Therefore, it is important to consider social, ethnic, and cultural backgrounds when conducting IPV studies. Similarly, individuals’ ethnic backgrounds must also be kept in mind when developing prevention and intervention strategies. Tailoring interventions with respect to ethnic and cultural backgrounds may be more instrumental in successfully assisting abused women. As found in this study, the risk factors for violence were dissimilar between African American and U.S. Caribbean Black women, and these women may have different needs, warranting different strategies and approaches for intervention and services. Differences in strategies and approaches are also essential when dealing with undocumented immigrant or Caribbean women who may experience an increase in violence and are trapped in relationships and may not seek help because of their immigration status. Consequently, a single approach may not be the most effective in developing policy, services, and interventions aimed at addressing the unique needs of abused women within a diverse Black population.

The study findings also provide suggestions for future studies including continued exploration of research that examines group differences. There is a particular need for studies on Caribbean women where research remains limited. Another area for future exploration is to understand what roles structural conditions, cultural context, and immigration status may play in IPV among U.S. Black and immigrant women. Also, studies should investigate a range of other risk factors for IPV, including a history of violence exposure in women’s family of origin through witnessing violence between parents or in the form of child abuse victimization and quality of the relationship (Gage, 2005). Finally, comparative studies are necessary to understand the experiences of severely abused Caribbean women within the Caribbean context itself and migrating countries, such as Canada, England, and other European nations.

Biographies

Krim K. Lacey, PhD, earned his degree in sociology from Wayne State University. He currently has research affiliation with the University of Michigan’s Institute for Social Research, Program for Research on Black Americans. His primary research interest is intimate partner violence, particularly focusing on minority and immigrant populations. He has also been engaged in research that addresses the influence of social context and cultural factors on the physical and mental well-being of Caribbeans residing in the United States and in the Caribbean region.

Carolyn M. West, PhD, is an associate professor of psychology in the Interdisciplinary Arts and Sciences Program at the University of Washington. She is an award-winning author and fellow of the American Psychological Association. She writes, trains, consults, and lectures internationally on interpersonal violence and sexual assault. She is author of numerous book chapters and journal articles, and is editor/contributor of Violence in the Lives of Black Women: Battered, Black, and Blue, which was published by Routledge in 2002.

Niki Matusko, BS, is a research area specialist for the Program for Research on Black Americans at the University of Michigan Institute for Social Research. She is well versed in advanced statistical procedures such as structural equation modeling, hierarchical linear modeling, time series, and autoregressive integrated moving average (ARIMA) forecasting. In addition, she has extensive experience in compiling and presenting secondary research such as market share quantification, in-migration and out-migration patterns, and various strategic intelligence analyses. Her roles have included client research analyst, health information specialist, and strategic market analyst. She has a bachelor of science degree in mathematics from the University of Michigan.

James S. Jackson, PhD, is the Daniel Katz distinguished university professor of psychology, professor of Afro-American and African Studies, and director of the Institute for Social Research, all at the University of Michigan. His research focuses on issues of racial and ethnic influences on life course development, attitude change, reciprocity, social support, and coping and health among Blacks in the Diaspora. He is currently directing the most extensive social, political behavior, and mental and physical health surveys on the African American and Black Caribbean populations ever conducted. He is the President of the Consortium of Social Science Associations.