Women’s experiences with violence: A national study

Article Outline

• Abstract
• Methods
  • Study Design
  • Instrument Development
  • Variable Definitions
    • Personal experiences with violence
    • Demographic variables
  • Data Collection
  • Data Analyses
• Results
  • Sociodemographic Characteristics of the Study Participants
  • Women’s Experiences With Violence
• Conclusions and Discussion
  • Study Limitations
  • Conclusion and Implications
• Acknowledgments
• References
• Biography
• Copyright

Background

Violence against women (VAW) is widespread and linked to negative public health and social outcomes. Research on VAW, however, has largely been limited to convenience samples and on variable definitions of violence, hindering our ability to fully characterize this important problem nationally and among subgroups of women.

Methods

Using a population-based national sample of noninstitutionalized women ages ≥18 (n = 1,800), we conducted a telephone survey on women’s experiences with 6 types of violence, including being followed and repeatedly contacted, as well as physical and sexual assault by intimate partners and others. We calculated adult lifetime and prior year prevalence of violent experiences, examined bivariate differences in experiences among groups of women, and employed logistic regression to model the odds of adult lifetime and prior year victimization.

Results

Sixty percent of the respondents experienced at least 1 form of violence since age 18; 10% reported violence in the previous year. Adult lifetime and prior-year prevalence varied widely by types of violence, and by respondents’ sociodemographic characteristics. Women under age 55, those receiving public assistance, and lesbian/bisexual women were at higher risk of experiencing violence in their adult lifetimes. Women age 18–24 had increased risks of victimization in the previous year.

Conclusions

To accurately reflect the chronic nature of partner violence, point estimates should be supplemented with adult lifetime estimates of victimization, including stalking behaviors. Ensuring adequate numbers of women from diverse backgrounds and developing measures that more completely assess the patterns and consequences of women’s experiences with violence are important next steps.

Violence against women (VAW) by both partners and nonpartners has serious long- and short-term physical and mental health sequelae. Physical and sexual assaults may result in fatal and nonfatal injuries, trauma-specific and generalized pain, unwanted pregnancies, sexually transmitted infections, and gynecologic problems (Brokaw et al 2002, Coker et al 2002, Greenfeld 1998, Hathaway et al 2000, Holmes et al 1998, Pastore and Maguire N.D., Tjaden and Thoennes 2000) Research has linked experiencing violence with gastrointestinal disorders, anxiety, depression, substance use, and suicide attempts (Bergman and Brismar 1991, Caetano et al 2000, Drossman et al 1995, Holmes et al 1998, Humphreys et al 2001, Kilpatrick et al 1997, Kilpatrick et al 1997, Wyshak and Modest 1996). Victimized women also view themselves as being less healthy, having more physical complaints and symptoms of emotional distress, and report lower levels of physical and mental well-being than women who have not been victimized (Brokaw et al 2002, Dickinson et al 1999, Hathaway et al 2000).

As with physical and sexual assault, stalking, defined most often as repeated nonconsensual harassing or threatening behavior (U.S. Department of Justice, Violence Against Women Grants Office, 1998), has a negative impact on women’s health and well-being. A national survey revealed that in 1996 30% of women stalked during the previous year had sought counseling, and 26% had missed time from work owing to their victimization (Tjaden & Thoennes, 2000). Stalking often accompanies intimate partner violence (IPV), and is associated with threats, and often physical and sexual assault (Hall 1998, Pathe 1997).

In addition to the impact on individual victims, VAW exacts an enormous social toll in the form of medical costs, decreased work productivity, homelessness, and an increased burden on the health, criminal justice, and human service systems. Miller, Cohen, and Wiersema (1996) estimated the tangible (property damage and loss, medical care, victims’ services, and police and fire services) and intangible costs (pain and suffering, reduced quality of life) of various crimes, and estimated that domestic violence and sexual assault resulted in $67 and $127 billion in annual losses, respectively, measured in 1993 dollars. The Centers for Disease Control and Prevention used data from the National Violence Against Women survey (NVAW) and estimated that the direct medical and mental health care services related to IPV exceeded $4.1 billion annually (National Center for Injury Prevention and Control, 2003).

Despite the significance of VAW as a public health problem, our knowledge about its prevalence and patterns is limited. Currently there are no permanent national or statewide surveillance systems devoted to VAW, so prevalence estimates must be gleaned from other sources, such as the National Crime Victimization Survey (NCVS; Rennison, 2003). With the exception of the National Family Violence Surveys conducted in 1975 and 1985, and the NVAWS, conducted in 1996 (Straus et al 1990, Straus et al 1996, Tjaden and Thoennes 2000), research has largely been limited to convenience samples, and has used variable definitions of violence, hindering our ability to fully characterize this important problem nationally and among subgroups of women. Prevalence estimates vary widely, due to differences in the types of violence included, samples considered, and methodologies employed.

Although women of all ages experience violence, young women, particularly those <25, have been found to be at greater risk of all types of violent victimization than older women (Coker et al 2000, Straus et al 1990, Straus et al 1996, Tjaden and Thoennes 2000, Verhoek-Oftedahl and Babcock 2000). Early research suggested a higher rate of violence among African American women and Latinas; however, more recent studies have found no difference in the prevalence of IPV and other forms of VAW between whites and minority women, after controlling for socioeconomic factors (Coker et al 2000, Harwell and Spence 2000, Hathaway et al 2000, Straus et al 1990, Tjaden and Thoennes 2000, Verhoek-Oftedahl and Babcock 2000). Previous research indicates that a history of victimization is a strong predictor of future physical and sexual assault, even after controlling for demographic variables (Acierno et al 1997, Coker et al 2000, Harwell and Spence 2000, Hathaway et al 2000, Kilpatrick et al 1997). Small sample sizes have often prohibited examining subpopulations and controlling for multiple covariates of women’s experiences with violence.

This study adds to our knowledge about violence in women’s lives by describing the prevalence and patterns of a variety of types of violence in a population-based national sample of adult women. Our purposes were to 1) describe the nature of women’s experiences with violence within their adult lifetimes and during a 12-month period; 2) measure adult lifetime and prior- year prevalence of women’s experiences with violence; 3) describe how these 2 measures vary across groups of women; and 4) examine associations between selected sociodemographic variables and women’s experiences with violence. Our purpose was to better understand the prevalence and patterns of women’s experiences with violence to identify potential avenues of identification and intervention.

Back to Article Outline

Methods 

Study Design 

In this research, we used qualitative and quantitative methods in a sequential, complementary manner. First, we conducted 13 focus group interviews to inform the content and suggest language for a telephone survey instrument. Then we used a random digit dial telephone interviewing procedure to provide a general probability sample of all households with telephones, including homes with unlisted numbers. We restricted eligibility to noninstitutionalized, English-speaking females at least 18 years old and living in the 48 contiguous U.S. states. Women in college or boarding school dormitories, military barracks, or other group home settings were included in the sample. If a household contained multiple eligible women, 1 was randomly selected, using a computer algorithm.

Instrument Development 

In the focus groups, we sought to include women with a wide range of sociodemographic characteristics and experiences with violence. We also wanted to make sure that certain groups of women who may have had experiences with violence, including battered women, women who lived in high crime, urban areas, lesbians and older women, were included in our focus group sample. Consequently, we used purposeful sampling, employing both “intensity” (recruiting information-rich cases) and “maximum variation” (recruiting participants representing a wide range of variation on dimensions of interest) strategies to recruit participants (Patton, 2001). Homogeneity is particularly important to ensure the comfort of the participants and to facilitate lively group discussion. Thus, although we recruited focus group participants according to a variety of criteria of interest, the groups themselves were homogeneous.

One hundred eleven women participated in 13 focus groups in North Carolina (n = 7 groups), Iowa (n = 2 groups), Washington state (n = 2 groups), and Maryland (n = 2 groups). We pilot tested the resulting survey instrument on a population-based sample of 143 North Carolina women, and revised the wording and format based on the results of the pilot study. We administered the survey to a national sample of 1,800 women between August and December 1997.

Variable Definitions 

We used the results of the focus group to guide the content of the interview questions, including those about respondents’ experiences with violence. Respondents were queried about different types of violent experiences they may have had since age 18, including being followed by a man in a way that frightened them; repeatedly contacted by someone after telling them to stop; physically attacked by a stranger; physically assaulted by a current or former intimate partner (asked only of women who had ever been married or cohabitated with a male or female intimate partner); raped or sexually assaulted (asked as “forced to have sex and/or do sexual things”) by a stranger; and raped or sexually assaulted by someone they knew. Women who had experienced violence were asked follow-up questions about the timing and frequency of the incident(s), as well as the identity of the perpetrator for the most recent incident.

We measured age, ethnicity, household income, educational attainment, and urban–rural status. We also asked women whether they were married or cohabitating with an intimate partner; considered themselves to be a lesbian or bisexual; and received any type of public assistance or lived with anyone who did. We also asked how many adults and children age <18 lived with them.

Data Collection 

Trained female interviewers conducted both the pilot and the national surveys using Computer Assisted Telephone Interviewing (CATI) procedures. They called randomly generated numbers during weekday afternoons and evenings, and on weekends, making a total of 6 attempts before classifying a household as nonresponsive. Potential respondents were informed that their telephone numbers had been randomly generated by a computer, and they were being recruited as part of a national sample of the general adult female population for a women’s health and safety study.

To ensure that respondents would not be endangered by participating in the study, we asked whether it was a safe time for them to talk, and told them they could use a code phrase (“I don’t care to contribute today, thank you”) if they needed to end the interview abruptly (e.g., if an abusive partner entered the room). Interviewers also gave toll-free numbers for the National Domestic Violence Hotline and the Rape, Incest and Abuse National Network to all study participants, regardless of whether they disclosed a history of victimization. Respondents remained anonymous and received a toll-free number they could use to call back at a safe time (for interrupted interviews) or to verify the study’s authenticity.

Data Analyses 

We computed adult lifetime prevalence as the percentage of women who experienced a specific type of violence since age 18, and prior year prevalence as the percentage of women experiencing violence during the 12 months prior to the survey. We used sampling and poststratification weights,^⁎ based on 1990 census data, to correct for unequal probabilities of being selected into the sample and to adjust for sampling bias. We compared our results using weighted versus unweighted data. We calculated rates of violence in the year prior to the survey by dividing the number of experiences in the 12 months before the survey by the number of women in the sample.

Using the χ² test of homogeneity, we compared violent experiences among selected subgroups of respondents. We used logistic regression to identify demographic and experiential factors associated with adult lifetime and prior year experiences with violence, controlling for confounding among covariates. Because 215 women did not supply information about their annual household income, we excluded the income variable from regression analyses. We also omitted cases with missing data for independent and dependent variables, which resulted in sample sizes of 1,767–1,782 for models of nonpartner violence and 1,650 for the model for partner violence. For the partner violence model, we considered only women who reported ever having been married or ever cohabitating with an intimate partner. For models of violent experiences in the previous year, we combined stranger physical and sexual assault because only a small number of women reported each of these types of violence. An α level of .05 was considered statistically significant for all analyses.

Back to Article Outline

Results 

Sociodemographic Characteristics of the Study Participants 

We had a response rate of 73% among eligible households. All the women who began the interviews, which lasted an average of 33 minutes, completed them. As shown in Table 1, the respondents ranged in age from 18–92 years, with a median age of 43. About one fifth (19%) described themselves as belonging to an ethnic minority group, 90% had at least a high school diploma or GED, and nearly half (45%) lived with children <18 years old. The study sample was demographically comparable to that of the March 1997 Current Population Survey, with the exception that women in our sample were, on average, older and more highly educated. Therefore, we used weighted data in reporting our results.

Table 1. Demographic Characteristics of Study Sample

Variable	Study Sample
	n	%
Age (years; n = 1,793)
18–24	149	8.3
25–34	348	19.4
35–44	449	25.0
45–54	315	17.6
55–64	237	13.2
65–74	185	10.3
≥75	110	6.1
Median age	43
Race/ethnicity (n = 1,796)
Caucasian/white	1465	81.6
African American	163	9.1
Biracial	35	1.9
Multiracial	34	1.9
Native American/American Indian	26	1.4
Asian/Asian American	21	1.2
Other	52	2.9
Latina/Hispanic⁎	86	4.8
Marital status (n = 1,750)
Never married	245	14.0
Married	1098	62.7
Separated/divorced	232	13.3
Widowed	175	10.0
Lesbian/bisexual (n = 1,793)	42	2.3
Works outside home†	1212	67.4
Educational attainment (n = 1,800)
Less than high school/GED	154	8.6
High school diploma/GED	548	30.4
More than high school/GED	1098	61.0
Annual household income (n = 1,534)
<$20,000	379	22.8
$20,001–$40,000	448	29.2
$40,001–$60,000	370	24.1
$60,001–$80,000	151	9.8
>$80,000	186	12.12

Women’s Experiences With Violence 

Nearly 60% of the respondents reported experiencing at least 1 form of violence listed in Table 2 since the age of 18. One in 8 (12.0%) of the respondents reported violent experiences in the previous year. Being followed, being repeatedly contacted, and being physically assaulted by a partner were the types of violence women most commonly reported experiencing both in their adult lifetimes and in the previous 12 months.

Table 2. Adult Lifetime and Prior Year Prevalence of Women’s Personal Experiences With Violence (n = 1,800)

	Adult Lifetime Prevalence⁎		Prior Year Prevalence†
	Unweighted Frequency (n)	Weighted Percentage of Women Who Have Experienced (95% CI)	Unweighted Frequency (n)	Weighted Percentage of Women Who Have Experienced (95% CI)
Type of Experience
Followed by a man	772	42.7(40.3–44.8)	91	5.9(4.8–7.0)
Repeatedly contacted	441	25.3(23.3–27.4)	82	5.5(4.4–6.5)
Physical assaulted by
Intimate partner‡ or stranger	477	25.6(23.7–26.4)	61	3.4(2.7–5.1)
Intimate partner‡	399	23.4(21.4–25.5)	45	2.6(1.7–3.4)
Stranger	134	7.5(6.3–8.7)	17	1.0(.57–1.5)
Sexually assaulted by
Known person or stranger	291	16.7(15.2–17.9)	19	1.1(.74– 1.46)
Known person	248	14.3(12.7–15.9)	15	.87(.44–1.3)
Stranger	89	5.6(4.3–6.4)	6	.36(.08–.64)
Any of the above types of experience with violence	1081	59.8(57.5–62.0)	188	12.0(10.4–13.4)

For all types of violence except being followed in a way that frightened them, where strangers were the most common (80%) perpetrators, women were most often victimized by men they knew, frequently their current or former intimate partners. Three times as many women reported physical assaults by partners than assaults by strangers (23.4% versus 7.5% for adult lifetime and 2.6% versus 1.0% for pervious year), and the difference was similar for sexual assaults. Among women who reported repeated unwanted contact, current (15.9%) and former (32.9%) intimate partners were the perpetrators in nearly half of the most recent incidents, followed by strangers (20.2%), acquaintances (19.5%), coworkers (5.1%), and family members (1.6%).

Almost one third (31.5%) of the women had experienced more than one type of violence (e.g., being followed and repeatedly contacted) in their lifetimes, and almost 3% had experienced multiple types of violence in the past year. In addition, women generally experienced multiple incidents of the same type of violence, both during both their adult lifetimes and in the prior year. Thus, the number of victimizations exceeded the number of women victimized for all types of violence (data not shown).

Women’s adult lifetime experiences with violence varied by their sociodemographic characteristics (Table 3). Women <55, especially those 18–34, were more likely to report having had experiences with violence than women ≥55. Caucasian and Asian/Asian American women appeared to be at lower adult lifetime risk of violence than African American, Hispanic, Native American, or biracial or multiracial women, although there was considerable variation among groups. Lesbians and bisexual women reported experiences in significantly higher proportions for most types of violence, with 82% having experienced some type of violence. Adult lifetime experiences with violence did not vary consistently by either income or educational attainment, although living in a household that received public assistance did appear to be associated with higher proportions of all types of violence, except being followed.

Table 3. Women’s Adult Lifetime Experiences With Violence by Sociodemographic Characteristics^⁎

Percent Ever Experienced (Weighted)†
Characteristic	Followed	Repeatedly Contacted	Physical Assault by Stranger	Physical Assault by Partner‡	Sexual Assault by Stranger	Sexual Assault by Known Person	Any Violence
Age (years)	p = .001	p = .0001	p = .0001	p = .0001	p = .0001	p = .0001	p = .0001
18–24	46.4	36.6	10.1	23.8	10.5	21.5	63.5
25–34	50.6	34.8	10.1	27.8	7.5	19.7	71.6
35–44	47.8	26.4	9.9	29.8	6.1	17.1	66.1
45–54	47.7	25.0	9.3	28.8	4.1	13.5	65.0
55–64	35.7	17.7	2.5	18.6	2.3	6.8	50.1
≥65	26.2	9.9	2.1	11.2	1.3	5.3	38.9
Race/ethnicity	p = .32	p = .013	p = .0001¶	p = .002	p = .29¶	p = .006¶	p = .065
Caucasian	41.6	24.7	6.6	22.6	5.2	13.5	58.5
African-American	46.1	25.9	10.4	35.3	6.6	17.5	62.8
Bi-/multiracial	48.2	42.7	23.3	36.9	8.2	38.3	75.3
Native American	40.8	36.5	10.2	22.5	8.9	23.7	66.9
Asian	42.8	7.0	5.5	5.2	0.0	0.0	51.2
Other	54.7	30.5	7.4	22.1	3.6	12.7	73.1
Hispanic§	p = .078	p = .008	p = .006	p = .28	p = .53¶	p = .57	p = .007
Yes	51.6	37.2	16.0	24.8	3.1	16.3	73.3
No	42.2	24.7	7.1	20.9	5.4	14.2	59.0
Lesbian/bisexual	p = .023	p = .008	p = .001	p = .13	p = .0001	p = .0001	p = .0007
Yes	58.3	40.9	29.0	33.4	22.2	47.1	82.4
No	42.3	24.7	6.8	23.0	4.7	13.1	59.0
Public assistance	p = .70	p = .002	p = .001	p = .001	p = .0001	p = .0002	p = .001
Yes	43.9	33.7	16.0	41.8	11.7	22.4	69.5
No	42.6	24.2	6.3	20.9	4.4	13.1	58.4
Children under 18 in household	p = .065	p = .001	p = .024	p = .001	p = .014	p = .020	p = .001
Yes	45.4	30.5	9.1	29.4	6.8	12.6	65.2
No	40.8	21.2	6.3	18.4	4.2	16.4	55.5
Urban–rural	p = .001	p = .95	p = .002	p = .04	p = .112	p = .50	p = .013
City	50.2	24.5	9.1	26.5	6.3	15.0	65.5
Suburb	48.4	25.2	9.3	20.2	5.1	12.9	62.7
Medium town	43.2	26.8	10.5	18.4	7.7	16.5	57.0
Small town	32.1	26.0	4.5	26.9	4.7	14.7	55.9
Rural area	39.3	24.3	4.3	22.6	2.9	11.8	55.9
Household income	p = .23	p = .18	p = .04	p = .006	p = .19	p = .02	p = .13
<$20,000	41.4	27.4	9.8	28.5	6.6	17.2	58.3
$20,001–$40,000	45.1	28.2	8.4	28.1	6.7	17.6	64.0
$40,001–$60,000	43.7	23.3	5.1	20.9	3.5	12.3	59.8
$60,001–$80,000	50.0	19.9	11.3	17.4	4.0	11.9	66.2
>$80,000	46.2	23.2	5.4	19.7	3.2	8.6	61.8
Education	p = .001	p = .206	p = .32	p = .001	p = .65	p = .009	p = .0001
Less than high school/GED	37.5	21.4	9.1	37.4	6.5	18.2	61.0
High school/GED	37.3	21.4	5.8	22.6	4.7	12.0	51.1
More than High School/GED	46.3	26.5	8.0	22.7	4.8	14.0	64.4

The multivariate analyses for adult lifetime violence (Table 4) confirmed many of the observed bivariate relationships. The coefficients reported are the odds ratios and 95% confidence intervals for experiencing versus not experiencing specific types of adult lifetime violence. Specifically, being <55 years old remained a significant risk factor for all types of violence, and living in a household that received public assistance remained significantly associated with all types of violence except being followed. Lesbians and bisexual women also reported higher levels of stranger physical and sexual violence, as well as sexual violence by known assailants. Ethnic minority group status was not a significant risk factor for any type of violence.

Table 4. Adjusted Odds Ratios and 95% Confidence Intervals for Women’s Adult Lifetime Experiences With Violence by Sociodemographic Characteristics

Adjusted Odds Ratio for Victimized Versus Not Victimized in Adult Lifetime (95% CI)
	Followed (n = 1,767)	Repeatedly Contacted (n = 1,778)	Physical Assault by Partnera (n = 1,650)	Physical Assault Stranger (n = 1,782)	Sexual Assault by Stranger (n = 1,782)	Sexual Assault by Known Person (n = 1,782)	Any Violence (n = 1,774)
Age
18–34	2.42(1.78–3.30)	4.54(3.00–6.87)	2.20(1.44–3.35)	3.69(1.61–8.44)	6.14(2.21–17.10)	4.03(2.34–6.97)	2.96(2.21–3.97)
35–54	2.34(1.76–3.27)	2.95(1.94–4.49)	3.54(2.34 –5.35)	4.00(1.75–9.15)	3.82(1.35–10.86)	3.27(1.88–5.68)	2.81(2.11–3.76)
55–64	1.57(1.09–2.27)	1.76(1.07–2.91)	1.69(1.03–2.76)	1.03(.34–3.13)	1.33(.35–5.02)	1.17(.57–2.38)	1.50(1.06–2.11)
>65 (ref)	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Race/ethnicity
African American	1.08(.76–1.53)	.87(.58–1.30)	1.18(.79–1.75)	1.17(.65–2.12)	.71(.33–1.51)	1.11(.70–1.77)	.95(.66–1.37)
Caucasian (ref)	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Otherb	1.26(.84–1.88)	.98(.62–1.53)	.97(.59–1.61)	1.39(.73–2.62)	.76(.30–1.96)	1.03(.60–1.80)	1.14(.74–1.75)
Hispanic	1.02(.63–1.64)	1.46(.88–2.41)	.79(.43–1.45)	1.68(.83–3.39)	.40(.11–1.45)	.85(.44–1.63)	1.29(.76–2.20)
Non-Hispanic (ref)	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Urban/rural status
Urbanc	1.44(1.15–1.80)	.86(.66–1.12)	1.11(.85–1.46)	1.11(.74–1.67)	1.32(.82–2.12)	.98(.71–1.35)	1.29(1.02–1.63)
Other (ref)	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Public assistance
Yes	1.06(.78–1.42)	1.65(1.19–2.28)	2.06(1.49–2.84)	2.84(1.80–4.50)	2.84(1.66–4.88)	1.62(1.11–2.38)	1.58(1.15–2.18)
No (ref)	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Education
<High School/GED	.98(.67–1.41)	.77(.50–1.19)	1.87(1.26–2.78)	1.08(.55–2.12)	.84(.385–1.82)	1.81(1.11–2.88)	1.56(1.06–2.25)
High School/GED	.78(.63–.97)	.96(.75–1.24)	1.29(.99–1.69)	.78(.51–1.21)	1.14(.70–1.87)	(.73–1.38)	.70(.56–.87)
>High School/GED (ref)	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Lesbian or bisexual
Yes	1.50(.85–2.68)	1.46(.82–2.60)	1.10(.57–2.14)	3.97(2.07–7.63)	3.89(1.90–8.00)	4.19(2.35–7.46)	2.28(1.10–4.75)
No (ref)	1.0	1.0	1.0	1.0	1.0	1.0	1.0

Being <55 years old was the only variable consistently associated with victimization in the previous year, with women age 18–24 at the highest risk of experiencing all types of violence (Table 5). Lesbian and bisexual women, and women receiving public assistance, were also more likely than other women to be sexually assaulted by a known assailant.

Table 5. Adjusted Odds Ratios and 95% Confidence Intervals for Women’s Victimization Experiences in Prior Year by Sociodemographic Characteristics

Adjusted Odds Ratio for Victimized vs. Not Victimized in Prior Year (95% CI)
Type of Violence	Age 18–24⁎	Age 25–54†	Minority‡	Urban¶	Public Assistance§	Less than High School/GED	Lesbian or Bisexual
Followed	34.19(13.2–88.4)	5.24(2.04–13.4)	.89(.48–1.37)	1.47(.94–2.30)	.53(.22–1.25)	2.68(1.01–7.16)	1.12(.44–2.85)
Repeatedly contacted	47.8(11.05–206.8)	16.23(3.95–66.70)	NS	NS	NS	NS	NS
Stranger physical or sexual violence	34.84(4.36–278.5)	NS	NS	NS	NS	NS	NS
Physical violence by partner∥	11.29(1.39–91.8)	9.75(1.36–69.8)	.90(.39–2.06)	1.56(.75–3.25)	1.27(.45–3.25)	.59(.20–1.75)	4.02(1.36–11.86)
Sexual assault by known person⁎⁎	NA	NA	1.34(.41–4.38)	.59(.16–2.16)	7.26(2.10–25.12)	1.82(.34–9.64)	9.12(2.77–30.0)
Any violence	32.33(16.1–65.0)	7.14(3.65–14.0)	1.02(.71–1.48)	1.13(.80–1.59)	.91(.55–1.48)	.96(.56–1.65)	1.90(.99–3.63)

Back to Article Outline

Conclusions and Discussion 

This study is one of the few national, population-based studies to examine women’s experiences with violence, and only the second to examine stalking experiences. Our annual prevalence estimates for some types of violence were as much as 6 times higher than those of the 1997 NCVS, possibly because, even though their questions were not limited to incidents reported to law enforcement, the NCVS asks about victimization in the context of a survey about crime. Much of the violence that women experience historically has not been considered aberrant or criminal behavior (Brownmiller 1975, Schecter 1982) and may not have been thought of as “crimes” by NCVS respondents. In addition, our methodological strategies, such as using female interviewers, ensuring respondent anonymity, and providing a code phrase to suspend the interview, may have facilitated women’s disclosure of violent experiences.

Our study was not as comprehensive or lengthy as the NVAWS, yet our findings on VAW prevalence and patterns are remarkably similar. Given the widely varying prevalence estimates for VAW, and the paucity of national population-based surveys on this topic, consistency in findings among population-based research is important to note, to refine our prevalence estimates, and accurately portray patterns and characteristic of subgroups, which in turn, could lead to the development of effective interventions and policies.

Since the mid-1990s, the Centers for Disease Control and Prevention have sought to improve the quality of VAW data by working with researchers, practitioners, and advocates in 5 states to standardize measures and facilitate the development and evaluation of population-based surveillance systems for IPV (Harwell and Spence 2000, Hathaway et al 2000, Saltzman et al 2000, Weinbaum et al 2001) Another strategy has been “surveillance partnerships” in which violence questions are included in ongoing surveillance systems such as the Behavioral Risk Factor Surveillance System and the Pregnancy Risk Factor Monitoring System. Recent findings from these efforts should provide needed guidance (Durant et al 2000, Harwell and Spence 2000, Hathaway et al 2000, Saltzman et al 2000, Vest et al 2002).

Being followed by a man in a way that frightened them and being repeatedly contacted, behaviors that fit within most definitions of stalking, were experiences noted by the largest percentage of respondents. The NVAWS, which defined stalking as behavior(s) that occurred on ≥2 occasions and resulted in reports of feeling frightened or fearing bodily harm, reported that 8.1% of American women had been stalked in their adult lifetimes and 1% had been stalked in the previous 12 months (Hall 1998, Tjaden and Thoennes 2000). However, stalking is neither typically defined as “violence,” nor assessed in crime or victimization surveys. As both the NVAW survey and this study indicate, stalking is an important component of women’s overall experiences with violence. Future research should explore the contexts and consequences of stalking in greater detail.

Because of the serious health consequences of IPV against women, in the past decade there has been a call for health care facilities to develop and implement IPV “screening” protocols for female patients has been endorsed by many public health and health care professional associations (American College of Emergency Physicians 1994, American College of Nurse Midwives 1995, American Medical Association 1992, American Medical Association 1995, American Nurses Association 1991). Despite these recommendations, surveys of health care providers in both acute and primary care settings indicate that fewer than half regularly ask their female patients about IPV, and even fewer ask about other types of violence (Feldhaus et al 2000, Friedman et al 1992, Hamberger et al 1998, Thompson et al 2000). Currently no “gold standard” screening instrument exists, and there has been only limited evaluation of health care-based identification and referral protocols for victims of violence (Chalk et al 1998, Cohn and Stobo 2002, Ramsay et al 2002, Wathen and MacMillan 2003). Thus, although encounters with health care providers may provide excellent opportunities to identify victims of violence, research examining the most effective screening methods, interventions, and referral protocols is sorely needed.

As confirmed in this study, the number and proportion of women who report violence during their adult lifetimes is considerable, yet the National Criminal Victimization Survey, and most past research, inquire only about events during the previous 12-month period. Numerous studies suggest that the effects of violent victimization are profound and enduring (Coker et al 2000, Drossman et al 1995, Kilpatrick et al 1997, Norris and Kaniasty 1997). Therefore, longitudinal research that details the “natural history” of violent experiences, including the contexts in which they occur and the nature and duration of their impact, will provide an empirical basis for a fuller understanding of the scope and long-term impact of VAW.

We found that women <35 years old were significantly more likely to report experiencing violence in their adult lifetimes and within the past year than women ≥35. Although a higher proportion of young women might be expected to report violent experiences in the previous year than older women, the finding that adult lifetime prevalence was also higher for women 18–34 may reflect the influence of experiences unique to women in different age groups (cohort effect), recall bias, or a reluctance to disclose past violent experiences among older women.

We found that women who live in households that receive some form of public assistance were more likely to report adult lifetime experience with all types of violence and sexual assault by a known perpetrator (most often a current or former partner) in the previous year, even after controlling for covariates. Because they have fewer resources, perhaps women who receive public assistance are potentially less able to leave violent relationships, quit abusive work situations, or move to safer neighborhoods.

In multivariate analyses, the differences in victimization between white and minority women was not significant. However, bivariate analyses revealed considerable variation among women of different ethnic backgrounds. The uniformly low rates among Asian women we found are consistent with findings from the NVAWS and the National Criminal Victimization Survey. These results are particularly intriguing, given that in our study over half the Asian women were under age 25, and hence theoretically at higher risk. Some researchers have suggested that traditional cultural values in Asian communities may discourage women from disclosing violent acts, especially those committed by partners (Bui and Merry 1999, Ho 1990, Tjaden and Thoennes 2000, Yoshihama 1999). However, the small number of both Asian women and their personal experiences with violence in the study sample limit our ability to examine this finding further. Also consistent with the NVAWS are our findings that higher proportions of biracial or multiracial women reported experiencing all types of violence, and that more Native American women reported physical and sexual assault by strangers and sexual assault by a known person (Tjaden & Thoennes, 2000). Again, sample size precluded more in-depth analyses of these groups’ experiences.

Lesbians and bisexual women reported high levels of violence, and sexual orientation remained a risk factor for some types of adult lifetime victimization after controlling for covariates. However, these groups’ small numbers in our sample, and our inability to establish the temporal sequence of women’s violent experiences necessitates cautious interpretation. Currently, little is known about unique victimization risk factors for lesbians and bisexual women. This area is ripe for future research, including information about whether these women are the targets of antigay violence.

Study Limitations 

Although our survey allowed us to examine women’s experiences with violence in some depth, our study’s generalizability is limited by several factors. First, our sample was restricted to English-speaking adult women who lived in homes with telephones. However, we used poststratification weights to adjust for sampling bias and any unequal probability of being selected into the sample. In addition, the cross-sectional nature of the survey limited our ability to discern the temporal relationships among variables; for example, whether receiving public assistance put women at higher risk of violence, or whether experiencing violence preceded women’s needing public assistance.

Our measures of violence were limited; we did not ask about all forms of violence (e.g., child abuse, elder abuse by a caregiver), nor were our variables explicitly behaviorally defined. Thus, we have likely underestimated women’s experiences. We also relied on women’s self-reports of violent experiences, which may have been subject to recall bias.

Conclusion and Implications 

To devise appropriate prevention and intervention strategies, a clear understanding of the patterns, context, chronology, and consequences of VAW is necessary. Future studies should rely on longitudinal designs, ensure that adequate numbers of women from diverse ethnic backgrounds are included, and collect information on stalking behaviors.

Back to Article Outline

Acknowledgments 

The authors thank Thomas B. Cole, Brenda DeVellis, and Sandra L. Martin for their assistance reviewing drafts of the manuscript, and Lisa Dulli and Michele Decker for their help in editing and preparing the manuscript for publication.

Supported by grants from the National Center for Injury Prevention and Control (R49/CCR4105-01 and R49-CCR402444-04-13).

Back to Article Outline

References 

1. Acierno R, Resnick HS, Kilpatrick DG. Health impact of interpersonal violence. 1: Prevalence rates, case identification, and risk factors for sexual assault, physical assault, and domestic violence in men and women. Behavioral Medicine. 1997;23(2):53–64
   • View In Article
2. American College of Emergency Physicians. Emergency medicine and domestic violence. Irving, TX: Author; 1994;
   • View In Article
3. American College of Nurse Midwives. Position statement on violence against women. Silver Spring, MD: Author; 1995;
   • View In Article
4. American Medical Association. Diagnostic and treatment guidelines on domestic violence. Archives of Family Medicine. 1992;1:39–47
   • View In Article
   • MEDLINE
5. American Medical Association. Strategies for the treatment and prevention of sexual violence. Chicago: Author; 1995;
   • View In Article
6. American Nurses Association. Position statement on physical violence against women. Washington, D.C: Author; 1991;
   • View In Article
7. Bergman B, Brismar B. A 5-year follow-up study of 117 battered women. American Journal of Public Health. 1991;81:1486–1489
   • View In Article
   • MEDLINE
   • CrossRef
8. Brokaw J, Fullerton-Gleason L, Olson L, Crandall C, McLaughlin S, Sklar D. Health status and intimate partner violence: A cross-sectional study. Annals of Emergency Medicines. 2002;39:31–38
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (91 KB)
   • CrossRef
9. Brownmiller S. Against our will: Men, women, and rape. New York: Bantam Books; 1975;
   • View In Article
10. Bui HN, Morash M. Domestic violence in the Vietnamese immigrant community: An exploratory study. Violence Against Women. 1999;5:769–795
    • View In Article
    • CrossRef
11. Caetano R, Cunradi CB, Clark CL, Schafer J. Intimate partner violence and drinking patterns among white, black, and Hispanic couples in the U.S. Journal of Substance Abuse. 2000;11:123–138
    • View In Article
    • MEDLINE
    • CrossRef
12. Chalk RA, King PA National Research Council (U.S.) Committee on the Assessment of Family Violence Interventions. Violence in families: assessing prevention and treatment programs. Washington, D.C: National Academy Press; 1998;
    • View In Article
13. In:  Cohn F,  Salmon ME,  Stobo JD editor. Confronting chronic neglect: The education and training of health professionals on family violence. Washington, D.C: National Academy Press; 2002;
    • View In Article
14. Coker AL, Davis KE, Arias I, Desai S, Sanderson M, Brandt HM, et al. Physical and mental health effects of intimate partner violence for men and women. American Journal of Preventive Medicine. 2002;23:260–268
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (132 KB)
    • CrossRef
15. Coker AL, Smith PH, Bethea L, King MR, McKeown RE. Physical health consequences of physical and psychological intimate partner violence. Archives of Family Medicine. 2000;9(5):451–457
    • View In Article
    • MEDLINE
    • CrossRef
16. Coker AL, Smith PH, McKeown RE, King MJ. Frequency and correlates of intimate partner violence by type: Physical, sexual, and psychological battering. American Journal of Public Health. 2000;90:553–559
    • View In Article
    • MEDLINE
    • CrossRef
17. Dickinson LM, deGruy FV, Dickinson WP, Candib LM. Health-related quality of life and symptom profiles of female survivors of sexual abuse. Archives of Family Medicine. 1999;8:35–43
    • View In Article
    • MEDLINE
    • CrossRef
18. Drossman DA, Talley NJ, Leserman J, Olden KW, Barreiro MA. Sexual and physical abuse and gastrointestinal illness (Review and recommendations). Annals of Internal Medicine. 1995;123:782–794
    • View In Article
    • MEDLINE
19. Durant T, Colley Gilbert B, Saltzman LE, Johnson CH. Opportunities for intervention: Discussing physical abuse during prenatal care visits. American Journal of Preventive Medicine. 2000;19:238–244
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (126 KB)
    • CrossRef
20. Feldhaus KM, Houry D, Kaminsky R. Lifetime sexual assault prevalence rates and reporting practices in an emergency department population. Annals of Emergency Medicines. 2000;36:23–27
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (63 KB)
    • CrossRef
21. Friedman LS, Samet JH, Roberts MS, Hudlin M, Hans P. Inquiry about victimization experiences (A survey of patient preferences and physician practices). Archives of Internal Medicine. 1992;152:1186–1190
    • View In Article
    • MEDLINE
22. Greenfeld LA, Rand MR, Craven D, Klaus PA, Perkins CA, Ringel C, et al. Violence by intimates: Analysis of data on crimes by current or former spouses, boyfriends and girlfriends. Washington, D.C: Bureau of Justice Statistics; 1998;[Publication NCJ 167237].
    • View In Article
23. Hall DM. The victims of stalking. In:  M JR editors. The psychology of stalking. San Diego: Academic Press; 1998;p. 113–137
    • View In Article
24. Hamberger LK, Ambuel B, Marbella A, Donze J. Physician interaction with battered women: the women’s perspective. Archives of Family Medicine. 1998;7:575–582
    • View In Article
    • MEDLINE
    • CrossRef
25. Harwell TS, Spence MR. Population surveillance for physical violence among adult men and women, Montana 1998. American Journal of Preventive Medicine. 2000;19:321–324
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (135 KB)
    • CrossRef
26. Hathaway JE, Mucci LA, Silverman JG, Brooks DR, Mathews R, Pavlos CA. Health status and health care use of Massachusetts women reporting partner abuse. American Journal of Preventive Medicine. 2000;19:302–307
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (269 KB)
    • CrossRef
27. Ho C. An analysis of domestic violence in Asian American communities: A multicultural approach to counseling. Women and Therapy. 1990;9:129–150
    • View In Article
28. Holmes MM, Resnick HS, Frampton D. Follow-up of sexual assault victims. American Journal of Obstetrics and Gynecology. 1998;179:336–342
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (111 KB)
    • CrossRef
29. Humphreys J, Lee K, Neylan T, Marmar C. Psychological and physical distress of sheltered battered women. Health Care for Women International. 2001;22:401–414
    • View In Article
    • MEDLINE
    • CrossRef
30. Kilpatrick DG, Acierno R, Resnick HS, Saunders BE, Best CL. A 2-year longitudinal analysis of the relationships between violent assault and substance use in women. Journal of Consulting and Clinical Psychology. 1997;65:834–847
    • View In Article
    • MEDLINE
    • CrossRef
31. Kilpatrick DG, Resnick HS, Acierno R. Health impact of interpersonal violence (3: Implications for clinical practice and public policy). Behavioral Medicine. 1997;23(2):79–85
    • View In Article
32. Miller TR, Cohen MA, Wiersema B. Victim costs and consequences: A new look. Washington, D.C: National Institute of Justice; 1996;
    • View In Article
33. National Center for Injury Prevention and Control. Costs of intimate partner violence against women in the United States. Atlanta, GA: Centers for Disease Control and Prevention; 2003;
    • View In Article
34. Norris FH, Kaniasty TM. The psychological consequences of crime: Findings from a longitudinal, population-based study. In:  Davis LA,  Skogan RC editor. Victims of Crime. Thousand Oaks, CA: Sage; 1997;p. 146–166
    • View In Article
35. Pastore, A. L., & Maguire, K. (N.D.) Sourcebook of criminal justice statistics. Available: from www.albany.edu/sourcebook. Accessed February 2, 2004.
    • View In Article
36. Pathe M. The impact of stalkers on their victims. British Journal of Psychiatry. 1997;170:12–17
    • View In Article
37. Patton MQ. In: Qualitative Research and Evaluation Methods. 3rd Ed.. Thousand Oaks, CA: Sage; 2001;
    • View In Article
38. Ramsay J, Richardson J, Carter YH, Davidson LL, Feder G. Should health professionals screen women for domestic violence? (Systematic review). BMJ. 2002;325(7359):314
    • View In Article
    • CrossRef
39. Rennison CM. Intimate partner violence, 1993–2001. Washington, D.C: Bureau of Justice Statistics; 2003;
    • View In Article
40. Saltzman LE, Green YT, Marks JS, Thacker SB. Violence against women as a public health issue. American Journal of Preventive Medicine. 2000;19:325–329
    • View In Article
    • Full Text
    • Full-Text PDF (137 KB)
    • CrossRef
41. Schecter S. Women and male violence: The vision and struggles of the battered women’s movement. Boston: South End Press; 1982;
    • View In Article
42. Straus MA, Gelles RJ, Smith C. Physical violence in American families: Risk factors and adaptations to violence in 8,145 families. New Brunswick, NJ: Transaction Publishers; 1990;
    • View In Article
43. Straus MA, Hamby SL, Boney-McCoy S, Sugarman DB. The Revised Conflict Tactics Scales (CTS-2). Journal of Family Issues. 1996;17:283–316
    • View In Article
44. Thompson RS, Rivara FP, Thompson DC, Barlow WE, Sugg NK, Maiuro RD, et al. Identification and management of domestic violence: A randomized trial. American Journal of Preventive Medicine. 2000;19:253–263
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (413 KB)
    • CrossRef
45. Tjaden P, Thoennes N. Full report of the prevalence, incidence and consequences of violence against women: Findings from the National Violence Against Women Survey. Washington, D.C: National Institute of Justice; 2000;
    • View In Article
46. U.S. Department of JusticeViolence Against Women Grants OfficeO. o. J. P.. Stalking and domestic violence: The third annual report to Congress under the Violence Against Women Act (No. 172204). Washington, D.C: Author; 1998;
    • View In Article
47. Verhoek-Oftedahl W, Babcock JC. Improving surveillance of intimate partner violence by use of multiple data sources. American Journal of Preventive Medicine. 2000;19:308–315
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (934 KB)
    • CrossRef
48. Vest JR, Catlin TK, Chen JJ, Brownson RC. Multistate analysis of factors associated with intimate partner violence. American Journal of Preventive Medicine. 2002;22:156–164
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (75 KB)
    • CrossRef
49. Wathen CN, MacMillan HL. Interventions for violence against women: Scientific review. Journal of the American Medical Association. 2003;289:589–600
    • View In Article
50. Weinbaum ST, Chavez G, Motylewski C, Barrera N, Courtney JG California Department of Human Services Women’s Health Survey Group. Female victims of intimate partner physical domestic violence (IPP-DV), California, 1998. American Journal of Preventive Medicine. 2001;21:313–319
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (79 KB)
    • CrossRef
51. Wyshak G, Modest GA. Violence, mental health, and substance abuse in patients who are seen in primary care settings. Archives of Family Medicine. 1996;5:441–447
    • View In Article
    • MEDLINE
52. Yoshihama M. Domestic violence against women of Japanese descent in Los Angeles: Two methods of estimating prevalences. Violence Against Women. 1999;5:869–897
    • View In Article
    • CrossRef

• ⁎ To calculate weights, the respondents were categorized into subgroups by their race, census region, and age. The respondent’s race and age were questions in the survey, and their census region was obtained by matching up their state with a map of census regions. For those cases where respondents did not provide their race or age (n = 11), values were imputed. Race was defined as white versus nonwhite. Census regions are Northeast, Midwest, South, and West, as defined by the U.S. Census. Age was defined by these 5 groups: 18–24 years old, 25–34 years old, 35–44 years old, 45–54 years old, and ≥55 years old. For subgroups where the sample size was too small, age categories were collapsed. In all of these cases, the race was nonwhite. These resulted in the following changes to the subgroups: Northeast, 18–34 years old and ≥35 years old; Midwest, 18–34 years old, 35–44 years old, and ≥45; and West, 18–34 years old, 35–44 years old, and ≥45 years old. This resulted in 33 subgroups for the weight calculation. First, the unadjusted weight (W_hi) for the i-th respondent in the h-th adjustment cell was calculated by dividing the number of eligible women in the household in which the hi-th respondent was found by the number of phone lines reaching the household in which the hi-th respondent was found. This information was obtained from the respondent by the questions PREINTRO and H78, respectively. Then, the poststratification ratio adjustment (A_h) for the h-th cell was calculated by dividing the proportion of the female population ≥18 years old in the adjustment cell by the sample weighted proportion of female respondents in the adjustment cell. The 1990 United States Census was used to find the proportion of the female population ≥18 years old in the adjustment cell. The final adjusted weight (W*_hi) is A_h multiplied by W_hi.