Effect of Face-to-Face Interview Versus Computer-Assisted Self-Interview on Disclosure of Intimate Partner Violence Among African American Women in WIC Clinics

Abstract

African American women in the United States report intimate partner violence (IPV) more often than the general population of women. Overall, women underreport IPV because of shame, embarrassment, fear of retribution, or low expectation of legal support. African American women may be especially unlikely to report IPV because of poverty, low social support, and past experiences of discrimination. The purpose of this article is to determine the context in which low-income African American women disclose IPV. Consenting African American women receiving Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) services in WIC clinics were randomized to complete an IPV screening (Revised Conflict Tactics Scales–Short Form) via computer-assisted self-interview (CASI) or face-to-face interview (FTFI). Women (n = 368) reported high rates of lifetime and prior-year verbal (48%, 34%), physical (12%, 7%), sexual (10%, 7%), and any (49%, 36%) IPV, as well as IPV-related injury (13%, 7%). Mode of screening, but not interviewer race, affected disclosure. Women screened via FTFI reported significantly more lifetime and prior-year negotiation (adjusted odds ratio [aOR] = 10.54, 3.97) and more prior-year verbal (aOR = 2.10), sexual (aOR = 4.31), and any (aOR = 2.02) IPV than CASI-screened women. African American women in a WIC setting disclosed IPV more often in face-to-face than computer screening, and race-matching of client and interviewer did not affect disclosure. Findings highlight the potential value of face-to-face screening to identify women at risk of IPV. Programs should weigh the costs and benefits of training staff versus using computer-based technologies to screen for IPV in WIC settings.

Keywords

African American women intimate partner violence screening response effects

Introduction

Each year in the United States, approximately 4.7 million women report physical violence by an intimate partner (Black et al., 2011). Overall, 36% of U.S. women report rape, physical violence, or stalking by an intimate partner during their lifetime (Black et al., 2011). Compared with other racial and ethnic groups, African American women are disproportionately more likely to experience intimate partner violence (IPV). In the United States, more African American (44%) than European American (35%) or Hispanic (37%) women report IPV during their lifetime, and African American women experience recurrent IPV at a rate 6 times higher than that of European American women (Black et al., 2011).

IPV predicts poor health outcomes and high health care costs for women and their children (Rivara et al., 2007; Sobkoviak, Yount, & Halim, 2012; Yount, DiGirolamo, & Ramakrishnan, 2011). The physical consequences of IPV range from minor injuries such as lacerations and contusions to more severe injuries, disability, and death (Black et al., 2011; Tjaden & Thoennes, 2000). Compared with women with no history of IPV, those who experience IPV are at an increased risk for gastrointestinal problems, chronic pain, sexually transmitted infections and vaginal bleeding, gynecological or pregnancy complications, and are more likely to engage in high-risk behaviors, such as substance abuse or smoking (Hankin, Smith, Daugherty, & Houry, 2010; Kramer, Lorenzon, & Mueller, 2004; Yount et al., 2011). IPV is also associated with lower mental health functioning (Plichta, 2004). Specifically among low-income African American women, IPV is associated with depression, posttraumatic stress symptoms, and suicidal ideation (Houry, Kaslow, & Thompson, 2005; Leiner, Compton, Houry, & Kaslow, 2008; Yount et al., 2011).

Accurate and reliable estimates of IPV perpetration and victimization are elusive because of the sensitivity of the questions (Yount & Li, 2012). IPV victimization often is underreported because women feel shame, embarrassment, fear retribution from the perpetrator, or do not expect to receive legal support (Ellsberg, Heise, Pena, Agurto, & Winkvist, 2001). African American women may be especially unlikely to report IPV or seek assistance because of poverty, low social support, or expectations of discrimination and mistreatment by service providers (Taft, Bryant-Davis, Woodward, Tillman, & Torres, 2009). Furthermore, IPV disclosure is sensitive to methodological factors, including the nature, wording, and length of the question as well as the context of the interview including privacy, interviewer skill, and opportunities to disclose (Ellsberg et al., 2001). Despite the existence of many IPV screening tools and research on their implementation (see Rabin, Jennings, Campbell, & Bair-Merritt, 2009, for review), the modes of screening for IPV are still understudied as well as procedures for follow-up after screening (Kottenstette & Stulberg, 2013). However, it is clear that the consistency and effectiveness of those who screen and methods of screening are key (Amar, Laughon, Sharps, Campbell, & Expert Panel on Violence, 2013).

IPV screening does not increase the likelihood of abuse, but a positive IPV screen does predict future victimization (Houry et al., 2004, 2008). Thus, screening is an important first step in offering assistance to women experiencing violence. Clinics that provide Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) services provide a unique opportunity to screen for IPV among women who are potentially unaware of available services or who may not otherwise seek IPV-related services. WIC clinics also serve a high percentage of low-income African American women, a population particularly at risk for IPV victimization (Cunradi, Caetano, & Schafer, 2002). These clinics also are an important, and often the only, point-of-contact between low-income African American women and the health care system. Thus, incorporating IPV screening into WIC intake procedures provides an opportunity to identify women experiencing IPV and to offer information, resources, and services.

Both interviewer–client race-matching and computer-assisted self-interview (CASI) show promise in maximizing disclosure of sensitive information although their success in eliciting disclosure of IPV has not been addressed. Social science research conducted in the 1970s suggests respondents are more open and frank with interviewers of their own race (Hatchett & Schuman, 1975-1976). More recently, research regarding disclosure of sensitive information suggests that interviewer characteristics, especially interviewer appearance, can influence participant survey responses (Dailey & Claus, 2001; Weeks & Moore, 1981). An integral part of interviewer appearance, the interviewer’s race, influenced participants’ responses in face-to-face, telephone, and self-administered surveys, and the effects were more dramatic when survey items asked about sensitive topics, including physical abuse and substance use. Participants responded in ways that might have been perceived more positively by a person of the interviewer’s race (Davis, Couper, Janz, Caldwell, & Resnicow, 2010). Effects of interviewer race may be particularly relevant to our study, which takes place in the southern United States. Research suggests African Americans in the South are more likely to disclose sensitive and potentially stigmatizing behaviors to African American interviewers compared with European American interviewers (Livert, Kadushin, Schulman, & Weiss, 1998).

CASI screening may influence disclosure of sensitive information although the effect of CASI on disclosure of IPV victimization is understudied. CASIs have been shown to increase disclosure of stigmatized behaviors, including HIV-status, history of tuberculosis, and nonprescription methadone use, in comparison with face-to-face administration of the same questionnaire (Newman et al., 2002). In addition, this technology seems ideally suited for the collection of sensitive data because of added privacy and anonymity (Rosenbaum, Rabenhorst, Reddy, Fleming, & Howells, 2006). CASI administration has elicited higher reported rates of illegal drug use and sexual behavior in comparison with traditional, face-to-face interview (FTFI) techniques (Turner et al., 1998).

In contrast, face-to-face interviewing elicits more frequent reporting of psychological distress, including feelings of hopelessness, worry, or depression (Newman et al., 2002). Other studies found no difference in disclosure of risk behaviors for transmission of HIV or gynecological history between CASI and FTFI (Hasley, 1995; Sanders et al., 1994).

More research is needed to investigate the contexts in which low-income African American women disclose IPV victimization. Our study explores the effect of race-matched versus non-race-matched interviews and CASI versus FTFI on disclosure of IPV victimization among African American clients of two WIC clinics in greater metropolitan Atlanta, Georgia.

Method

Study Design

Between July 17, 2012, and September 21, 2012, we conducted a cross-sectional survey of self-reported IPV victimization in African American women receiving WIC services. Questionnaires were administered via FTFI or CASI with either an African American or European American interviewer. We computed the prevalence of IPV victimization overall and by type in the general WIC population and compared levels of disclosure across interview mode and interviewer race. Telephone follow-up interviews were conducted with study participants at 2 weeks to ask about their experience with and preference for screening method.

Setting

The study took place in two WIC clinics (hereafter called Clinic 1 and Clinic 2) located in metropolitan Atlanta, Georgia. Metropolitan Atlanta has a population of more than 5.3 million people and is the ninth largest metropolitan area in the United States (U.S. Census Bureau, 2010). The city of Atlanta has a population of about 432,000, just more than half (54%) of whom are African American (U.S. Census Bureau, 2010). Nineteen percent of families in the city of Atlanta live below the federal poverty line, which is defined as a yearly income of US$23,050 or less for a family of four (Department of Health and Human Services, 2012), and 26% of these families have children younger than 5 years of age. Almost half (45.7%) of families within the city of Atlanta with children younger than 18 years old receive food stamps, and a large majority of the families receiving food stamps are African American (91.7%; U.S. Census Bureau, 2010).

WIC is funded by the U.S. Department of Agriculture, which extends grants to states to provide supplemental food, health care referrals, and nutrition education to low-income women, infants, and children up to 5 years of age (Martinez-Schiferl, Zedlewski, & Giannarelli, 2013). To be eligible for WIC, families must fall below 185% of the U.S. Poverty Income Guidelines (Martinez-Schiferl et al., 2013). WIC clinics are a common health care contact for low-income women and children, and thus an important opportunity to identify at-risk clients and to give them resources.

In this study, Clinic 1 serves, on average, 436 women per month, 91% of whom are African American. Clinic 2 serves, on average, 709 women per month, 80% of whom are African American. The Emory University Institutional Review Board and the Division of Health and Wellness of the health department of the county in which the study was conducted approved the study.

Sample Inclusion and Exclusion Criteria and Power

To be enrolled in the study, a woman had to be at least 18 years old, eligible to receive WIC services, English speaking, and literate. Based on the average monthly caseload in each clinic, we estimated that we could recruit 700 total participants during the study period. To make the sample proportional to monthly clinic volume, the total desired sample size was adjusted to 704, with 256 participants to be recruited from Clinic 1 and 448 participants to be recruited from Clinic 2. Assuming a two-sided Type I error rate of 0.05, 80% power, and a 40% reported lifetime prevalence of IPV victimization in the FTFI group (Black et al., 2011), our study was able to detect an effect if reported lifetime prevalence of IPV victimization was less than 30% or greater than 51% in the CASI group.

Instruments

Study instruments included brief modules on demographics, general health behaviors, tobacco use, alcohol use (TWEAK: Tolerance, Worried, Eye-opener, Amnesia, Cut down; Russell, 1994), substance use (Drug Abuse Screening Test; Skinner, 1982), contraceptive use, and women’s lifetime and prior-year IPV victimization. Questions from the Revised Conflict Tactics Scales–Short Form (CTS2S) were used to screen for IPV victimization. The CTS2S asks 10 questions, two each in the five domains of negotiation, psychological IPV, physical IPV, sexual IPV, and IPV-related injury. The negotiation items included a woman’s partner (a) explaining an opinion or suggesting a compromise and (b) showing respect or caring for her feelings during a disagreement. Psychological IPV items included a woman’s partner (a) insulting, shouting, swearing, or yelling at her and (b) threatening to hit her or destroying her belongings. Physical IPV items included a woman’s partner (a) pushing, shoving, or slapping her and (b) punching, kicking, or beating her up. Sexual IPV items included a woman’s partner (a) using force (hitting, holding down, using a weapon) to make her have sex and (b) insisting on sex, or insisting on sex without a condom, without using physical force. IPV-related injury included a woman (a) having a sprain, bruise, small cut, or feeling pain after a fight with her partner and (b) needing to see a doctor after a fight with her partner (Straus & Douglas, 2004). Each of the 10 items was measured on a 7-point scale, ranging from 2 (once in the past year) to 7 (more than 20 times in the past year) and includes choices for “not in the past year, but this has happened before” (1) and “this has never happened” (0). Thus, each item captured the frequency of specific acts and behaviors, rather than attitudes or causes of violence (Straus & Douglas, 2004). Each of the five domains was coded dichotomously (yes, no [reference]) based on women’s responses to domain-specific CTS2S items. The CTS2S has good construct validity with the full 78-item Revised Conflict Tactics Scale (CTS2; Straus, Hamby, Boney-McCoy, & Sugarman, 1996), with item correlations between the CTS2 and CTS2S for IPV victimization ranging from 0.65 to 0.94 (Straus & Douglas, 2004). The Cronbach’s alpha for internal consistency of the CTS2S in our study sample was 0.83.

Data Collection

Interviewers were trained on survey administration, including the study introduction, and completed sensitivity training for administering questionnaires related to IPV (Watts, Heise, Ellsberg, & Moreno, 2001). Interviewers were present in the WIC clinic waiting rooms from 8:30 a.m. to 5:00 p.m. during the study period to identify and approach potential participants. Interviewers identified potential participants by reviewing the clinic registration log. All potentially eligible participants were approached and taken to a private room to obtain informed consent. Potential participants were informed that the survey asked questions about their general health and about their relationship with their partner.

Participants could stop the interview if they were called to receive their WIC services or if they chose not to finish. All participants who completed a FTFI received a brochure with resources related to WIC, child health insurance, healthy relationships, and contact information for local domestic violence agencies and hotlines. All participants who completed a CASI received an equivalent printed list of resources for any health risk behaviors they disclosed.

By the end of the study period, 1,137 women had been approached before the desired sample of 704 women was reached, for an overall participation rate of 61.9%. The majority (90.8%) of study participants were African American. As race-matched interviews and non-race-matched interviews took place only at Clinic 2, only African American women interviewed at Clinic 2 were included in the present analysis (n = 402). Of the participants who self-identified as African American, 368 (91.5%) had complete data for variables of interest and were included in the analysis. The women who did not complete the full survey were similar to women who completed the full survey with respect to education, relationship status, employment, age, and number of children (results available on request). One hundred seventeen women (31.8%) completed the follow-up survey 2 weeks after their interview.

Characteristics of the Sample

The mean age of respondents was 27.4 years (SD = 7.77; Table 1). The majority of respondents were single (40%) or in an unmarried relationship (45%). Fourteen percent of respondents completed some high school, and 30% received a high school degree. There were no significant differences in age, relationship status, education, employment, or number of children between women interviewed by an African American versus a European American interviewer. Reported relationship status did differ by interview mode. Relatively, more women completing a CASI reported being single (47%) than did women completing a FTFI (33%), whereas more women completing a FTFI reported being in an unmarried relationship (47%) or married (20%) than did women completing a CASI (42% and 11%, respectively). A higher percentage of women completing a FTFI reported having a job outside the home (50%) than did women completing a CASI (40%).

Table 1.

Characteristics of the Sample, Overall and by Interviewer Race and Mode of Interview, African American WIC Clients (n = 368).

Characteristics	Interviewer Race				Interview Mode
Characteristics	Total Sample	African American	European American	p	Computer	Face-to-Face	p
Age, in years, mean (SD = 7.77)	27.42	27.39	27.53	.883	27.67	27.20	.565
Relationship status (ref: single)	0.40	0.42	0.32	.178	0.47	0.33	.007
Unmarried relationship	0.45	0.42	0.53	—	0.42	0.47	—
Married	0.15	0.16	0.15	—	0.11	0.20	—
Education (ref: ≤ high school)	0.44	0.44	0.45	.052	0.42	0.46	.755
Some college	0.33	0.31	0.41	—	0.35	0.32	—
Completed college or higher	0.22	0.25	0.14	—	0.23	0.22	—
Job outside the home	0.45	0.44	0.47	.579	0.40	0.50	.050
Number of children, mean (SD = 1.65)	2.35	2.27	2.58	.112	2.35	2.34	.954

Note. p values less than .05 are noted in italics. WIC = Women, Infants, and Children.

Variables

Outcomes

We created 12 dichotomous (yes, no [reference]) outcomes for disclosure of lifetime and prior-year: (a) negotiation skills, (b) exposure to psychological IPV, (c) exposure to physical IPV, (d) exposure to sexual IPV, (e) exposure to any IPV (psychological, physical, or sexual), and (f) IPV-related injury. The outcomes on negotiation were included because they are part of the CTS2S and are arguably less sensitive items about relationship dynamics and so permitted us to gauge to what extent differences in disclosure by mode of interview depended on the sensitivity of the relationship item being asked. A summative score of IPV victimization for each participant was created with scores from 0 to 7 for each item about psychological, physical, or sexual IPV or any IPV-related injury.

Exposures

Exposure variables included mode of screening (FTFI or CASI [reference]) and interviewer race (European American [reference] or African American). Participants were randomized to participate in a FTFI or to enter their answers on a tablet computer (CASI). Participants were recruited by and completed the questionnaire with either a race-matched, African American interviewer or a non-race-matched, European American interviewer. Matching respondents and interviewers on race was not formally randomized, but rather determined by interviewer availability. Women interviewed by an African American interviewer were similar to those interviewed by a European American interviewer with respect to education, relationship status, employment, age, and number of children (results available on request).

Covariates

Measured covariates included the participant’s relationship status (married, unmarried relationship, or single [reference]), education (less than or equal to high school [reference], some college, or college), and currently employed outside the home (yes, no [reference]). We also collected data on the participant’s age in years and total number of children ever born.

Data Analysis

Demographic attributes of the sample across interview mode and interviewer race were compared via chi-square tests and t tests to determine any differences between these subsamples. Chi-square tests were conducted to compare reported rates of lifetime and prior-year IPV victimization, by type, across interview mode, and interviewer race. Summative scores for lifetime and prior-year IPV victimization were compared across interview mode and interviewer race using the Wilcoxon–Mann–Whitney test (which is appropriate for nonnormally distributed variables). We estimated 24 simple logistic regression models to assess first the unadjusted odds of reporting IPV victimization, overall and by type, by interview mode, and interviewer race. We then estimated 12 multiple logistic regression models to assess the associations of the exposure variables with each outcome, adjusted for participant characteristics that varied across interview mode and interview race, and that other studies have shown to be associated with the odds that women report IPV victimization (relationship status, education, and employment; Breiding, Black, & Ryan, 2008; Coker, Smith, McKeown, & King, 2000; Jewkes, 2002). Multiple logistic regression, which allows for the inclusion of multiple covariates in a regression model, is appropriate for the analysis of a dichotomous outcome predicted from a series of categorical independent variables. All multiple logistic regression models included exposure variables for race of interviewer and mode of interviewer, along with control variables, and did not include interaction terms. All analyses were conducted using STATA 12 (STATACorp LP, College Station, TX).

Results

Exposure Variables

One fourth of participants (25.3%) were interviewed by an African American interviewer, and 74.7% were interviewed by a European American interviewer (n = 368). Almost half (48.1%) completed the survey via CASI, and just more than half (51.9%) completed the survey via FTFI (n = 368).

Rates of IPV Disclosure by Race of Interviewer and Mode of Interview

The majority of women reported negotiation both in their lifetime (94%) and during the prior-year (88%). Forty-nine percent of women reported any lifetime IPV victimization, and 36% reported any IPV victimization during the prior-year. The most commonly reported type of IPV was psychological victimization (48% lifetime; 34% prior-year). A notable minority (13%) of women reported an IPV-related injury during their lifetime, and 7% reported an IPV-related injury during the previous year (Table 2).

Table 2.

Distribution of Self-Reported Negotiation Skills and Intimate Partner Violence, by Interviewer Race and Mode of Interview, African American WIC Clients (n = 368).

Outcome	Interviewer Race				Interview Mode
Outcome	Total Sample	African American	European American	p ^a	CASI	FTFI	p ^a
Negotiation
Lifetime negotiation (yes = 1)	0.94	0.94	0.96	.499	0.89	0.99	.000
Prior-year negotiation (yes = 1)	0.88	0.87	0.90	.385	0.80	0.95	.000
Lifetime IPV victimization, by domain
Psychological IPV (yes = 1)	0.48	0.49	0.43	.310	0.42	0.52	.055
Physical IPV (yes = 1)	0.12	0.12	0.11	.679	0.12	0.12	.788
Sexual IPV (yes = 1)	0.10	0.11	0.06	.245	0.08	0.11	.313
Any lifetime IPV (yes = 1)^b	0.49	0.50	0.45	.402	0.44	0.54	.046
IPV-related injury (yes = 1)	0.13	0.13	0.15	.568	0.16	0.11	.173
Prior-year IPV victimization, by domain
Psychological IPV (yes = 1)	0.34	0.37	0.27	.084	0.26	0.42	.001
Physical IPV (yes = 1)	0.07	0.07	0.06	.789	0.06	0.08	.308
Sexual IPV (yes = 1)	0.07	0.08	0.03	.136	0.03	0.10	.006
Any prior-year IPV (yes = 1)^b	0.36	0.39	0.30	.144	0.28	0.44	.002
IPV-related injury (yes = 1)	0.07	0.07	0.08	.745	0.07	0.06	.686
IPV summative scale^c
Lifetime (range = 0-56), mean	2.85	3.05	2.27	0.278^d	2.36	3.31	.018^d
Prior-year (range = 0-48), mean	1.75	1.92	1.26	0.175^d	1.30	2.17	.001^d

Note. p values less than .05 are noted in italics. WIC = Women, Infants, and Children; FTFI = face-to-face interview; CASI = computer-assisted self-interview; IPV = intimate partner violence;

Differences tested using chi-square test, unless otherwise noted.

Includes any verbal, physical, or sexual IPV or IPV-related injury.

Rank sum differences tested using Wilcoxon–Mann–Whitney test.

During both their lifetime and the prior-year, participants reported higher levels of negotiation during a FTFI (99%, 95%) than during a CASI (89%, 80%). There was no association between interviewer race and reporting negotiation and any type of lifetime or prior-year IPV victimization; however, there were different rates of disclosure by interview mode. Respondents who completed a FTFI reported significantly higher levels of any lifetime IPV victimization (54%) and any prior-year IPV victimization (44%) than did women who completed a CASI (44%, 28%; Table 2). The mean summative lifetime and prior-year IPV victimization scores were significantly higher among women participating in a FTFI (3.31, 2.17) than women participating in a CASI (2.36, 1.30; Table 2).

Multivariate Results

In unadjusted models, interviewer race was not associated with participants’ reporting of negotiation, any lifetime IPV victimization, or any prior-year IPV victimization. However, women participating in a FTFI had 1.52 times higher unadjusted odds of reporting any lifetime IPV victimization than women who completed a CASI (95% confidence interval [CI] = [1.01, 2.29]). Also, women interviewed via FTFI had 1.99 times higher unadjusted odds of reporting any prior-year IPV victimization than women interviewed via CASI (95% CI = [1.29, 3.08]). The odds of women completing a FTFI reporting prior-year psychological or sexual IPV victimization were 2.05 and 3.80 times the odds of reporting among women completing a CASI (95% CI = [1.32, 3.19] and [1.39, 10.41], respectively). Similarly, the odds of women interviewed via FTFI reporting both lifetime and prior-year negotiation were 11.36 and 4.46 times the odds of reporting among women completing a CASI (95% CI = [2.61, 49.54] and [2.14, 9.32], respectively; Table 3).

Table 3.

Unadjusted and Adjusted Odds Ratios (95% Confidence Intervals) of Reporting Negotiation Skills and IPV by Type, by Interviewer Race and Mode of Interview, African American WIC Clients (n = 368).

Exposure	African American Interviewer (ref: European American)		FTFI (ref: CASI)		African American Interviewer (ref: European American)		FTFI (ref: CASI)
Outcome	uOR	95% CI	uOR	95% CI	aOR^a	95% CI	aOR^a	95% CI
Negotiation
Lifetime negotiation	1.47	[0.48, 4.47]	11.36	[2.61, 49.54]	1.11	[0.34, 3.58]	10.54	[2.37, 46.84]
Prior-year negotiation	1.41	[0.65, 3.04]	4.46	[2.14, 9.32]	1.03	[0.45, 2.35]	3.97	[1.85, 8.53]
Lifetime IPV victimization, by domain
Psychological IPV	0.78	[0.49, 1.26]	1.49	[0.99, 2.26] p = .056	0.75	[0.46, 1.22]	1.48	[0.97, 2.25] p = .072
Physical IPV	0.85	[0.40, 1.80]	0.92	[0.49, 1.72]	0.80	[0.37, 1.73]	0.93	[0.49, 1.79]
Sexual IPV	0.59	[0.23, 1.46]	1.44	[0.71, 2.92]	0.52	[0.21, 1.32]	1.48	[0.72, 3.07]
Any lifetime IPV^b	0.82	[0.51, 1.31]	1.52	[1.01, 2.29]	0.79	[0.49, 1.28]	1.50	[0.98, 2.28] p = .061
IPV-related injury	1.22	[0.62, 2.37]	0.66	[0.36, 1.21]	1.20	[0.60, 2.38]	0.66	[0.35, 1.22]
Prior-year IPV victimization, by domain
Psychological IPV	0.63	[0.37, 1.07]	2.05	[1.32, 3.19]	0.58	[0.34, 1.00]	2.10	[1.33, 3.32]
Physical IPV	0.88	[0.34, 2.26]	1.53	[0.67, 3.46]	0.78	[0.30, 2.06]	1.44	[0.62, 3.36]
Sexual IPV	0.40	[0.12, 1.38]	3.80	[1.39, 10.41]	0.33	[0.09, 1.16]	4.31	[1.54, 12.07]
Any prior-year IPV^b	0.69	[0.41, 1.14]	1.99	[1.29, 3.08]	0.64	[0.38, 1.08]	2.02	[1.29, 3.16]
IPV-related injury	1.16	[0.47, 2.88]	0.85	[0.38, 1.91]	1.15	[0.45, 2.93]	0.79	[0.34, 1.83]

Note. Significant odds ratios and significant confidence intervals are noted in italics. IPV = intimate partner violence; WIC = Women, Infants, and Children; FTFI = face-to-face interview; CASI = computer-assisted self-interview; uOR = unadjusted odds ratios; aOR = adjusted odds ratios.

Models include both exposure variables for race of interviewer and mode of interview and adjust for relationship status, education, and job outside the home.

Includes any verbal, physical, or sexual IPV or IPV-related injury.

In models controlling for interview mode and demographic characteristics, interviewer race was not associated with participants’ reporting of negotiation, lifetime IPV victimization, or prior-year IPV victimization, but interviewer race was associated with the odds of disclosing prior-year psychological IPV victimization. Specifically, women interviewed by an African American interviewer had 42% lower odds of reporting psychological IPV than women interviewed by a European American interviewer (odds ratio [OR] = 0.58, 95% CI = [0.34, 1.00]; Table 3).

When controlling for interviewer race and demographic characteristics, we found no difference in the adjusted odds of reporting lifetime IPV victimization (in any domain) based on the mode of interview (Table 3). In contrast, the odds of reporting any prior-year IPV victimization were 2.02 times higher for women who completed a FTFI compared with women who completed a CASI. Also, compared with women who completed a CASI, women who completed a FTFI had 2.10 and 4.31 times the adjusted odds of disclosing prior-year psychological and sexual IPV victimization as well as 10.54 and 3.97 times the odds of reporting lifetime and prior-year negotiation, respectively (Table 3).

Discussion

Rates of reported IPV victimization among African American women are disproportionally higher than reported rates among women of other races and ethnicities (Black et al., 2011), even though African American women may be less likely to report IPV (Taft et al., 2009). To understand the burden of IPV victimization among African American women, accurate IPV screening procedures are needed; however, few studies exist that examine the contexts in which low-income African American women disclose IPV victimization. Although we hypothesized that race-matched interviews and CASIs would maximize IPV disclosure, our results suggest that low-income African American women are more likely to report IPV victimization when interviewed face-to-face rather than via CASI, but race-matching of the interviewer and respondent does not affect disclosure.

Overall, African American women in a WIC setting disclosed high rates of both lifetime (49%) and prior-year (36%) IPV victimization. This reported lifetime prevalence is consistent with reported rates (44%) among African American women in the 2010 National Intimate Partner and Sexual Violence Survey (NISVS; Black et al., 2011). However, our estimate of prior-year IPV victimization (36%) is much higher than NISVS estimates of prior-year IPV (5.9%) among African American women (Black et al., 2011). This difference could be a result of sample variation, with our sample including only women who are 185% of the poverty line and registered for WIC services. Research suggests that low-income African American women experience IPV at a higher rate than do women in households with higher incomes (Benson & Fox, 2004). Furthermore, African American women living in impoverished neighborhoods are more likely to experience IPV than European American women with a similar economic status (Benson & Fox, 2004; Cunradi, Caetano, Clark, & Schafer, 2000).

Interviewer race did not affect disclosure of IPV victimization. Prior research suggests that the race of the interviewer may affect disclosure of sensitive information (Davis et al., 2010; Livert et al., 1998). For example, one study among substance users found that interviewer race affected disclosure of physical and sexual abuse, in that respondents were more likely to disclose past abuse to a European American interviewer; however, matching clients with interviewers of their same race did not affect disclosure (Dailey & Claus, 2001). Thus, interviewer and client race-matching may not be necessary to get accurate assessments of sensitive information on IPV.

Mode of interview, however, was related to disclosure. FTFI enhanced disclosure over CASI for less sensitive items on relationship dynamics and more sensitive items on sexual IPV.

Specifically, women completing a FTFI reported more lifetime (adjusted odds ratio [aOR] = 10.54, 95% CI = [2.37, 46.84]) and prior-year (aOR = 3.97, [1.85, 8.53]) negotiation than women completing a CASI. Women completing a FTFI reported more prior-year IPV victimization than did women completing a CASI (aOR = 2.02, 95% CI = [1.29, 3.16]), but there was no difference in disclosure of lifetime IPV victimization by interview mode (aOR = 1.50, 95% CI = 0.98, 2.28]). The difference between prior-year and lifetime reporting suggests disclosure of recent IPV is more sensitive to interview mode. Higher rates of prior-year IPV disclosure may be due to participants feeling more comfortable disclosing IPV because of trust and rapport developed between the participant and the interviewer. Indeed, a follow-up survey revealed that 91.5% of women who completed a FTFI reported feeling comfortable giving their answers to an interviewer (response rate = 61.3%). Feelings of connection during in-person interviews have also led to more disclosure of study habits, substance use, and physical and sexual aggression, victimization, and perpetration (Rathod, Minnis, Subbiah, & Krishnan, 2011). The importance of interviewer—respondent connection is recognized in IPV-related research protocols that stress the importance of making the respondent feel comfortable during the interview by building rapport and remaining interested and nonjudgmental (Ellsberg et al., 2001; Watts et al., 2001). Indeed, researchers who receive this type of training are more successful in eliciting disclosure of IPV victimization (Jansen, Watts, Ellsberg, Heise, & Garcia-Moreno, 2004).

Potential barriers to IPV disclosure during any type of interview are respondents’ perceptions that they were not asked directly about their experiences with violence, beliefs that interviewers lack time and interest in discussing abuse, fears about involving police and courts, and concerns about confidentiality (Rodriguez, Sheldon, Bauer, & Perez-Stable, 2001). Perhaps women screened via CASI had concerns about the confidentiality of their responses or felt that the computer-based screening was being administered because the interviewer was not interested in discussing or lacked the time to discuss potential abuse. In addition, IPV disclosure may have been hindered among women completing a CASI because of low-literacy or low-computer-literacy. Women interviewed via CASI may have become frustrated with technical issues and finished the survey more quickly. Another possibility is that women interviewed via FTFI may have experienced the interviewer’s compassion and interest, which built rapport and may have made participants feel more comfortable disclosing IPV victimization.

Limitations

The study was conducted in two WIC clinics in greater metropolitan Atlanta; therefore, our results may not be generalizable to other regions, smaller communities, or clinics that serve a different race/ethnic clientele. Our study sample did not include women who did not speak and read English, so our results and estimated prevalences of IPV victimization may miss a group of women who are at particular risk for IPV, including women who are illiterate or immigrants. Future research may address these limitations by testing IPV screening in languages other than English and testing audio-CASI devices to accommodate low-literate women who are less familiar with computers. In our study, the matching of interviewers and participants by race was not randomized. Although women interviewed by a European American interviewer and women interviewed by an African American interviewer did not differ with regard to age, number of children, relationship status, education, or employment status, the lack of randomization could affect the results insofar as the race of the interviewer may have proxied for a range of unmeasured characteristics associated with disclosure. Unmeasured variables, for example, interview length, may have differed across interview mode, race-matched interview, or reporting of IPV and biased estimates of interest.

The majority of participants were interviewed by a European American interviewer, suggesting weak power to compare the race-matching variable. However, our focus on African American women, a marginalized population at high risk for IPV, is important. Although the correlation between the CTS2S and the CTS2 ranges from 0.65 to 0.94, the CTS2S has a lower sensitivity than the full CTS2 because it screens for only a subset of IPV-related behaviors (Straus & Douglas, 2004); thus, some participants who would have disclosed IPV using the longer instrument may have been misclassified as nonexposed (Straus & Douglas, 2004). Still, the 10 items from the CTS2S about IPV victimization are an economical way to screen for IPV in a busy clinic setting, such as WIC. Finally, our IPV screening took place in the context of a research study. Patient–provider relationships are different from participant–researcher relationships, so our results and conclusions may not directly translate into a clinical setting.

Conclusion

In a WIC setting, low-income African American women disclosed high rates of IPV victimization. This population of women is at an elevated risk for IPV victimization and, as a group, may experience substantial barriers to the disclosure of IPV. Thus, appropriate procedures for screening that encourage disclosure among low-income African American women are needed to understand the prevalence of IPV and to offer information, resources, and services. FTFI performed better in terms of enhancing disclosure for less sensitive items, including relationship negotiation, as well as more sensitive item such as sexual IPV. Our findings highlight the potential value of face-to-face screening to identify women at risk of IPV victimization. Programs should consider the costs and benefits of training staff versus using computer-based IPV screenings in WIC settings. Future research may investigate the use of other technologies that may enhance disclosure, including audio-CASI, which may be more appropriate for women with lower literacy skills.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The project was funded by the Georgia Department of Public Health, Maternal and Child Health Program; Emory Center for Injury Control (Grant – R49 CE001494 and PH2011120G); and the Hubert Department of Global Health, Rollins School of Public Health, Emory University.

Author Biographies

Danielle Fincher, MPH, is a graduate of the Rollins School of Public Health at Emory University. She received her bachelor of science degree in neuroscience from Rhodes College in Memphis, Tennessee, and is currently pursuing her MD at the University of Alabama School of Medicine in Birmingham, Alabama. Her past work has focused on women’s health and intimate partner violence.

Kristin VanderEnde, MSN, CNM, PhD, is a postdoctoral fellow in the Hubert Department of Global Health, Rollins School of Public Health at Emory University. She completed clinical training in nurse-midwifery at Vanderbilt University and received her doctorate in nursing from Emory University. She has experience in research on gender issues in Bangladesh, Egypt, Vietnam, and the United States. She has published in Social Science and Medicine, and her research has focused on the relationships between characteristics of communities and violence against women, the measurement of women’s agency, and scale development.

Kia Colbert, MPH, CHES, is a senior research project coordinator with the Department of Emergency Medicine and Center for Injury Control at Emory University. She completed her master of public health degree at Rollins School of Public Health at Emory University with a focus in prevention science. Prior to Rollins, she received her bachelor of science degree in health sciences at Clemson University. She has experience in research focused on intimate partner violence and translational methods.

Debra Houry, MD, MPH, is vice chair for research and associate professor in the Department of Emergency Medicine at Emory University School of Medicine and in the Department of Behavioral Science and Health Education and Department of Environmental Health at the Rollins School of Public Health. She is the director of the Emory Center for Injury Control and principal investigator on the Centers for Disease Control and Prevention (CDC) Injury Control Research Center grant. She has authored more than 90 peer-reviewed publications and book chapters on injury prevention and violence and is the president-elect for the Society for Advancement of Violence and Injury Research.

L. Shakiyla Smith, MPH, is the deputy director of the Emory Center for Injury Control and the deputy director for research in the Department of Emergency Medicine at Emory University’s School of Medicine. In her current work, she manages the infrastructure and functioning of a comprehensive research development and support program for the Department of Emergency Medicine and manages the organizational and programmatic activities and evaluation of the Emory Center for Injury Control. She has more than 9 years of public health experience, with a focus on violence and injury prevention.

Kathryn M. Yount, PhD, is Asa Griggs Candler Chair of Global Health and a faculty member in the Hubert Department of Global health and Department of Sociology at Emory University. Trained at Johns Hopkins University, she has 20 years of experience in research on gender issues, gender-based violence, and women’s empowerment in Africa, Asia, Latin America, the Middle East, and the United States. Her research has been supported by grants from the National Institutes of Health, National Science Foundation, United Nations Development Program, World Bank, World Health Organization, and several private foundations. She is an author of more than 100 publications, including 38 articles on violence against women and girls.

References

Amar

Laughon

Sharps

Campbell

& Expert Panel on Violence. (2013). Screening and counseling for violence against women in primary care settings. Nursing Outlook, 61, 187-191.

Benson

M. L.

Fox

G. L.

(2004). When violence hits home: How economics and neighborhood play a role (National Institute of Justice Research in Brief; NCJ 205004). Washington, DC: U.S. Department of Justice.

Black

M. C.

Basile

K. C.

Breiding

M. J.

Smith

S. G.

Walters

M. L.

Merrick

M. T.

. . . Stevens

M. R.

(2011). The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 Summary Report. Atlanta, GA: Centers for Disease Control and Prevention.

Breiding

M. J.

Black

M. C.

Ryan

G. W.

(2008). Prevalence and risk factors of intimate partner violence in eighteen U.S. States/territories, 2005. American Journal of Preventive Medicine, 34, 112-118.

Coker

A. L.

Smith

P. H.

McKeown

R. E.

King

M. J.

(2000). Frequency and correlates of intimate partner violence by type: Physical, sexual, and psychological battering. American Journal of Public Health, 90, 553-559.

Cunradi

C. B.

Caetano

Clark

Schafer

(2000). Neighborhood poverty as a predictor of intimate partner violence among White, Black, and Hispanic couples in the United States: A multilevel analysis. Annals of Epidemiology, 10, 297-308.

Cunradi

C. B.

Caetano

Schafer

(2002). Socioeconomic predictors of intimate partner violence among White, Black, and Hispanic couples in the United States. Journal of Family Violence, 17, 377-389.

Dailey

R. M.

Claus

R. E.

(2001). The relationship between interviewer characteristics and physical and sexual abuse disclosures among substance users: A multilevel analysis. Journal of Drug Issues, 31, 867-888.

Davis

R. E.

Couper

M. P.

Janz

N. K.

Caldwell

C. H.

Resnicow

(2010). Interviewer effects in public health surveys. Health Education Research, 25, 14-26.

10.

Department of Health and Human Services. (2012). Annual update of the health and human services poverty guidelines. Federal Register, 77, 4034-4035.

11.

Ellsberg

Heise

Pena

Agurto

Winkvist

(2001). Researching domestic violence against women: Methodological and ethical considerations. Studies in Family Planning, 32, 1-16.

12.

Hankin

Smith

L. S.

Daugherty

Houry

(2010). Correlation between intimate partner violence victimization and risk of substance abuse and depression among African-American women in an urban emergency department. Western Journal of Emergency Medicine, 11(3), 252-256.

13.

Hasley

(1995). A comparison of computer-based and personal interviews for the gynecologic history update. Obstetrics & Gynecology, 85, 494-498.

14.

Hatchett

Schuman

(1975-1976). White respondents and race-of-interviewer effects. Public Opinion Quarterly, 39, 523-528.

15.

Houry

Feldhaus

Peery

Abbott

Lowenstein

S. R.

al-Bataa-de-Montero

Levine

(2004). A positive domestic violence screen predicts future domestic violence. Journal of Interpersonal Violence, 19, 955-966.

16.

Houry

Kaslow

N. J.

Kemball

R. S.

McNutt

L. A.

Cerulli

Straus

. . . Rhodes

K. V.

(2008). Does screening in the emergency department hurt or help victims of intimate partner violence? Annals of Emergency Medicine, 51, 433-442.

17.

Houry

Kaslow

N. J.

Thompson

M. P.

(2005). Depressive symptoms in women experiencing intimate partner violence. Journal of Interpersonal Violence, 20, 1467-1477.

18.

Jansen

H. A. F. M.

Watts

Ellsberg

Heise

Garcia-Moreno

(2004). Interviewer training in the WHO multi-country study on women’s health and domestic violence. Violence Against Women, 10, 831-849.

19.

Jewkes

(2002). Intimate partner violence: Causes and prevention. Lancet, 359, 1423-1429.

20.

Kottenstette

J. B.

Stulberg

(2013). PURLs: Time to routinely screen for intimate partner violence? The Journal of Family Practice, 62(2), 90-92.

21.

Kramer

Lorenzon

Mueller

(2004). Prevalence of intimate partner violence and health implications for women using emergency departments and primary care clinics. Women’s Health Issues, 14, 19-29.

22.

Leiner

A. S.

Compton

M. T.

Houry

Kaslow

N. J.

(2008). Intimate partner violence, psychological distress, and suicidality: A path model using data from African American women seeking care in an urban emergency department. Journal of Family Violence, 23, 473-481.

23.

Livert

Kadushin

Schulman

Weiss

(May 1998). Do interviewer-respondent race effects impact the measurement of illicit substance use and related attitudes? Paper presented at the 53rd Annual Conference of the American Association for Public Opinion Research, St Louis, MO.

24.

Martinez-Schiferl

Zedlewski

Giannarelli

(2013). National and state-level estimates of Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) Eligibles and Program Reach, 2010 (Vol. 1). Alexandria, VA: U.S. Department of Agriculture, Food and Nutrition Service, Office of Research and Analysis.

25.

Newman

J. C.

Des Jarlais

D. C.

Turner

C. F.

Gribble

Cooley

Paone

(2002). The differential effects of face-to-face and computer interview modes. American Journal of Public Health, 92, 294-297.

26.

Plichta

S. B.

(2004). Intimate partner violence and physical health consequences: Policy and practice implications. Journal of Interpersonal Violence, 19, 1296-1323.

27.

Rabin

R. F.

Jennings

J. M.

Campbell

J. C.

Bair-Merritt

M. H.

(2009). Intimate partner violence screening tools: A systematic review. American Journal of Preventive Medicine, 36, 439-445.e434.

28.

Rathod

S. D.

Minnis

A. M.

Subbiah

Krishnan

(2011). ACASI and face-to-face interviews yield inconsistent estimates of domestic violence among women in India: The Samata Health Study 2005-2009. Journal of Interpersonal Violence, 26, 2437-2456.

29.

Rivara

F. P.

Anderson

M. L.

Fishman

Bonomi

A. E.

Reid

R. J.

Carrell

Thompson

R. S.

(2007). Healthcare utilization and costs for women with a history of intimate partner violence. American Journal of Preventive Medicine, 32, 89-96.

30.

Rodriguez

M. A.

Sheldon

W. R.

Bauer

H. M.

Perez-Stable

E. J.

(2001). The factors associated with disclosure of intimate partner abuse to clinicians. Journal of Family Practice, 50(4), 338-344.

31.

Rosenbaum

Rabenhorst

M. M.

Reddy

M. K.

Fleming

M. T.

Howells

N. L.

(2006). A comparison of methods for collecting self-report data on sensitive topics. Violence and Victims, 21, 461-471.

32.

Russell

(1994). New assessment tools for risk drinking during pregnancy: T-ACE, TWEAK, and others. Alcohol Health and Research World, 18, 55-61.

33.

Sanders

G. D.

Owens

D. K.

Padian

Cardinalli

A. B.

Sullivan

A. N.

Nease

R. F.

(1994). A computer-based interview to identify HIV risk behaviors and to assess patient preferences for HIV-related health states. Proceedings of the Annual Symposium on Computer Application in Medical Care, 20-24.

34.

Skinner

H. A.

(1982). The drug abuse screening test. Addictive Behavior, 7, 363-371.

35.

Sobkoviak

R. M.

Yount

K. M.

Halim

(2012). Domestic violence and child nutrition in Liberia. Social Science & Medicine, 74, 103-111.

36.

Straus

M. A.

Douglas

E. M.

(2004). A short form of the Revised Conflict Tactics Scales, and typologies for severity and mutuality. Violence and Victims, 19, 507-520.

37.

Straus

M. A.

Hamby

S. L.

Boney-McCoy

Sugarman

D. B.

(1996). The Revised Conflict Tactics Scales (CTS2): Development and preliminary psychometric data. Journal of Family Issues, 17, 283-316.

38.

Taft

C. T.

Bryant-Davis

Woodward

H. E.

Tillman

Torres

S. E.

(2009). Intimate partner violence against African American women: An examination of socio-cultural context. Aggression and Violent Behavior, 14, 50-58.

39.

Tjaden

Thoennes

(2000). Full report of the prevalence, incidence, and consequences of violence against women: Findings from the National Violence Against Women Survey. Washington, DC: National Institute of Justice.

40.

Turner

C. F.

Rogers

S. M.

Lindberg

L. D.

Pleck

J. H.

Sonenstein

F. L.

(1998). Adolescent sexual behavior, drug use, and violence: Increased reporting with computer survey technology. Science, 280, 867-873.

41.

U.S. Census Bureau. (2010). Atlanta (city), Georgia (American Fact Finder). Washington, DC: U.S. Department of Commerce. .

42.

Watts

Heise

Ellsberg

Moreno

C. G.

(2001). Putting women first: Ethical and safety recommendations for research on domestic violence against women. Geneva, Switzerland: World Health Organization, Family and Community Health, Department of Gender and Women’s Health.

43.

Weeks

M. F.

Moore

R. P.

(1981). Ethnicity-of-interviewer effects on ethnic respondents. Public Opinion Quarterly, 45, 245-249.

44.

Yount

K. M.

DiGirolamo

A. M.

Ramakrishnan

(2011). Impacts of domestic violence on child growth and nutrition: A conceptual review of the pathways of influence. Social Science & Medicine, 72, 1534-1554.

45.

Yount

K. M.

(2012). Spousal disagreement in the reporting of physical violence against wives in Southern Egypt. Journal of Family Issues, 33(11), 1540-1563.