Sociodemographic and Incident Variables as Predictors of Victim Injury From Intimate Partner Violence: Findings From Police Reports

Abstract

Predictors of victim injury from intimate partner violence (IPV) were investigated using 1,292 police reports collected in South Carolina in 2009/2010. All cases were opposite sex adults. Results from bivariate statistics showed that IPV cases with (n = 649) and without visible injuries (n = 643) differed on victim gender, victim race, type of relationship, and perpetrator’s alcohol use. Results from a logistic regression analysis predicting victim injury showed higher odds ratios for males, Whites, and couples identified as cohabitants. Although most victims, including most injured victims, were Black women, males and Whites were overrepresented in the injured group.

Keywords

intimate partner violence physical violence gender

Introduction

To date, several nationally representative studies have been completed in the United States in regard to prevalence rates of domestic violence (DV) and intimate partner violence (IPV). Of note, DV and IPV are overlapping constructs that differ in some significant ways. DV can be defined as a violent encounter between two or more people who share a household, whereas IPV can be defined as a violent encounter between two intimate partners, typically current or former intimate partners or spouses. Although DV and IPV can include several different types of violence, they typically include physical and sexual violence and sometimes psychological or emotional violence. However, DV and IPV tend to focus on physical and/or sexual violence. For the sake of continuity from here forward, we are going to use the term IPV when referring to both IPV and DV. This study utilized police reports from IPV incidents that met the legal definition of criminal DV in the state of South Carolina at the time of data collection. These definitions vary from state to state but often include physical violence aspects of IPV constructs. The current project’s aim is to predict visible injury in IPV victims from a large sample of police reports.

Several population-based studies have been conducted on IPV prevalence rates in the United States. One of them is the National Violence Against Women Survey (NVAWS), published by the U.S. Department of Justice and the Centers for Disease Control and Prevention (CDC), which found that 7.7% of women and 0.3% of men had been raped at some point during their life by an intimate partner (Tjaden & Thoennes, 2000). For physical IPV, 22.1% of women and 7.4% of men endorsed being victims during their lifetime. In terms of stalking, 4.8% of women and 0.6% of men reported being victims of stalking by an intimate partner during their lifetime. Of all women who had been raped, physically assaulted, and/or stalked at some point in their lives, 64.0% experienced violence inflicted by an intimate partner. In comparison, 16.2% of men reported that their perpetrator was an intimate partner.

The more recent National Intimate Partner and Sexual Violence Survey (NISVS), published by the CDC, revealed similar findings; one in three women and one in four men endorsed lifetime IPV victimization (e.g., physical violence, rape, and/or stalking; Black et al., 2011). More specifically, 32.9% of women and 28.2% of men endorsed lifetime physical IPV victimization, 10.7% of women and 2.1% of men endorsed lifetime stalking by an intimate partner, and 9.4% of women endorsed rape by an intimate partner. The rape prevalence rate for men was too low to produce reliable estimates. Another study the National Crime Victimization Survey (NCVS), published by the Bureau of Justice Statistics (BJS), found that IPV accounts for one in five nonfatal crime victimizations in the United States, and IPV for three in 20 (Truman & Morgan, 2014).

IPV rates vary by type of violence with the highest rates reported for psychological violence followed by physical; the lowest rates are for sexual violence (Black et al., 2011). In the NISVS, more than half of women and men endorsed psychological IPV victimization (Black et al., 2011). IPV has been linked to several negative outcomes that involve both those directly affected and society at large. For example, studies have estimated that IPV costs U.S. society US$6 billion a year; the majority are health-related costs (Centers for Disease Control and Prevention, 2003; Max, Rice, Finkelstein, Bardwel, & Leadbetter, 2004). In Kothari and Rhodes’s (2006) study with 964 police-identified female victims of IPV, eight out of 10 victims utilized the emergency department (ED) between 1 and 71 times during a 3-year period (median = 4). IPV victimization has also been linked to a range of negative psychosocial outcomes such as posttraumatic stress disorder, depression, substance use, poor physical health, and missed work days (Black et al., 2011; Coker et al., 2002).

IPV Reported to the Police

According to the NCVS, 55% of IPV incidents are reported to the police (Truman & Morgan, 2014). The most common reason victims of IPV did not report to the police was fear of retaliation or getting the perpetrator in trouble (38%; Langton, Berzofsky, Krebs, & Smiley-McDonald, 2012). Similarly, Felson, Messner, Hoskin, and Deane (2002) found that IPV victims were less likely to call the police because of concerns with privacy, creating problems for the perpetrator, and fear of reprisal. They also found that women were more likely to want or think they need protection and less likely to view the incident as trivial and a private/family matter, and therefore, female victims were overall more likely to call the police compared with male victims. Another possible reason for victims not calling the police is the possibility of dual arrest (e.g., arresting both partners). As there has been increased pressure to criminalize IPV, one of the solutions has been mandatory arrest. Chesney-Lind (2002), among others, has expressed concerns with mandatory arrest requirements, especially dual arrests, in which victims might be arrested because police are unable to determine who the perpetrator is. Thus, victims might not report IPV to the police because of concerns that they will be arrested. Studies have found that dual arrest policies disproportionally affect women (Durfee, 2012; Gerstenberger & Williams, 2012). Moreover, outcomes from Miller’s (2001) qualitative study with professionals involved with IPV dual arrests (e.g., law enforcement, social services) suggest that the IPV incidents are often taken out of context, ignoring factors such as previous history of victimization and motivation behind the violence (e.g., self-defense).

Researchers have found that several sociodemographic and incident factors affect the likelihood that IPV is reported to the police. In Ackerman and Love’s (2014) study with heterosexual female IPV victims, from the NCVS, they found that victims were more likely to report to the police if (a) they were racial/ethnic minorities (e.g., non-White), (b) the perpetrator had consumed alcohol, (c) the victim was injured, and (d) a commonly identifiable weapon was involved. Moreover, socioeconomic status (SES) mediated the relationship between ethnicity/race and reporting to police in such a way that ethnic/racial minority women with lower SES were more likely to report the IPV to the police than ethnic/racial minority women with higher SES. Regardless of SES, IPV victims were also less likely to report the violence if they had experienced prior IPV by the same perpetrator and if they lived together with the perpetrator.

Akers and Kaukinen (2009) found similar results when utilizing a Canadian sample of female IPV victims. Ethnic/racial minorities were more likely to report IPV to the police. Other factors that increased the likelihood of reporting to the police were if children witnessed the violence, weapons were used, the victim suffered any injuries, there was property damage, and the perpetrator had consumed alcohol. Younger women and married women were less likely to report to the police.

Rhodes, Dichter, and colleagues (2011) investigated the overlap between police and medical professionals’ involvement in IPV incidents by collecting data on “all female IPV victims identified by police and prosecutors during the year 2000” (p. 895) in a semirural Midwestern county (N = 993). The victims were 16 years or older and the IPV occurred within heterosexual relationships. After identifying these cases, the researchers tracked down all interactions with the police and the emergency room (ER) during 1999 and 2002. During these 4 years, the police had been called 3,426 times (median = 3). Those who used the ER following an IPV incident were more likely to be African American and unmarried. They were also more likely to have a police report related to a subsequent IPV incident on file as well as more calls to the police than IPV victims who did not use the ER. Of the IPV victims who were seeking medical help at the ER, those who were identified by medical providers as victims were more likely to be uninsured or on Medicaid and less likely to have any children. These victims were also more severely abused and more likely to call the police. Using the same sample, the role of children in IPV victims’ contact with the police and the ER was explored (Rhodes, Kothari, et al., 2011). Six out of 10 IPV victims had at least one child under the age of 18. Those with children were more likely to have prior IPV incidents (76% vs. 68%). According to the police reports, children were present during the IPV incidents 75% of the time.

In addition to factors that affect the likelihood of reporting IPV to the police, researchers have also reported on characteristics of IPV from police reports including demographic descriptions of perpetrators and victims as well as situation-specific factors (e.g., incident variables). As mentioned previously, IPV incidents from police reports are based on state-specific laws about what counts as DV.

Demographic variables

In police reports of IPV, the majority of perpetrators are males (Melton & Sillito, 2012; Truman & Morgan, 2014), and they are more likely to be intimate partners as opposed to other family members (Truman & Morgan, 2014). This is true for both simple assault and serious violent crimes (e.g., “rape, sexual assault, robbery, and aggravated assault,” Truman & Morgan, 2014, p. 5). There is also a gender effect for IPV victims in police reports, as females are more likely to be identified as the victims of IPV. As for the perpetrators, this gender effect applies to both simple assault and serious violent crimes (Truman & Morgan, 2014). In the NCVS, 82% of IPV victims were females (Truman & Morgan, 2014). Moreover, young adults (18-24 years old) are at heightened risk for IPV, especially young females. Individuals identifying as non-Hispanic African American or non-Hispanic biracial are also more likely to report being victims of IPV (Truman & Morgan, 2014). Similarly, IPV victims who used the ER following an IPV incident were more likely to be African American and never married (Rhodes, Dichter, Kothari, Marcus, & Cerulli, 2011).

Incident variables

There are several situation-specific variables that appear to be related to IPV reported to police. Previous research has shown that IPV is more likely to occur at people’s homes. In the NCVS, three out of four IPV incidents occurred in or around the victim’s home (Truman & Morgan, 2014). Individuals who have separated are also at heightened risk (Truman & Morgan, 2014).

Approximately one in five IPV incidents reported in the NCVS involved a weapon (Truman & Morgan, 2014). Most common weapon reported was a knife (7.8% of all IPV cases) followed by “other” (6.3%), gun (3.7%), and “unknown” (2.0%). Perpetrators’ use of firearms is a risk factor for more severe IPV, including homicide (Campbell, Glass, Sharps, Laughon, & Bloom, 2007). In a large sample of police reports (N = 1,421; all perpetrators were men), Folkes, Hilton, and Harris (2013) found that male perpetrators’ use of any weapon was associated with more severe violence (with and without weapons) and repeated incidents of IPV. Perpetrators who had access to firearms were more likely to use a weapon (but not necessarily a firearm) toward their partners. Moreover, there are numerous studies showing that substance use increases the likelihood for IPV. A recent meta-analysis found a small to medium effect for substance use and male-perpetrated violence and a small effect for substance use and female-perpetrated violence in heterosexual relationships (Foran & O’Leary, 2008).

IPV injuries

Data from the NCVS between 2003 and 2012 show that injuries were reported in 45% of the IPV incidents (Truman & Morgan, 2014). The majority of physical injuries were mostly minor such as “bruises or cuts” (Truman & Morgan, 2014, p. 8). Slightly more than one third of the injured victims received medical care. Moreover, Muelleman, Lenaghan, and Pakieser (1996) found that female IPV victims were more likely to have injuries to their head, face, neck, thorax, and abdomen compared with non-IPV injured women who presented to an ED. More than one in 10 women who presented to the ED with injuries were classified as IPV-related.

In nationally representative samples, researchers have found that women are more likely to be injured compared with men (Black et al., 2011; Tjaden & Thoennes, 2000). Black and colleagues (2011) found that 14.8% of women and 4.0% of men reported IPV-related injury. In the NVAWS (Tjaden & Thoennes, 2000), 41.5% of female victims and 19.9% of male victims had been injured during their most recent physical IPV incident. Most of the injuries were classified as minor (“scratch, bruise, welt,” Tjaden & Thoennes, 2000, p. 50). Women are also more likely to be killed by their intimate partners compared with men (39% vs. 3%; Catalano, 2013).

Tjaden and Thoennes (2000) reported on both gender-neutral risk factors (e.g., affected both male and female victims) and gender-specific risk factors for being injured from physical IPV. Gender-neutral risk factors were if the perpetrator threatened the victim or others close to the victim and if the perpetrator was under the influence of any substances. A unique risk factor for women was if the perpetrator was a former or current intimate partner, and unique risk factors for men were if a weapon was used and if the victim was under the influence of any substances (Tjaden & Thoennes, 2000).

Other studies have also compared injured with noninjured IPV victims. A study of women presenting with injuries to ED found that women injured by their male partners were more likely to be in a relationship, to be living with their former partner, and to have partners who abuse alcohol and/or drugs, have unstable employment, and have lower education (Kyriacou et al., 1999). Data from the NCVS showed that when perpetrators were intimate partners, victims were more likely to report injuries and they were more likely to report serious injuries such as “sexual violence injuries, gunshots, knife wounds, internal injuries, unconsciousness, and broken bones” (Truman & Morgan, 2014, p. 8). In another study, Walton-Moss, Manganell, Frye, and Campbell (2005) compared physically abused women with and without injuries. They found that those with injuries had been in the relationship longer (more than 1 year) and their partners were more controlling, had prior arrest(s) related to IPV, and had fair/poor mental health including suicidality.

Moreover, Thompson, Saltzman, and Johnson (2003) investigated predictors of physical injury in two national samples of women (NVAWS, Tjaden & Thoennes, 2000; Canadian Violence Against Women Survey [CVAWS], Statistics Canada, 1994). Across both samples, significant predictors of injury were if children witnessed the IPV, the perpetrator had been drinking alcohol, previous IPV (in the same relationship), and if the victim reported that she feared injury or death. In the Canadian sample, emotional IPV victimization was also a predictor of physical injury. Finally, as mentioned previously, Folkes and colleagues (2013) found that male perpetrators’ use of any weapon was associated with more severe violence (with and without weapons), which included injury.

In summary, previous studies show that injured IPV victims differ from noninjured IPV victims on a range of demographic and incident variables. The current study expands this research by predicting visible injury from demographic (e.g., victim gender, race, age), relationship type, and incident/perpetrator variables (e.g., suspect present or absent, use of an object, perpetrator’s use of alcohol) available in a large sample of police reports from IPV incidents.

Hypotheses

We hypothesized that IPV cases in which the victim had visible injuries would differ significantly from cases without visible injuries on a number of demographic and incident variables. Based on previous research, we hypothesized that female gender, minority status (e.g., non-White), younger age, relationship status (e.g., current vs. former relationship), perpetrator’s alcohol use, and use of an object would predict injury. We also included an additional variable that has not been examined in previous studies that we wanted to explore as a possible predictor of IPV-related injuries—the presence of the suspect when law enforcement arrived. We hypothesized that the suspect being gone when the police arrives might be an indicator of more severe violence (e.g., due to concerns that he or she will be arrested).

Method

Participants

Participants were perpetrators and victims from police reports on IPV incidents collected from seven different police districts in South Carolina between July 1, 2009, and June 30, 2010. There were a total of 1,451 police reports. The current study focused on adult opposite-sex IPV cases with single arrests, which resulted in 1,292 police reports. We excluded 10 child cases, 15 same-sex dyads (e.g., adult nonromantic family members of the same sex), and 134 dual arrest cases. The largest number of police reports (62%; n = 797) was from an area in which 47% of the population is Black and 42% White (U.S. Census, 2010). In the second largest number of police reports (12%; n = 160), census data (2010) show that 70% were White and 25% Black.

Procedure

A university institutional review board approved all study procedures. This project was part of a larger community-based program funded by the Office for Violence Against Women (OVW) to develop a coordinated community response to DV. Staff collected all IPV incident reports from each of the individual law enforcement agencies for incidents occurring from July 1, 2009-June 30, 2010. Upon a second review by staff of potential reports, no reports appeared to be missing. These reports include only those involving charges of criminal DV (and related assault charges) following South Carolina statutes of definition of this term (see below). South Carolina does not recognize same-sex relationship violence as criminal DV, but rather as simple assault. As such, for the purpose of the larger project focusing on South Carolina police reports on IPV incidents, same-sex relationship cases were not included in data collection.

The collection process varied depending on the agency. Although some agencies were able to easily print out their small number of reports and hand them over, other agencies provided a computer and printer for staff to physically go to the agency and search through their incident databases for criminal DV charges and manually print out each report. Reports included forced entry fields (e.g., victim race) in addition to a written narrative by the officer regarding the incident. The narrative was either handwritten or typed by the officer, depending on the jurisdiction. Incident report data from each agency were entered into separate Excel databases. Data were stripped of identifying information and provided with an ID number. All data sets were initially coded in Excel and then merged into one SPSS database for analyses. A codebook was created based on coding and classification from prior relevant studies (e.g., Folkes et al., 2013; Muelleman et al., 1996; Truman & Morgan, 2014). A number of string variables from the written police report narrative were coded for analyses including type of object, lethality of object used, place of injury, and severity of injury. Two trained research assistant volunteers coded all narrative data and demonstrated substantial interrater reliability (Cohen’s kappa = 0.79). After 25 reports were coded, oversampling for cases that included physical injury and objects used, one of the coauthors reviewed coding and discussed discrepancies. During this initial reliability checking/training phase, the majority of discrepancies arose in the type of object category and object lethality category. Definitions of each category were clarified. The remaining 1,267 reports were coded. Those items with discrepancies between coders were discussed among the research team (coauthors) and decided by consensus (approximately 30 total). Given this was a records review of community-based law enforcement police reports, we expected a higher rate of missing data.

Measures

Before operationalizing predictor and outcome variables, it is important to define IPV, which in the current project is based on the legal definitions of criminal DV in South Carolina during the year of data collection (July 1, 2009-June 30, 2010). Criminal DV as defined by South Carolina statutes includes inflicting physical harm or attempt to cause injury under circumstances to create fear or imminent peril to a household member. A household member includes a spouse, former spouse, persons who have a child in common, or male and female who are cohabitating or formerly have cohabited. Consequently, for this project, the definition of IPV is from the criminal justice system’s perspective.

Predictors

Predictor variables consisted of victim demographic variables, perpetrator behaviors, and types of relationships; perpetrator demographics were not examined as predictors due to their high overlap with victim demographics. For the victim demographic variables, gender, age, and race were included. The perpetrator behaviors examined were the use of an object, the presence of the perpetrator at the scene, and whether the officer noted alcohol use by the perpetrator. Whether an officer noted alcohol use by the perpetrator was coded dichotomously. Incident report forms allow officers to acknowledge if alcohol use is uncertain; however, it appeared that many officers did not mark either “no” or “unknown” when they did not detect alcohol use. As a result, “no,” “unknown,” and missing responses were coded as “not noted by the officer” and “yes” responses were coded as “noted by the officer.” The type of relationship was coded as spouse, cohabitant, and other. The other category captured those cases that were currently not in a romantic relationship (e.g., former partners/cohabitants, child in common).

Use of object included any object that was used to cause someone pain. This did not include body parts such as fists or hands. The objects were further classified based on type and lethality. In terms of types of objects, they were classified as blunt (e.g., firearm, baseball bat, pipe, hammer, and bottles), sharp (e.g., knife, scissors, and broken glass), and other (e.g., belt, garden hose, shoe, and pepper spray). Lethality was classified as minimally, moderately, and highly lethal. Minimally lethal were smaller objects that typically would not inflict much harm (e.g., cell phone, book, bottle, and letter opener). Moderately lethal were larger objects that had the potential to cause harm (e.g., baseball bat, door, cane, and hammer). Finally, highly lethal were objects that were likely to cause severe harm or death such as guns, knives, and motor vehicles.

Outcome

Visible victim injury was the dependent variable in this study. Answer options were dichotomous (i.e., yes/no). Cases with injuries were also coded based on where the injury was inflicted and how severe the injury was. Body regions were divided into six categories: face/head, neck/throat, torso/chest/abdomen, spine/back, upper extremities, and lower extremities. Injury severity was coded based on research by Ellsberg, Heise, Peña, Agurt, and Winkvist (2001). Moderate injury included cases in which the victim was slapped, grabbed, pushed, shoved, scratched, bit, or cut. It also included situations in which the perpetrator had threatened the victim and/or thrown an object at the victim. Severe injury were incidents in which the victim had been beaten, burned, kicked, dragged, hit with an object, hit with a fist, or slammed against a wall/large object. This also included cases in which the victim had been shot or stabbed.

Analytic Approach

To examine descriptive characteristics and correlates of cases in which victims sustained injuries, descriptive statistics and a series of chi-square and t tests were used to compare IPV cases with and without visible injuries. Following this, a hierarchical logistic regression analysis was conducted to examine unique predictors of injury status. Victim demographics were entered as predictor variables in the first step. These included the victims’ race, age, and gender. The type of relationship was entered as a predictor in the second step. Finally, perpetrator behaviors were entered as predictors in the third step, which included whether perpetrator alcohol use was noted by the officer, the presence of the perpetrator at the scene of the incident, and the use of an object in the assault. Finally, prior to analyses, missing data and analytical assumptions were examined. Multiple variables contained significant degrees of missingness. Whether the perpetrator remained at the scene was missing in 4.0% of cases (n = 26). These cases were estimated using “hot decking,” a missing data estimation approach for categorical data that appears to optimally reduce bias relative to other estimation techniques (Furlow, Fouladi, Gagne, & Whittacker, 2007). Perpetrator alcohol use was also significantly missing (n = 459, 35.5%); however, upon further examination of incident reports, it appeared that officers often did not respond to items related to alcohol use if alcohol use was not apparent. In fact, multiple jurisdictions rarely indicated an absence of alcohol or other substances (i.e., less than 10% of reports). In one jurisdiction, officers never confirmed an absence of alcohol or other substances on any report. This contrasts with jurisdictions with more complete data in that alcohol and substances were noted as absent by officers in a majority of cases. To account for this, analyses were examined with alcohol use coded as “present” if officers noted alcohol use and coded as “absent” if values were missing or if officers indicated that alcohol use was not present or unknown. Analyses were also examined only with confirmed present/absent responses. Relations between alcohol and victim injury did not differ between these two analyses. As a result, analyses are presented with the full sample in which unknown and confirmed absent categories were combined into a single category. No other variables contained significant missing values (i.e., less than 2% missing). Multicollinearity diagnostics also suggested multicollinearity of predictors was within acceptable ranges (variance inflation factors’ [VIF] values were all less than 2.0).

Results

Descriptive Statistics

Descriptive statistics for predictor and dependent variables will be provided for the overall sample and separately for IPV cases with and without injuries.

Overall sample

See Table 1 for descriptive data of predictor variables for all IPV cases. Perpetrators were significantly more likely to be men, χ²(1) = 656.33, p < .001, whereas victims were significantly more likely to be women, χ²(1) = 637.25, p < .001. Results from a paired-samples t test also show that perpetrators are significantly older than the victims, t(1,217) = 5.38, p < .001. Age of perpetrators ranged from 16-79, whereas age of victims ranged from 17-82 years old. For both perpetrators and victims, most were in the age categories of 25-34 (35.2% and 32.6%, respectively) followed by 18-24 (23.1% and 29.5%) and 35-44 (21.6% and 19.0%). The majority of perpetrators and victims were Black followed by White, χ²(2) = 699.03, p < .001, and χ²(2) = 630.42, p < .001, respectively. Perpetrator–victim pairs were more likely to be separated as opposed to currently in a relationship as spouses or cohabitants, χ²(2) = 33.49, p < .001.

Table 1.

Descriptive Statistics of Predictor Variables for All IPV Cases As Well As Separately for Cases With and Without Visible Injuries.

	All cases (N = 1,292)	Injuries (n = 649)	No injuries (n = 643)	Test statistic^a	p
	M (SD)/n (%)	M (SD)/n (%)	M (SD)/n (%)	Test statistic^a	p
Perpetrator gender				χ²(1) = 7.43	.006
Male	1,101 (85.2%)	536 (82.6%)	565 (87.9%)
Female	183 (14.2%)	109 (16.8%)	74 (11.5%)
Victim gender				χ²(1) = 7.08	.008
Male	181 (14.0%)	110 (16.9%)	71 (11.0%)
Female	1,076 (83.3%)	539 (83.1%)	537 (83.5%)
Age
Perpetrator	33.39 (11.18)	33.03 (11.03)	33.76 (11.33)	t(1,260) = 1.16	.247
Victim	32.22 (11.16)	32.02 (11.22)	32.44 (11.10)	t(1,241) = 0.66	.510
Perpetrator race				χ²(2) = 1.71	.426
White	492 (38.1%)	258 (39.8%)	234 (36.4%)
Black	776 (60.1%)	381 (58.7%)	395 (61.4%)
Other	12 (0.9%)	5 (0.1%)	7 (1.1%)
Victim race				χ²(2) = 10.06	.007
White	526 (40.7%)	297 (45.8%)	229 (35.6%)
Black	714 (55.3%)	346 (53.3%)	368 (57.2%)
Other	13 (1.0%)	4 (0.6%)	9 (1.4%)
Type of relationship				χ²(2) = 7.93	.019
Married	382 (29.6%)	186 (28.7%)	196 (30.5%)
Cohabitant	366 (28.3%)	209 (32.2%)	157 (24.4%)
Other	519 (40.2%)	250 (38.5%)	269 (41.8%)
Suspect gone at scene				χ²(1) = 0.81	.369
Yes	629 (48.9%)	324 (50.1%)	335 (52.4%)
No	659 (51.2%)	325 (49.9%)	304 (47.6%)
Use of object				χ²(1) = 1.23	.268
Yes	220 (17.0%)	118 (18.2%)	102 (15.9%)
Not noted	1,072 (83.0%)	531 (81.8%)	541 (84.1%)
Alcohol use (perpetrator)^b				χ²(1) = 12.09	.001
Noted by officer	347 (26.9%)	202 (31.1%)	145 (22.6%)
Not noted	945 (73.1%)	447 (68.9%)	498 (77.4%)

Note. IPV = intimate partner violence.

Compares cases with injuries versus no injuries.

In approximately one in five cases, it was noted in the police report that an object was used during the IPV incident. For those cases that did involve an object, the majority of them used one object (n = 201; 91.4%). The total number of objects reported was 239; 50.2% (n = 120) were classified as blunt, 28.9% (n = 69) as sharp, 19.7% (n = 47) as other, and 1.3% (n = 3) as unknown (i.e., no description provided). Out of the objects classified as blunt, 20.0% (n = 24) were considered minimally lethal, 40.8% (n = 49) moderately lethal, and 39.2% (n = 47) highly lethal. For those classified as sharp, 4.3% (n = 3) were considered minimally lethal and 95.7% (n = 66) highly lethal. In the “other” category, all objects were classified as minimally lethal.

In close to half of all IPV cases, the suspect was not at the scene when the police arrived (compared with being at the scene; p = .384). More than one in four perpetrators were noted by the officer to be under the influence of alcohol.

Cases with and without injuries

Half of the police reports indicated that the victim had sustained at least one injury (n = 649; 50.2%). The number of injuries sustained ranged from one to five. When injuries were noted, most often only one injury was noted (n = 453; 69.8%) and two injuries were also frequently noted (n = 158; 24.3%). Less than 10% of those who sustained injuries had sustained three or more injuries. There were a total of 894 injuries reported; 47.5% (n = 425) of the injuries were inflicted to victims’ face/head, 23.9% (n = 214) to the victims’ upper extremities, 11.2% (n = 100) to the victims’ neck/throat, 8.3% (n = 74) to the victims’ torso/chest/abdomen, 6.8% (n = 61) to the victims’ lower extremities, and 2.2% (n = 20) to the victims’ spine/back.

More than half of all cases with injuries were classified as severe injuries (56.4%; n = 366). The rest were classified as moderate injuries (43.6%; n = 283). Moreover, in cases where victims had severe injuries, they also had significantly more injuries (M = 1.54; SD = 0.75) than those classified as moderate (M = 1.17; SD = 0.43), t(647) = −8.07, p < .001, corrected for significant Levene’s test. For those classified as severe injuries, 50.5% (n = 285) were inflicted to victims’ face/head, 21.6% (n = 112) to victims’ upper extremities, 13.7% (n = 77) to victims’ neck/throat, 8.0% (n = 45) to victims’ torso/chest/abdomen, 6.4% (n = 36) to victims’ lower extremities, and 1.6% (n = 9) to victims’ spine/back. For those classified as moderate injuries, 42.4% (n = 140) were inflicted to victims’ face/head, 30.9% (n = 102) to victims’ upper extremities, 8.8% (n = 29) to victims’ torso/chest/abdomen, 7.0% (n = 23) to victims’ neck/throat, 7.6% (n = 25) to victims’ lower extremities, and 3.3% (n = 11) to victims’ spine/back.

See Table 1 for descriptive data and bivariate statistics on predictor variables comparing IPV cases with and without injuries. Cases with and without injuries varied significantly by perpetrator and victim gender. Among the male victims, 60.8% had injuries and 39.2% did not. However, female victims were more equally distributed between victims with (50.1%) and without injuries (49.9%). It is important to keep in mind that perpetrators were still significantly more likely to be males and victims significantly more likely to be females. These findings suggest that female victims reported a higher frequency of injury, although male victims were more likely to be injured.

Ad hoc analyses were completed to further explore these gender differences. Among the injured group, there was no difference in injury severity between male and female victims, χ²(1) = 0.39, p = .531; however, an object was more likely involved when a male victim had been injured (43.6% of the cases) compared with a female victim (13.0%), χ²(1) = 57.69, p < .001. Again, it is important to note that most victims were women and most injured victims were women. Similarly, among the 220 cases that involved the use of an object, 68.2% (n = 150) of the victims were women. Narrowing it down further, among the 118 cases in which an object was used and the victim was injured, 59.3% (n = 70) of the injured victims were women.

Cases with and without injuries differed by victim race. The majority of victims were Black or White; however, Black victims were less likely to have injuries noted (48.5% of victims) relative to White victims (56.5% of victims; p = .007). In bivariate chi-square analyses, cases with injured victims were more likely to be classified as being cohabitants (53.6%) compared with those without injuries (42.9%), especially compared with those classified as spouses (48.9% injured vs. 54.7% noninjured). The other relationship types had a more even distribution between injured and noninjured cases (e.g., 48.2 vs. 51.8%). In most instances in which officers noted that perpetrators had used alcohol, officers also noted injuries among victims (58.2% of alcohol-related instances).

Logistic Regression

Results from logistic regression suggested that the first step, which utilized victim demographics as predictor variables, significantly predicted injury status (Step χ² = 16.55, df = 4, p = .002, Nagelkerke r² = .02). In the second step, relationship type marginally improved predictions of injury status (Step χ² = 5.16, df = 2, p = .076). In the third step, perpetrator behavior variables did not significantly improve predictions of injury status (Step χ² = 5.70, df = 3, p = .127). See Table 2 for results on individual variables.

Table 2.

Hierarchical Logistic Regression Predicting Visible Victim Injury.

	Step 1			Step 2			Step 3
	OR	95% CI	p	OR	95% CI	P	OR	95% CI	p
Male victim	1.50	[1.08, 2.09]	.017	1.52	[1.09, 2.12]	.013	1.50	[1.06, 2.12]	.021
Victim age	0.99	[0.98, 1.00]	.175	0.99	[0.98, 1.00]	.223	0.99	[0.98, 1.00]	.106
Victim race
White	1.00	NA	NA	1.00	NA	NA	1.00	NA	NA
Black	0.72	[0.57, 0.91]	.006	0.73	[0.58, 0.93]	.010	0.78	[0.61, 1.00]	.046
Other	0.38	[0.12, 1.25]	.110	0.41	[0.12, 1.35]	.143	0.39	[0.12, 1.31]	.127
Type of relationship
Married				1.00	NA	NA	1.00	NA	NA
Cohabitant				1.37	[1.02, 1.84]	.038	1.34	[0.99, 1.80]	.057
Other				1.05	[0.79, 1.41]	.724	1.08	[0.80, 1.45]	.620
Suspect gone^a							0.91	[0.72, 1.14]	.403
Use of object							1.03	[0.76, 1.40]	.847
Alcohol use (perpetrator)^b							1.34	[1.02, 1.76]	.034
Model χ² (df)	16.55 (4)**			21.71 (6)***			27.41 (9)***
Step χ² (df)	16.55 (4)**			5.16 (2)			5.70 (3)
Nagelkerke r²	.02			.02			.03

Note. OR = odds ratio; CI = confidence interval.

Missing values for whether the perpetrator was present at the scene was estimated with a hot decking procedure.

Given that officers appeared to not comment on substance use when substance or alcohol use was not present, instances in which officers did not report substance use were combined with instances in which officers confirmed a lack of substance use.

p < .05. **p < .01. ***p ≤ .001.

Discussion

In this sample of 1,292 police reports of IPV incidents, the majority of perpetrators were Black males in their mid-30s and the majority of victims were Black females in their mid-30s, which is fairly consistent with the demographic profile of the region. Perpetrators were significantly older than the victims, and pairs were more likely to be separated as opposed to currently in a relationship. In close to one out of five cases, an object was involved. In half of these cases, the objects were classified as blunt (e.g., baseball bats, guns, and motor vehicles) and in nearly one third as sharp (e.g., knives, scissors, and broken glass). In more than one out of four cases, the perpetrator was under the influence of alcohol.

Half of all victims had visible injuries with the majority of their injuries to the face/head and upper extremities, and more than half classified as severe. In terms of our hypotheses, there were several differences between IPV cases with visible injuries compared with those without, but not all differences were in the expected direction. Among the injured victims, the majority of perpetrators were males and the majority of victims were females, which is similar to previous studies (e.g., Black et al., 2011; Tjaden & Thoennes, 2000). However, among the male victims, a greater proportion were in the injured group than the noninjured group compared with female victims, who were more equally distributed between injured and noninjured victims. This finding might indicate that male victims are more likely to report the incident to the police if they have been injured. Although we found no differences in injury severity between male and female victims, cases with male victims were more likely to indicate that an object had been used, which could be an indicator of more severe IPV (Folkes et al., 2013). Findings from the current study also corroborate results from the NVAWS, which found that weapon use was a unique predictor of injuries for male victims (Tjaden & Thoennes, 2000).

In addition, although we predicted a significant difference for victim race, our results were in the opposite direction of the expected, with White victims more likely to have visible injuries than Black victims.

Moreover, when comparing IPV cases with and without injuries, victims who were unmarried cohabitants with their perpetrators were more likely to sustain injuries than victims who were married or not cohabitating with the perpetrator (e.g., former partners). This was also a significant predictor of visible victim injury after accounting for demographic variables. Few studies have investigated the role of current versus former partners. In the NCVS, they found that IPV perpetrated by an intimate partner is more likely to result in injuries than IPV perpetrated by a nonintimate partner (e.g., family member; Truman & Morgan, 2014). IPV was also more likely to lead to more severe injuries such as gunshot and knife wounds. Moreover, studies have found that IPV is more likely to be perpetrated by a current or former partner (Tjaden & Thoennes, 2000; Truman & Morgan, 2014). On the contrary, some studies have found that being a cohabitant is a risk factor for physical IPV, especially for female victims (O’Leary, Tintle, & Bromet, 2014). Subsequent studies should consider the effect of a coparenting relationship as an additional variable.

As predicted and supported by other studies (Kyriacou et al., 1999; Thompson et al., 2003; Tjaden & Thoennes, 2000), perpetrator alcohol use was found to be a risk factor for victim injury. In the injured group, the police reports were more likely to state that the perpetrator had been drinking alcohol. Unfortunately, we did not have any information about victims’ alcohol use, which has been found to be a predictor of injury among male victims (Tjaden & Thoennes, 2000).

Although we hypothesized that perpetrators might be less likely to be at the scene when the police arrived if the victim had sustained any injuries, this was not supported by the results. There were actually no significant differences between the injured and noninjured groups on this variable. However, given that other studies have found associations between IPV and presence/absence of suspect at the scene, this variable merits further attention. Hirschel and Buzawa (2013) found that perpetrators who are not at the scene when police arrived were 5.5 times less likely to be arrested. Perpetrators’ presence at the scene was a stronger predictor of arrest than presence of evidence or indications of more severe violence (e.g., injury and weapons). Although more information about previous arrests and/or IPV incidents can shed more light on this relationship between perpetrators’ presence at the scene and resulting arrest, officers may consider implementing more standardized follow-up procedures to locate perpetrators who are not present upon their arrival (especially repeat IPV perpetrators).

Although the prevalence rate of using an object in the current study was similar to findings from the NCVS (e.g., one in five; Truman & Morgan, 2014), there were some other noticeable differences between the current study and previous studies. First, in the current study, it was impossible to determine if police reports with no documentation of objects indicated that no objects were used or if this information was not documented (e.g., missing). Second, the operationalization of use of an object or weapon differs significantly between studies, which could explain the mixed outcomes. In the current study, use of an object was operationalized as using any object, except for body parts (e.g., fists and hands), that could cause someone pain. This definition is similar to definitions from other studies (“hit you with something that could have hurt you” and/or “used or threatened to use a gun or knife,” Akers & Kaukinen, 2009, p. 164; “any object other than hands or feet that could cause injury,” Folkes et al., 2013, p. 1147). The operationalization of object use in the current study also overlaps but is not identical to how object use is identified in the NCVS. In the NCVS, answer options are classified into firearm, knife, other, and unknown (Ackerman & Love, 2014; Truman & Morgan, 2014), which is based on asking:

has anyone attacked or threatened you in any of these ways—(a) with any weapon, for instance, a gun or a knife (b) with anything like a baseball bat, frying pan, scissors, or stick (c) by something thrown, such as a rock or bottle. (BJS, 2012)

Overall, these varying definitions complicate interpretations of results across studies.

Limitations and Future Directions

There were several limitations with the current study. Overall, variables for the current project were based on the information provided in the police reports, which is linked to several limitations. There was limited standardization of what was recorded in the police reports. For instance, police officers might vary on what they take note of and how much detail they provide. There could also be differences between the police districts, as this project included seven different police districts in South Carolina. This potential variability between documentation practices increases the likelihood for confounding factors. Another limitation with using police reports for data collection was the ambiguity about missing information. Several of the variables examined were documented in the narrative of the police report, as there were no checklists or yes/no options for some of the IPV characteristic variables within the police report. Multiple items on the incident report (including items not examined in the current article) were frequently left missing if the variable was unremarkable. When no information was provided about a certain variable, it was impossible to determine if that was because this information was missing (e.g., an object was used but it was not documented) or because it was absent (e.g., no object was used and therefore nothing was documented about it). For variables taken from the narrative of the officer (e.g., use of an object), it is difficult to determine if an object was not present or simply not mentioned in the narrative of the report.

Furthermore, the variables for the current project were created based on what was available from the police reports. Therefore, operationalization of variables was based on the information available as opposed to previously standardized and psychometrically sound measures or items. The results from the current study may also not be generalizable to all IPV cases. The current sample is limited by demographic and regional variables. Unfortunately, there was no information about SES in the current sample. This would be an important variable to consider in future studies, especially as Ackerman and Love (2014) found that SES mediated the relationship between race/ethnicity and reporting IPV to the police; minorities with lower SES were more likely to report IPV to the police. To our knowledge, there are no previous studies that have investigated race/ethnicity, SES, and risk for IPV-related injuries. There was also limited information about children. From the information provided in the police reports, we were unable to create any reliable variables in regard to whether children were present during the IPV incident or when police arrived, and/or if there were children in the household. Another limitation to the current dataset was information about strangulation. Strangulation is another type of victim injury that has been linked to more severe IPV and homicide (Campbell et al., 2007). In the current sample, 13.2% (n = 170) of the police reports indicated that the victim had been strangled, which is substantially lower than rates reported in other studies (47%, Block et al., 2000; 27%; Glass et al., 2008). Unfortunately, we were unable to determine if this information was truly absent in the other reports or not documented.

Summary and Implications

In this sample of 1,292 police reports of IPV incidents, there were significant differences between cases with and without visible victim injuries. Among the injured victims, the majority of perpetrators were males and the majority of victims were females. However, among male victims they were overrepresented in the injured group, whereas the female victims were more equally distributed between the injured and noninjured groups. These findings might suggest that male victims are more likely to report IPV incidents to the police, or the police are more likely to consider the man to be the victim if he has visible injuries. Moreover, although the majority of victims were Black, White victims were overrepresented in the injured group. A potential explanation is that the main effect of race on likelihood of injury is really reflecting the effect of SES, as previous research has shown that minorities with lower SES were more likely to report IPV to the police (Ackerman & Love, 2014); SES was not measured in the current study. Moreover, in the current study, perpetrator–victim pairs categorized as cohabitants were overrepresented in the injured group, and perpetrators were more likely to have used alcohol if the victim had visible injuries.

Based on findings from the current study and from our experience using police reports as a data source, we have several suggestions for IPV policy and police involvement in IPV cases. In the current study, we coded some variables based on reading the narratives from the police reports. One major issue was our inability to determine if absence of information (e.g., no mention of injuries) meant true absence (e.g., no injuries) or if this information was missing (e.g., possibly present but not documented). As several of the factors included in the current study have been linked to more severe IPV (e.g., injuries, weapon use), and possible homicide (Campbell et al., 2007), it is crucial for future studies to have a more systematic way of measuring these variables. It would also be important for law enforcement to use systematic data gathering rather than narrative police report for documentation of the presence or absence of these variables to be able to determine future risk. We suggest that law enforcement use IPV-specific police reports to include check boxes for these variables to increase the likelihood that police officers ask and document these IPV characteristics, both for documentation of relevant evidence and to assist in assessing potential future risk for IPV victims. Factors that would be important to document are victim injuries, including strangulation; weapon use; substance use by perpetrators and victims; and presence of children (as witnesses, in the household).

Consistent documentation of the identified IPV incident characteristics may also result in greater success for prosecution of these cases, which, in turn, could serve as a deterrent for some perpetrators to revictimize. Asking about children’s involvement could facilitate additional assessment of their needs, such as a referral to a local child advocacy center for trauma-focused interventions. Furthermore, training for law enforcement on IPV risk factors could improve their ability to respond to victims’ needs once they have gathered and documented the incident details recommended above. Although the current article focuses on IPV that is reported to law enforcement, the majority of IPV goes unreported. Thus, prevention efforts should continue to focus heavily on settings and individuals with whom victims come in contact outside of an acute episode of violence (e.g., health providers and their offices). Another recommendation is for IPV laws to be more inclusive such as including same-sex relationships and broader definitions of IPV (e.g., including nonphysical violence).

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: This research was supported by Grant 2013-HI-AX-K009 from the DOJ Office of Violence Against Women. Views contained in this article do not necessarily represent those of the DOJ Office of Violence Against Women or its associated agencies.

Author Biographies

Marie E. Karlsson, PhD, is an assistant professor in clinical psychology at Murray State University and a licensed psychologist in the Commonwealth of Kentucky. She completed her doctoral studies at the University of Arkansas and her predoctoral internship at the Medical University of South Carolina (MUSC). Her research focuses on interpersonal trauma, abusive relationships, female victimization, and trauma sequelae and treatments. Within these domains, she has a special interest in underserved and at-risk populations.

Kathryn Reid Quiñones, PhD, is a licensed psychologist and the director of Grants and Program Evaluation at the Dee Norton Child Advocacy Center. Her expertise includes evidence-based treatments for trauma-related mental health issues. She is the coinvestigator on several federally funded service grants to provide evidence-based interventions to child victims of abuse. Her research interests include predictors of youth risk and resilience following trauma exposure.

Cristina M. López, PhD, is a clinical psychologist and assistant professor at MUSC. She completed her doctoral degree in clinical psychology at Florida State University. Her research interests include increasing engagement of underserved populations in mental health services, with an emphasis on retention of trauma-exposed women and youth in evidence-based treatments. She is the clinical supervisor for trauma-focused psychology interns at a school-based mental health clinic and has also presented her findings related to trauma-focused mental health at over 20 national conferences and peer-reviewed publications.

Arthur R. Andrews III is an assistant professor with dual appointment in the Department of Psychology and Institute for Ethnic Studies at the University of Nebraska–Lincoln. He received his PhD from the University of Arkansas in 2014 and subsequently completed a postdoctoral fellowship in traumatic stress at the National Crime Victims Research and Treatment Center (NCVC) at MUSC. He has dual lines of research. The first focuses on mental health disparities with an emphasis on trauma and violence-related disparities among Latinas/os. In the second, he seeks to develop and evaluate novel treatment approaches to reduce disparities.

Megan M. Wallace is a licensed master’s social worker at the NCVC within the Department of Psychiatry and Behavioral Sciences at MUSC. She completed a B.S. in psychology at the College of Charleston and her MSW from the College of Social Work at the University of South Carolina. She has been trained in evidence-based treatment of trauma-related mental health issues and has expertise in traumatic loss and community-based mental health approaches.

Alyssa Rheingold is a licensed clinical psychologist and professor at the NCVC within the Department of Psychiatry and Behavioral Sciences at MUSC. She is the director of clinical operations at the NCVC. Her expertise includes evidence-based treatment of trauma-related mental health issues, child sexual abuse prevention, interpersonal violence, and loss by homicide. She is the principal investigator (PI) of several federally funded service grants to improve resources for survivors of homicide, and a service grant to provide services for Hispanic interpersonal violence victims.

References

Ackerman

Love

T. P.

(2014). Ethnic group differences in police notification about intimate partner violence. Violence Against Women, 20, 162-185. doi:10.1177/1077801214521327

Akers

Kaukinen

(2009). The police reporting behavior of intimate partner violence victims. Journal of Family Violence, 24, 159-171. doi:10.1007/s10896-008-9213-4

Black

M. C.

Basile

K. C.

Breiding

M. J.

Smith

S. G.

Walters

M. L.

Merrick

M. T.

et al . (2011). The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 summary report. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention.

Block

C. R.

Devitt

C. O.

Fonda

Fugate

Marting

McFarlane

et al . (2000). The Chicago Women’s Health Study: Risk of serious injury or death in intimate violence: A collaborative research project. Washington, DC: U.S. Department of Justice, National Institute of Justice.

Bureau of Justice Statistics. (2012). National crime victimization survey. Retrieved from http://www.bjs.gov/content/pub/pdf/ncvs1_2012.pdf

Campbell

J. C.

Glass

Sharps

P. W.

Laughon

Bloom

(2007). Intimate partner homicide: Review and implications of research and policy. Trauma, Violence & Abuse, 8, 246-269.

Catalano

(2013). Intimate partner violence: Attributes of victimization, 1993-2011. Washington, DC: U.S. Department of Justice.

Centers for Disease Control and Prevention. (2003). Costs of intimate partner violence against women in the United States. Retrieved from http://www.cdc.gov/violenceprevention/pdf/IPVBook-a.pdf

Chesney-Lind

(2002). Criminalizing victimization: The unintended consequences of pro-arrest policies for girls and women. Criminology & Public Policy, 2, 81-90. doi:10.1111/j.1745-9133.2002.tb00108.x

10.

Coker

A. L.

Davis

K. E.

Arias

Desai

Sanderson

Brandt

H. M.

et al . (2002). Physical and mental health effects of intimate partner violence for men and women. American Journal of Preventative Medicine, 23, 260-268. doi:10.1016/S0749-3797(02)00514-7

11.

Durfee

(2012). Situational ambiguity and gendered patterns of arrest for intimate partner violence. Violence Against Women, 18, 64-84. doi:10.1177/1077801212437017

12.

Ellsberg

Heise

Peña

Agurt

Winkvist

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

13.

Felson

R. B.

Messner

S. F.

Hoskin

A. W.

Deane

(2002). Reasons for reporting and not reporting domestic violence to the police. Criminology, 40, 617-647.

14.

Folkes

S. E. F.

Hilton

N. Z.

Harris

G. T.

(2013). Weapons use increases the severity of domestic violence but neither weapon use nor firearm access increases the risk of severity of recidivisim. Journal of Interpersonal Violence, 28, 1143-1156. doi:10.1177/0886260512468232

15.

Foran

H. M.

O’Leary

K. D.

(2008). Alcohol and intimate partner violence: A meta-analytic review. Clinical Psychology Review, 28, 1222-1234.

16.

Furlow

C. F.

Fouladi

R. T.

Gagne

Whittacker

T. A.

(2007). A Monte Carlo study of the impact of missing data and differential item functioning on theta estimates from two polytomous Rasch family models. Journal of Applied Measurement, 8, 388-403.

17.

Gerstenberger

C. B.

Williams

K. R.

(2012). Gender and intimate partner violence: Does dual arrest reveal gender symmetry or asymmetry? Journal of Interpersonal Violence, 28, 1561-1578. doi:10.1177/0886260512468325

18.

Glass

Laughon

Campbell

J. C.

Block

R. B.

Hanson

Sharps

P. S.

et al . (2008). Non-fatal strangulation is an important risk factor for homicide of women. Journal of Emergency Medicine, 35, 329-335. doi:10.1016/j.jemermed.2007.02.065

19.

Hirschel

Buzawa

E. S.

(2013). The impact of offenders leaving the scene on the police decision to arrest in cases of intimate partner violence. Violence Against Women, 19, 1079-1103. doi:10.1177/1077801213501843

20.

Kothari

C. L.

Rhodes

K. V.

(2006). Missed opportunities: Emergency department visits by police-identified victims of intimate partner violence. Annals of Emergency Medicine, 47, 190-199. doi:10.1016/j.annemergmed.2005.10.016

21.

Kyriacou

D. N.

Anglin

Taliaferro

Stone

Tubb

Linden

J. A.

et al . (1999). Risk factors for injury to women from domestic violence against women. New England Journal of Medicine, 341, 1892-1898. doi:10.1056/NEJM199912163412505

22.

Langton

Berzofsky

Krebs

Smiley-McDonald

(2012). Victimizations not reported to the police, 2006-2010. Washington, DC: U.S. Department of Justice. Retrieved from http://www.bjs.gov/content/pub/pdf/vnrp0610.pdf

23.

Max

Rice

D. P.

Finkelstein

Bardwell

R. A.

Leadbetter

(2004). The economic toll of intimate partner violence against women in the United States. Violence and Victims, 19, 259-272.

24.

Melton

H. C.

Sillito

C. L.

(2012). The role of gender in officially reported intimate partner abuse. Journal of Interpersonal Violence, 27, 1090-1111. doi:10.1177/0886260511424498

25.

Miller

S. L.

(2001). The paradox of women arrested for domestic violence: Criminal justice professionals and service providers respond. Violence Against Women, 7, 1339-1376. doi:10.1177/10778010122183900

26.

Muelleman

R. L.

Lenaghan

P. A.

Pakieser

R. A.

(1996). Battered women: Injury locations and severity. Annals of Emergency Medicine, 28, 486-492.

27.

O’Leary

K. D.

Tintle

Bromet

(2014). Risk factors for physical violence against partners in the U.S. Psychology of Violence, 4, 65-77. doi:10.1037/a0034537

28.

Rhodes

K. V.

Dichter

M. E.

Kothari

C. L.

Marcus

S. C.

Cerulli

(2011). The impact of children on legal actions taken by women victims of intimate partner violence. Journal of Family Violence, 26, 355-364. doi:10.1007/s10896-011-9370-8

29.

Rhodes

K. V.

Kothari

C. L.

Dichter

Cerulli

Wiley

Marcus

(2011). Intimate partner violence identification and response: Time for a change in strategy. Journal of General Internal Medicine, 26, 894-899. doi:10.1007/s11606-011-1662-4

30.

Statistics Canada. (1994). Violence against women survey: Public use microdata file documentation and user’s guide. Ottawa, Ontario: Author.

31.

Thompson

M. P.

Saltzman

L. E.

Johnson

(2003). A comparison of risk factors for intimate partner violence-related injury across two national surveys on violence against women. Violence Against Women, 9, 438-457. doi:10.1177/1077801202250955

32.

Tjaden

Thoennes

(2000). Full report of the prevalence, incidence, and consequences of violence against women. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/183781.pdf

33.

Truman

J. L.

Morgan

R. E.

(2014). Nonfatal domestic violence, 2003-2012. Washington, DC: U.S. Department of Justice. Retrieved from http://www.bjs.gov/content/pub/pdf/ndv0312.pdf

34.

United States Census (2010). 2010 Census Interactive Population Search: SC – South Carolina. Retrieved from https://www.census.gov/2010census/popmap/ipmtext.php?fl=45

35.

Walton-Moss

B. J.

Manganell

Frye

Campbell

J. C.

(2005). Risk factors for intimate partner violence and associated injury among urban women. Journal of Community Health, 30, 377-389. doi:10.1007/s10900-005-5518-x