Offense Type and the Arrest Decision in Cases of Intimate Partner Violence

Abstract

Although previous research has examined arrests for intimate partner violence (IPV), most of these analyses focus exclusively on physical assault and intimidation. Research on arrests for sexual assault have examined arrests for cases of stranger and/or acquaintance sexual assault, but have not included sexual IPV. Using data from the 2010 National Incident-Based Reporting System, this analysis is the first to calculate and compare arrest rates for sexual IPV, physical IPV, and intimidation. Results indicate that after controlling for other factors, police are less likely to make an arrest in cases of sexual IPV than in cases of physical IPV or intimidation. These findings are discussed in the context of the consequences of sexual assault on IPV victims.

Keywords

intimate partner violence domestic violence arrest sexual assault

Introduction

Over the past 40 years, activists and advocates have worked to shift domestic violence from a private family matter to one of public concern. One key element of this movement has been the criminalization of domestic violence, including hiring more police, providing more extensive training for officers, removing the spousal exemption for rape, passing mandatory arrest laws, implementing no-drop prosecution policies, and creating criminal no-contact orders and domestic violence civil protection orders. The criminalization of domestic violence was hailed as a “symbolic” victory for women (Chesney-Lind, 2002), and advocates of mandatory and pro-arrest laws argued that their implementation sent a “powerful message” to both batterers and society about the seriousness of domestic violence (Goodmark, 2004, p. 31; see also S. L. Miller, 2001).

One of the more controversial elements of the criminalization of domestic violence is the implementation of mandatory and pro-arrest laws. Mandatory arrest laws require law enforcement to make an arrest when an act of intimate partner violence (IPV) has occurred; pro-arrest laws make arrest the “preferred response” in cases of IPV (Hirschel, Buzawa, Pattavina, & Faggiani, 2007). Since the implementation of mandatory and pro-arrest laws, arrest rates for IPV reported to law enforcement have increased steadily, and arrest rates for IPV are higher in jurisdictions that have mandatory and pro-arrest policies than in jurisdictions that do not (Eitle, 2005; Hirschel et al., 2007; Pattavina, Hirschel, Buzawa, Faggiani, & Bently, 2007; Pattavina, Buzawa, Hirschel, & Faggiani, 2007).

Yet, what is noticeably absent from the discussion of and research on mandatory arrest laws is the fact that the type of violence reported by the victim may influence police decision making. While it is clear that officers are making more arrests in cases of physical abuse since the passage of mandatory and pro-arrest laws, it is unclear whether they are making arrests in cases of sexual assault. The Federal Bureau of Investigation (2008) reports that officers are far less likely to make an arrest in a case of forcible rape than in a case of aggravated assault; however, their data does not allow for the identification of cases of IPV, making it impossible to know whether this is true for victimizations in intimate relationships reported to the police. Previous research examining determinates of arrest in cases of IPV either focus exclusively on physical violence such as intimidation, simple assault, and aggravated assault (Hirschel et al., 2007; Pattavina et al., 2007; Pattavina, Hirschel et al., 2007) or fail to distinguish between sexual assault and physical assault (Eitle, 2005). Furthermore, most research focusing on sexual assault have limited their analyses to those sexual assaults perpetrated by strangers and acquaintances (Temple, Weston, Rodriguez, & Marshall, 2007) and ignored sexual assaults committed by intimate partners.

The only previous research providing estimates of arrest rates for IPV disaggregated by violence using national data comes from the National Violence Against Women Survey (NVAWS). The NVAWS estimates rely on self-report data from victims, which may or may not accurately reflect officer decision making. Even more problematic is that the data were collected in 1995 and 1996 (and referenced victimizations that could have happened long before data collection). In the 1970s, there were no laws criminalizing marital rape in the United States, and it was not until 1993 that marital rape and sexual assault were criminalized in all 50 states (Martin, Taft, & Resick, 2007; Zorza, 1993). Furthermore, since the mid-1990s, many Thus, estimates of arrest rates from 1995 and 1996 are unlikely to be reflective of current patterns of officer decision making. We simply do not know how often police make arrests in reported cases of sexual IPV and whether reported cases of sexual IPV are less likely to result in an arrest than are reported cases of physical IPV or intimidation. If there are differences in the rates of arrest by type of violence, we do not know if these differences are due to the type of offense or are instead due to other characteristics of the victim, offender, and/or incident.

This article examines arrest rates and determinates of arrest in cases of IPV reported to law enforcement, focusing on the role that offense type has on the likelihood of arrest. We first estimate arrest rates for cases of intimidation, sexual assaults, and physical assaults and test whether any differences observed in arrest rates are statistically significant. We then test whether offense type is a significant predictor of arrest after controlling for state arrest policy, incident characteristics, the relationship of the victim and the offender, and demographic characteristics of the victim and the offender. Are officers less likely to make an arrest in a case of sexual IPV than in cases of physical IPV or intimidation?

Reporting Sexual IPV to Police

Before officers can make an arrest, the incident must be reported to law enforcement, and previous research suggests that cases of sexual IPV may be less likely to be reported to the police than are other cases of IPV. Rape is one of the most underreported crimes in the United States (Jordan, 2004), and physical assault survivors are 2 times more likely to report their victimization to police than are survivors of sexual assault (Chen & Ullman, 2010; Felson & Pare, 2007). In addition, women who are sexually assaulted are less likely to report the incident to the police if they know the assailant (J. S. Jones, Alexander, Wynn, Rossman, & Dunnuck, 2009). According to Tjaden and Thoennes (2000), 27% of women who were physically assaulted by an intimate partner reported the assault to police, while only 17% of women who were sexually assaulted by an intimate partner reported the assault. This disparity in reporting holds true for cases of severe IPV; while 70% of the women who experienced “severe” physical abuse (as measured by the Conflict Tactics Scale [CTS-2]) reported their victimization to the police, only 20% of women who had experienced “severe” sexual abuse contacted law enforcement (Bonomi, Holt, Martin, & Thompson, 2006), despite the fact that reporting victimization to law enforcement often has positive outcomes for women. McFarlane et al. (2005) found that women who report sexual assault by an intimate partner to police were significantly less likely to be reassaulted than women who did not report the assault to police.

Sexual assault is heavily stigmatized, and victims of sexual assault often anticipate that they will face negative reactions if they disclose their abuse (Jordan, 2004; Ullman, Starzynski, Long, Mason, & Long, 2008). Some sexual assault victims feel that they experience a “double trauma”—one when they are initially victimized and a second when they report their victimization to the police (Nicholls, 2012). One of the most common reasons why female sexual assault survivors do not report the assault to law enforcement is because they felt that the “police would be insensitive and blame me” (J. S. Jones et al., 2009). Sexual assault victims may also fear reprisal from the perpetrator if they contact the police (Martin et al., 2007). Finally, the determination of culpability for a sexual assault may also be more complex than in cases of physical assault—victims may be more likely to be held responsible in cases of sexual assault than in physical assault.

Arrests for Intimate Partner Violence

Once an incident is reported to law enforcement, officers must make a determination as to whether an arrest is warranted. The arrest policy of a particular state has a large impact on an officer’s decision to arrest in cases of IPV. Most jurisdictions adopted mandatory and pro-arrest policies after the 1984 Minneapolis Domestic Violence Experiment (MDVE). After the results of the MDVE were published, more than one third of all police departments adopted some form of a mandatory or pro-arrest domestic violence policy (Maxwell, Garner, & Fagan, 2001). Mandatory arrest policies require law enforcement to make an arrest if there is evidence that an act of domestic violence has taken place or that there is an imminent threat of physical or sexual harm; pro-arrest policies do not require that law enforcement officers make an arrest but instead explicitly state that arrest is the “preferred” or “encouraged” outcome. Discretionary arrest policies allow law enforcement to make an arrest if there is probable cause that an act of domestic violence has occurred, but do not mandate arrest. As of 2000, all states had authorized warrantless arrests for IPV if there is probable cause that an offense has been committed by the arrestee (N. Miller, 2004).

Since the implementation of mandatory and pro-arrest policies, arrests for IPV have increased dramatically. Hirschel and Buzawa (2002) estimate that arrest rates for IPV in the 1970s and 1980s (before mandatory arrest laws) were 7% to 15%, while studies in the 1990s (after the passage of mandatory and pro-arrest laws) reported rates of 33% to 50% (see also Bourg & Stock, 1994; Eitle, 2005; Hirschel et al., 2007; Jordan, 2004; Mignon & Holmes, 1995; Pattavina et al., 2007; Pattavina, Hirschel et al., 2007). Previous research by Durfee (2012), Pattavina et al. (2007), Pattavina, Hirschel et al., 2007, Hirschel et al. (2007), and Eitle (2005) confirmed that arrest rates were significantly higher in jurisdictions with mandatory and pro-arrest laws as compared with jurisdictions with officer discretion policies.

However, despite the implementation of these policies many incidents of IPV reported to law enforcement do not end in arrest. Tjaden and Thoennes (2000), using self-report data from the NVAWS, found that less than half of the IPV incidents reported to law enforcement resulted in an arrest (see also Coulter, Kuehnle, Byers, & Alfonso, 1999). Other analyses using law enforcement data from 2000 have found that officers made arrests in approximately 50% of reported cases of IPV (Eitle, 2005; Hirschel et al., 2007). It is important to note that arrest rates for IPV are still higher than the arrest rates for assaults by strangers or acquaintances. According to Hirschel et al. (2007), only 35% of assaults by strangers and 29% of assaults by acquaintances resulted in an arrest.

Given that officers are not making arrests in many cases of reported IPV, even in states with mandatory arrest policies, what other factors influence an officer’s decision to arrest or not arrest an offender? Prior research indicates that the relationship between the victim and offender, the demographic characteristics of the victim and the offender (age, race, and gender), and the characteristics of the incident influence the likelihood that an incident will end in an arrest. With regard to relationship status, arrests are more likely when the victim is single, divorced, or separated (Bachman & Coker, 1995) or when the couple is cohabiting (Robinson & Chandek, 2000).

As for demographic characteristics, the likelihood of arrest increases with the victim’s and offender’s age (Eitle, 2005; Hirschel et al., 2007; Pattavina et al., 2007; Pattavina, Hirschel et al., 2007). Furthermore, after controlling for the offender’s race, White victims are more likely (Avakame, Fyfe, & McCoy, 1999) and Black victims are less likely (Eitle, 2005; Ferraro, 1989; Hirschel et al., 2007) to have their offenders arrested. White offenders are less likely to be arrested than are other offenders after controlling for the victim’s race (Hirschel et al., 2007; Pattavina et al., 2007; Pattavina, Hirschel et al., 2007). Similarly, even though there has been much rhetoric about gender and the arrest decision, most research with administrative and law enforcement data indicates that after controlling for other individual and incident characteristics, there is not a statistically significant relationship between gender and arrests for IPV (Eitle, 2005; Hirschel et al., 2007; D. A. Jones & Belknap, 1999; Simpson, Bouffard, Garner, & Hickman, 2006). However, analyses with victim self-report data from the NVAWS suggests that female offenders are treated more leniently than male offenders in cases of IPV (Felson & Pare, 2007; Tjaden & Thoennes, 2000). This disparity with regard to the effects of gender on arrests highlights concerns about exclusively relying on victim self-report data to analyze arrests for sexual versus physical IPV (discussed in the introduction of this article).

Finally, research shows that the characteristics of an incident influence officer decision making. Offenders who use a weapon (Eigenberg, McGuffee, Berry, & Hall, 2003; Eitle, 2005; Felson & Ackerman, 2001; Houry, Reddy, & Parramore, 2006) or injure their victim (Bachman & Coker, 1995) are more likely to be arrested. Incidents where officers believe that either the victim or the offender recently used alcohol and/or drugs are also more likely to result in “police action” and/or the arrest of the offender (Avakame & Fyfe, 2001; Eitle, 2005; D. A. Jones & Belknap, 1999; Mignon & Holmes, 1995; Rajan & McCloskey, 2007; Robinson & Chandek, 2000). Finally, cases of intimidation are less likely to result in an arrest than are cases of simple or aggravated assaults (Hirschel, Hutchinson, & Shaw, 2010; Pattavina et al., 2007; Pattavina, Hirschel et al., 2007).

Yet, we do not know whether cases of sexual IPV are more or less likely to result in arrest than cases of intimidation or physical IPV. Most of the studies cited above have either aggregated physical and sexual IPV or focused exclusively on physical IPV. Analyses of arrests for sexual assault have not distinguished between sexual IPV and sexual violence outside of intimate partnerships. Felson and Pare (2007) compared all cases of assault, including both domestic assault and incidents between strangers and incidents between acquaintances, and found that incidents of rape were more likely to result in an arrest than incidents of physical abuse and other types of sexual assault. Jordan (2004) compared arrests for rapes between intimate partners and strangers and found that offenders who raped their intimate partner were more likely to be arrested than offenders who raped strangers—most likely because the offender could be identified by the victim. However, Jordan did not examine forms of sexual IPV other than rape, nor did Jordan compare arrests for physical IPV and sexual IPV. Thus, we focus here on the role that offense type plays in the likelihood of arrest in cases of IPV after controlling for the state arrest policy, as well as the characteristics of the incident, victim, and offender.

Data and Method

For this analysis, we used the 2010 National Incident-Based Reporting System (NIBRS) Victim Extract. In the NIBRS Victim Extract file, the unit of analysis is the victim. The NIBRS is more appropriate than the Uniform Crime Report (UCR) for analyses of IPV for several reasons. First, with data from the UCR it is not possible to determine the relationship between the victim and the offender, making it impossible to determine which cases are incidents of IPV. Second, unlike the UCR, which contains summary counts of only the most serious offense in a criminal incident, NIBRS captures information on every offense in a criminal incident, allowing for researchers to identify cases of IPV that might otherwise be classified as non-IPV incidents. In addition, the UCR only collects information on Part I crimes: murder and non-negligent manslaughter, forcible rape, robbery, aggravated assault, burglary, larceny-theft, motor vehicle theft, and arson (U.S. Department of Justice, 2000). NIBRS contains information about 46 crimes (and arrest information about an additional 8 crimes), allowing researchers to look at a wider array of IPV offenses, thereby providing a more comprehensive analysis of IPV. For example, cases of intimidation do not appear in the UCR, but they do appear in the NIBRS. In addition to forcible rape, which is identified by the UCR, NIBRS data identifies three other forcible sex offenses—forcible sodomy, sexual assault with an object, and forcible fondling. Finally, NIBRS contains detailed information on the circumstances of the incident, weapon use, victim and offender characteristics, the relationship between the victim and offender, and whether an arrest was made. These are all key factors when attempting a study on arrest decisions in incidents of IPV.

Sample

Our analysis includes the 220,506 cases recorded in NIBRS that have an adult female victim in an incident of IPV—a sexual assault, physical assault, or intimidation (threat of an assault)—where the male offender was a spouse, common-law spouse, ex-spouse, or boyfriend. Due to the structure of the NIBRS, it is possible for a single incident to have multiple victims and multiple offenders. Only those victims in cases where there was a single victim and a single offender were retained. Similarly, because multiple offenses can be recorded for the same incident, it is possible for victims to have experienced multiple forms of victimization during the same incident. We identified 15 cases in which the victim suffered both a physical assault and a sexual assault; these victims were removed from the analyses.

Dependent Variable

Table 1 contains the descriptive statistics for the independent and dependent variables included in this study. The dependent variable for our analysis is whether the offender was arrested for the incident. An arrest could include the offender being taken into custody, an on-view arrest, or the offender receiving a summons/citation. An offender is taken into custody when an arrest is made based on an existing warrant or a previously submitted incident report. An on-view arrest occurs when the offender is taken into custody without a warrant. When an offender receives a summons or citation, that person is not taken into custody, but must appear before a court for the committed offense(s). All three of these options are defined as arrests in NIBRS, and were therefore included in the dependent variable of this study. An arrest for an assaultive offense was made in 46.8% of the 220,506 cases in the sample.

Table 1.

Descriptive Statistics for Study Variables (N = 220,506).

Variable	Frequency	Percent
Arrest (dependent variable)
No arrest charge	117,307	53.2
Arrest for assault offense	103,199	46.8
Offense type
Sexual assault^a	2,530	1.1
Intimidation	33,796	15.3
Physical assault	184,180	83.5
State arrest policy
Discretionary arrest provisions^a	57,504	26.1
Preferred arrest provisions	55,580	25.2
Mandatory—all relationships	8,366	3.8
Mandatory—except dating relationships	99,056	44.9
Weapon use
No weapon^a	48,078	21.8
Firearm	2,169	1.0
Knife	5,710	2.6
Personal weapons (hands, feet, teeth)	151,385	68.7
Miscellaneous weapon	8,272	3.8
Multiple weapons	2,585	1.2
Unknown weapon	2,307	1.0
Substance use by offender
Alcohol	37,347	16.9
Drugs	4,430	2.0
Injury to victim
No injury^a	110,002	49.9
Minor injury	105,154	47.7
Major injury	5,350	2.4
Victim demographics
Age		32 years^b
White^a	146,813	66.6
Black	67,481	30.6
American Indian	1,652	0.7
Asian	1,624	0.7
Unknown race	2,936	1.3
Offender demographics
Age		34 years^b
White^a	133,305	60.5
Black	81,553	37.0
American Indian	1,665	0.8
Asian	1,325	0.6
Unknown race	2,658	1.2
Relationship
Boyfriend/girlfriend^a	129,787	58.9
Spouse	69,671	31.6
Ex-spouse	9,557	4.3
Common-law spouse	11,491	5.2

Reference category.

Mean score.

Covariates

Several covariates of arrest were included in this article. These include different arrest policy types across states, incident variables (such as the type of offense, use of weapons, injuries sustained by the victim, and substance use), demographic variables for both the victim and offender, and the relationship between the victim and the offender. These covariates are listed in Table 1 along with their descriptive statistics. Because of the categorical nature of many of these covariates, dichotomous variables were created. Those covariates with more than two categories had a reference category that was omitted from the analysis.¹

Offense type

The major focus of this article is whether officers are less likely to make an arrest for a sexual assault against an intimate partner compared with a physical assault or an act of intimidation. Therefore, offense type was comprised of the categories of sexual assault, physical assault, or intimidation. Sexual assaults were comprised of the specific offenses of forcible rape, forcible sodomy, sexual assault with an object, and forcible fondling. Physical assaults were comprised of both aggravated assaults and simple assaults. Intimidation offenses comprised its own category. In the sample, 1.1% of the cases were sexual assaults, 15.3% were intimidation offenses, and 83.5% were physical assaults (37.4% were simple assaults and 9.9% were aggravated assaults). In the full logistic regression model, sexual assaults are the reference category.

Arrest policy

Because the NIBRS data used in this study come from 35 states, it is possible that state laws may influence the decision for arrest. Typically, arrest policies concerning violence between domestic partners can take one of three forms—discretionary, preferred, or mandatory. Although local policies may be more restrictive on officer decision making than state policies, no individual jurisdiction may be less restrictive than state policy (i.e., all jurisdictions in a mandatory arrest state must have a mandatory arrest policy; Hirschel, 2008). Thus, our variable captures the least restrictive policy possible in a specific state.

We conducted a thorough review of state arrest policies at the time the data were collected and created a classification system containing four possibilities. We found that mandatory arrest policies could be separated based on the intimate partner relationships included in the arrest policy statute. Some states had mandatory arrest policies that did not include dating relationships such as boyfriend/girlfriend, while other states included all relationships. Thus, four variables were created for arrest policy—discretionary (the reference category), preferred, mandatory (all relationships), and mandatory (except dating). A listing of states by arrest policy type can be found in Table 2. Twelve states (34%) had discretionary arrest policies, 5 states (14%) had preferred arrest policies, 13 states (37%) had mandatory arrest policies except in cases of IPV in dating relationships, and 5 states (14%) had mandatory arrest policies that applied to all cases of IPV.

Table 2.

Type of Arrest Provisions for Domestic Violence by State (35 States).

Discretionary (n = 12)	Preferred (n = 5)	Mandatory	Mandatory
Discretionary (n = 12)	Preferred (n = 5)	Except dating (n = 13)	All relationships (n = 5)
Alabama	Arkansas	Arizona	Maine
Delaware	Massachusetts	Colorado	Oregon
Georgia	Montana	Connecticut	Rhode Island
New Hampshire	North Dakota	Iowa	Washington
Idaho	Tennessee	Kansas	Washington, D.C.
Illinois		Louisiana
Kentucky		Missouri
Michigan		Ohio
Nebraska		South Carolina
Texas		South Dakota
Vermont		Utah
West Virginia		Virginia
		Wisconsin

Most victimizations in the sample (44.9%) occurred in states with mandatory arrest provisions that did not include dating relationships, while the least number of victimizations (3.8%) occurred in jurisdictions with mandatory arrest provisions that included all intimate relationships. The remainder of the sample was equally distributed between jurisdictions with discretionary (26.1%) and preferred arrest provisions (25.2%).

Weapons use

A second incident characteristic included as an independent variable was the use of a weapon in the offense committed against the victim. Due to the nature of NIBRS, it is possible for the officer to record more than one weapon used in an offense. In the Victims Extract file, up to three weapons could have been identified for each offense, and up to three offenses could have been identified for each case. Therefore, it was possible to have a maximum of nine weapons used for each individual case. For the majority of cases in this study (99%), only one type of weapon was used (if at all). The 1% of cases in which more than one weapon had been used were classified as “multiple weapons used.” A dichotomous variable was created for each weapon type used in the study. These weapons variables included no weapon used (21.8%); firearm (1.0%); knife/cutting/blunt instrument (2.6%); personal weapons—hands, feet, teeth, etc. (68.7%); miscellaneous weapons (3.8%); multiple weapons (1.2%); and unknown weapon (1.0%). The reference category for weapon use is no weapon used.

Substance use

The NIBRS data permits for the identification of substance use by the offender. Two variables were created in this study pertaining to suspected substance use by the offender—one for use of alcohol and one for the use of drugs. More incidents had offenders suspected of being under the influence of alcohol (16.9%) than drugs (2.0%).

Victim injury

The NIBRS collects detailed information on any injury that the victim may have suffered. We created three dichotomous injury variables from this information. These included no injury (the reference category), minor injury, and major injury. Major injury included apparent broken bones, possible internal injury, loss of teeth, severe laceration, unconsciousness, and other major injury. In approximately half of the incidents, the victim sustained no injury (49.9%). When victims suffered a physical injury in an incident, the majority suffered only a minor injury or injuries (47.7% of reported incidents); in only 2.4% of incidents did the victim suffer a major injury.

Victim demographics

Victim demographic variables for age and race were included in this study. One of the limitations of the data is that there is no way to determine whether any particular victim or offender was involved in more than one incident. However, as there is only one victim and one offender for each incident, for the ease of interpretation, we will discuss the distributions of the data and the regression results as if each incident had a separate victim and offender. Victim age was measured in years, and the average age for victims was 32 years. Racial categories in NIBRS include White, Black, American Indian, and Asian. Because there were sufficient cases for each racial group, categorical race variables were created for each one separately, with White as the reference category. In the sample, most victims were White (66.6%), followed by Black (30.6%), American Indian (0.7%), and Asian (0.7%). Race was unknown for 2,936 victims (1.3%).

Offender demographics

As with victims, offender demographics were included in the study. The age of the offender, measured in years, was included in the analysis. The average age for offenders in the study was 34, slightly higher than the average age of victims in the study. Categorical race variables were also created for the offender in this study. These included White (60.5%), Black (37.0%), American Indian (0.8%), and Asian (0.6%). Race was unknown for 2,658 of the offenders (1.2%).

Victim–offender relationship

Although this sample includes only victims who were current and former intimate partners of the offenders, we distinguished between specific victim–offender relationships in the analysis. This was done to determine whether there were any differences in arrest based on the nature of the relationship between the victim and offender. Relationship variables included spouse, common-law spouse, boyfriend/girlfriend (the reference category), and ex-spouse. Most of the victims in the sample had a boyfriend/girlfriend relationship (58.9%); the rest were spouses (31.6%), common-law spouses (5.2%), or ex-spouses (4.3%).

Analytic Technique

To determine whether officers are less likely to make an arrest in cases of sexual IPV, we first estimated clearance rates for sexual, physical, and intimidation IPV. We then conducted a chi-square to test whether any observed differences in arrest rates are statistically significant. Finally, we conducted a logistic regression to determine whether offense type is a statistically significant predictor of arrest after controlling for demographic characteristics of the victim and the offender, the relationship between the victim and the offender, incident characteristics, and the state arrest policy.

Results

Descriptive Statistics

Table 3 contains the arrest rates for IPV by offense type. The total number of offenses, total number of arrests, and percent of incidents ending in arrest is shown for each specific offense as well as for the three offense types.

Table 3.

Arrest Rates for Intimate Partner Violence by Offense Type (N = 220,506).

Top offense	Total incidents	Number of arrests	% of incidents ending in arrest
Sexual assault	2,530	663	26.2
Forcible rape	2,060	532	25.8
Forcible sodomy	158	34	21.5
Forcible fondling	312	97	31.1
Physical assault	184,180	94,948	51.6
Aggravated assault	21,723	12,369	56.9
Simple assault	162,457	82,579	50.8
Intimidation	33,796	7,588	22.5
Total	220,506	103,199	46.8

Physical assaults were found to have the highest arrest rate of 51.6%. Within this offense type, aggravated assaults had an arrest rate higher than that of simple assaults (56.9% vs. 50.8%). Sexual assaults had the second highest arrest rate of 26.2%, which was less than half of that of physical assaults but more than the arrest rate for intimidations (22.5%). Based on these descriptive statistics, it appears that officers are least likely to make arrests in cases of intimidation and most likely to make arrests in cases of physical assault.

Chi-Square Test

To test whether the observed differences in arrest rates were statistically significant, we conducted a chi-square test. Table 4 shows the results of the contingency table analysis with the chi-square test. The results show that physical assaults between intimate partners are more likely to end in an arrest than are sexual assaults or intimidation. Sexual assaults and intimidations are more likely to end in no arrest than an arrest. These differences are statistically significant at the .001 level. It is possible, however, that these differences are due to characteristics of the victim, the offender, the incident, or the arrest policy of the particular state in which the incident occurred.

Table 4.

Contingency Table for Offense Type by Arrest Decision (N = 220,506).

	Decision to arrest
Offense type	No	Yes	Total
Sexual assault	1,867 (73.8%)	663 (26.2%)	2,530 (100.0)
Physical assault	89,232 (48.4%)	94,948 (51.6%)	184,180 (100.0%)
Intimidation	26,208 (77.5%)	7,588 (22.5%)	33,796 (100.0%)
Total	117,309 (53.2%)	103,200 (46.8%)	220,509 (100.0%)

Note. χ² = 10,148.02; df (2); p < .001. Likelihood ratio = 10,724.47; df (2); p < .001.

Logistic Regression Model

To determine whether differences in arrest rates are due to these factors, we use logistic regression. After controlling for victim, offender, incident, and policy-level factors, we find that officers are less likely to make an arrest in cases of sexual assault than in cases of physical assault or intimidation. As shown in Table 5, the odds of an arrest for a case of physical assault are greater than that for a sexual assault (OR = 2.920, p < .001). Cases of intimidation are also more likely to result in an arrest than are cases of sexual assault between intimate partners (OR = 1.310, p < .001).

Table 5.

Logistic Regression Model for Arrests in Incidents of Intimate Partner Violence Reported to Police (N = 220,506).

Variable	B	SE	z	Odds ratio	%
Offense type
Intimidation	0.270***	0.051	5.30	1.310***	31.0
Physical assault	1.072***	0.047	22.87	2.920***	192.0
State arrest policy
Preferred arrest provisions	0.099***	0.013	7.72	1.104***	10.4
Mandatory—All relationships	0.511***	0.025	20.21	1.667***	66.7
Mandatory—Except dating relationships	0.192***	0.011	17.11	1.212***	21.2
Weapon use
Firearm	0.306***	0.048	6.33	1.358***	35.8
Knife	0.450***	0.033	13.47	1.568***	56.8
Personal weapons (hands, feet, teeth)	0.134***	0.019	7.21	1.143***	14.3
Miscellaneous weapon	0.145***	0.029	5.02	1.157***	15.7
Multiple weapons	0.320***	0.045	7.04	1.377***	37.7
Unknown weapon	0.006	0.047	0.13	1.006	0.6
Substance use by offender
Alcohol	0.478***	0.012	38.41	1.613***	61.3
Drugs	−0.268***	0.032	−8.24	0.765***	−23.5
Injury to victim
Minor injury	0.652***	0.010	64.99	1.919***	91.9
Major injury	0.706***	0.030	23.22	2.025***	102.5
Victim demographics
Age	0.007***	0.001	8.97	1.007***	0.7
Black	−0.425***	0.017	−25.36	0.654***	−34.6
American Indian	−0.091	0.063	−1.44	0.913	−8.7
Asian	−0.012	0.061	−0.20	0.988	−1.2
Unknown race	0.256***	0.041	6.21	1.292***	29.2
Offender demographics
Age	0.001	0.001	1.21	1.001	0.1
Black	−0.213***	0.016	−13.22	0.809***	−19.1
American Indian	−0.037	0.063	−0.60	0.963	−3.7
Asian	0.092	0.068	1.36	1.096	9.6
Unknown race	−0.471***	0.046	−10.27	0.625***	−37.5
Relationship
Spouse	0.234***	0.011	22.20	1.263***	26.3
Ex-spouse	−0.340***	0.025	−13.67	0.712***	−28.8
Common-law spouse	0.091***	0.021	4.40	1.095***	9.5

Note. The log odds were transformed to the percentage change in the odds of the dependent variable with the formula: 100 [exp(log odds) − 1].

***

p < .001.

As predicted, cases in jurisdictions with mandatory and preferred arrest policies were more likely to result in arrest than in jurisdictions with officer discretion statutes (the reference category). After controlling for all other variables, the odds of an arrest were 10% greater in jurisdictions with preferred arrest provisions for IPV than jurisdictions with discretionary policies (OR = 1.104, p < .001). The odds were even greater for jurisdictions with mandatory arrest provisions excluding dating relationships (OR = 1.212, p < .001), and greatest when mandatory arrest provisions included all intimate partner relationships (OR = 1.667, p < .001). Thus, mandatory and preferred arrest laws appear to have the desired effect of increasing the number of arrests for IPV.

Several characteristics of the incident, the victim, and the offender were also statistically significant. Overall, our analysis shows that an arrest is significantly more likely when a weapon is used than if no weapon is used. In comparison with not using any weapon, the odds of an arrest are 36% greater when the offender used a firearm (OR = 1.358, p < .001); 57% greater when an offender used a knife (OR = 1.568, p < .001); 14% greater when offender used his hands, feet, or teeth (OR = 1.143, p < .001); 16% greater when a miscellaneous weapon was used (OR = 1.157, p < .001); and 38% more likely when multiple weapons were used (OR = 1.377, p < .001).

Offenders who were suspected of being under the influence of alcohol had a significantly greater likelihood of an arrest for an assault-related offense in comparison with offenders that had not been drinking (OR = 1.613, p < .001). However, if the offender was suspected of having been using drugs, the odds of being arrested were actually 24% lower than for an offender not suspected of using drugs (OR = .765, p < .001). Injury to the victim was also related to the arrest decision. Incidents where the victim suffered a minor injury (OR = 1.919, p < .001) or a major injury (OR = 2.025, p < .001) were more likely to result in an arrest than incident where the victim was not injured.

The age of the victim had a statistically significant relationship with decisions to arrest. Although not strong in magnitude, each 1 year increase in the age of the victim increased the odds of an arrest by almost 1% (OR = 1.007, p < .001). As to victim race, incidents with Black victims are less likely to result in an arrest (OR = 0.654, p < .001) and incidents with victims of an unknown race are more likely to result in an arrest (OR = 1.292, p < .001) than are incidents with White victims. Similarly, the odds of an arrest of a Black offender are 19% lower (OR = 0.809, p < .001) and the odds of an arrest of an offender of an unknown race are 38% lower (OR = 0.625, p < .001) than that of a White offender.

Our results show that the outcome of an arrest is more apt to occur when the victim suffers any type of injury, whether it is minor or major. Controlling for all other variables, the odds of an arrest are 92% greater when the victim has visible minor injuries than when the victim suffers no injury. When the victim suffers any type of major injury from an offense, the odds of an arrest increase by a factor of two (OR = 2.025, p < .001). This highlights the importance of controlling for visible injuries when comparing arrest rates for sexual IPV with arrest rates for physical IPV, as victims of sexual assaults may be less likely to have injuries to parts of the body visible to officers and/or may be unwilling to show officers their injuries.

Using the relationship of boyfriend/girlfriend as the reference category and holding all other variables constant, the other victim-to-offender relationships in the study have statistically significant relationships with arrest decisions for IPV. Being married increases the odds of an arrest in an assault, despite whether the marriage was official or common law. If the victim was the offender’s spouse, the odds of the offender being arrested were 26% greater than if they had only been dating (OR = 1.263, p < .001). Likewise, the odds of an arrest occurring were about 10% greater if the victim-to-offender relationship was that of common-law spouse compared with boyfriend/girlfriend (OR = 1.095, p < .001). If the couple was divorced, the odds of an arrest occurring actually were lower than if the couple were dating (OR = 0.712, p < .001).

Discussion

There has been scant research on sexual assault and arrests for IPV, and what has been done often leaves more questions than answers. There has also been little discussion among policy makers, activists, and the advocacy community about law enforcement, arrest policies, and officer decision making in cases of sexual IPV. The main purpose of this study was to determine how often police make arrests in cases of sexual IPV and whether officers were less likely to make arrests in cases of sexual IPV than in cases of physical IPV or intimidation.

Findings from this study showed that there are statistically significant differences in the likelihood of arrest depending on the type of offense, and the odds of arrest are greater when the offense is a physical assault than if the assault is sexual in nature. While 51.6% cases of physical IPV resulted in an arrest, only 26.2% of reported sexual IPV result in an arrest. This disparity holds true even when we control for other incident and demographic variables—the odds of an arrest for physical IPV is 2.92 times that of sexual IPV. The differences in arrest rates for physical and sexual IPV are consistent with differences in arrest rates for all forcible rapes and aggravated assaults—cases of forcible rapes reported to the police are less likely to result in an arrest than are aggravated assaults reported to the police (Federal Bureau of Investigation, 2008). Although cases of intimidation have the lowest arrest rate of all three offense types (22.5%), the results of the full logistic regression model show that after controlling for incident, victim, and offender characteristics and the state arrest policy, cases of intimidation are far more likely to result in an arrest than cases of sexual IPV (OR = 1.310, p < .001).

That reported sexual IPV is less likely than reported physical IPV to result in an arrest is contrary to the findings of Felson and Pare (2007). In their analysis of data from the NVAWS, they reported that rapes and other sexual assaults reported to law enforcement were more likely to result in an arrest than were physical assaults. The difference may be that their analysis uses victim self-report data while this analysis uses law enforcement data. Their sample is also broader than our more restrictive sample, and included crimes committed by intimates, non-strangers, and strangers. While they added several variables to their model to indicate the relationship between the victim and the offender, the structure of their analysis does not allow them to directly compare rates of arrest for different types of IPV controlling for victim, offender, and incident characteristics. Restricting our sample exclusively to cases of IPV allows us to test differences in arrest rates for IPV not possible with their broader sample.

The relatively low arrest rates for sexual IPV may have to do with the nature of the crime, the stigma associated with sexual assault, and the types of injuries inflicted during a sexual assault. Because of the stigma of sexual assault, victims often do not call the police for assistance. Victims may anticipate that others may not believe that the assault occurred or may believe that the sexual contact was consensual and not forced, that the assault was the victim’s fault, that sexual victimization is inconsequential compared with physical victimization, that it is not possible to “rape” one’s intimate partner, that “good” women are not sexually assaulted, or that the victim has ulterior motives and is trying to manipulate or “get back” at the alleged assailant. Victims may also be skeptical about the ability of the police to take any meaningful action or the state to prosecute the case. Furthermore, victims are often fearful of reprisal by the perpetrator if they contact the police. Finally, victims may not feel that a sexual victimization is IPV—that IPV is limited to physical abuse. However, all of the cases of IPV in this sample were reported to law enforcement, so the different rates of reporting physical IPV and sexual IPV cannot be the cause of the lower arrest rate for incidents of sexual IPV included in this study.

A second explanation for the low arrest rates for sexual IPV may be that the injuries sustained by victims during a sexual assault are very different from those sustained in a physical assault. McFarlane et al. (2005) documented that victims of sexual IPV often have severe gynecological injuries, and these injuries are not easily seen by officers investigating incidents of IPV, which may lead to fewer arrests for sexual assaults. When victims are hesitant to discuss their victimization or show officers their injuries, officers may not have the evidence to support an arrest. As one sexual assault victim told Wolf, Ly, Hobart, and Kernic (2003),

Well, I was raped by my husband. There was no evidence except for bruises on the inside of my legs or the pain on my breasts, and you just can’t prove it . . . there was no way I was going to have some man come to my house that I didn’t know and tell him that my husband had raped me and show him bruises . . . I’d have to pull down my pants. (p. 123)

A victim’s hesitation to describe in detail a sexual assault or show officers her injuries is compounded by efforts by perpetrators to convince officers that the victim is lying about the assault (DeJong, Burgess-Proctor, & Elis, 2008). However, there was enough evidence that a sexual assault occurred that the responding officers categorized the incident as a sexual assault in their records (and thus the incident is coded in NIBRS as a sexual assault). Furthermore, arrests were more likely for cases of intimidation than for sexual IPV. Given that by definition there is no physical contact between an offender and victim in cases of intimidation, the lower arrest rates for sexual IPV cannot be solely attributed to a victim’s reluctance to show the physical injuries inflicted during a sexual assault. It is possible that the officers believed that a sexual assault occurred but felt that they did not have sufficient evidence to support an arrest (and/or for a prosecutor to file charges)—thus they took a report but did not make an arrest.

The most obvious explanation for the marked difference in arrest rates is that officers consider sexual IPV to be a less serious offense than physical IPV or intimidation. This suggests that efforts to educate police about sexual IPV have not been as effective as education about physical IPV and intimidation. Officers should be trained to recognize the signs of sexual IPV, educated about the severity of sexual IPV and the detrimental impacts that it has on victims, and taught how to interview victims in such a way that victims feel comfortable describing their abuse. Partnerships with local agencies that specialize in working with victims of sexual IPV would be beneficial not only to victims but also to officers who may not have the skills and/or resources to respond effectively to cases of sexual IPV.

It is important to understand the arrest decision in cases of sexual IPV as research suggests that sexual assault by a current or past intimate partner is more traumatic than sexual assault by other individuals, and is a significant predictor of health problems, including post-traumatic stress disorder (PTSD), stress, and dissociation (Temple et al., 2007). Victims who are sexually assaulted by intimate partners are more likely to have severe injuries than victims of stranger rape, are at greater risk of homicide than are victims of physical IPV, and attempted suicide rates are 5 times higher for victims of sexual IPV than for victims of physical IPV (McFarlane et al., 2005). In addition, the likelihood of reassault is significantly higher than that for victims assaulted by an intimate partner as compared with non-partners, likely because the victim is often living with the abusive partner. Tjaden and Thoennes (2000) found that women sexually assaulted by an intimate partner were, on average, assaulted by that partner 4.5 times, and that 15% of female sexual assault victims reported that they had been sexually victimized more than 10 times.

There are some important limitations of this analysis. First, the sample only includes female victims and male perpetrators. We chose to restrict our analysis by gender as at the time the data were collected (2010), the FBI’s definition of rape limited it to action perpetrated against women. Thus, sexual assaults against men were not classified as rape. Given the inherently problematic nature of this definition, we opted to exclude male victims from the analysis. The FBI expanded their definition of rape and sexual assault in 2012 to include assaults against men (Federal Bureau of Investigation, 2013). However, the new definition did not go into effect until January 1, 2013; thus, analyses including incidents with male victims are not possible with data collected prior to 2013.

Second, the relationship categories in the NIBRS also limit this analysis, especially given state variations in legal definitions of intimate relationships. In 2010, the intimate relationships specified by the NIBRS included spouse, “common-law” spouse,² ex-spouse, boyfriend/girlfriend, and homosexual relationship. As the data are collected by law enforcement and reported to the FBI, the data utilizes legal definitions of relationships. Differences between states in the legal definitions of marriage and “common-law” relationships can result in the same couple being classified as different relationship types in different states. A couple who would be classified as having a “common-law” relationship in one state may be classified as “boyfriend/girlfriend” in another. There is also no category for “ex-boyfriend/ex-girlfriend” in the 2010 NIBRS, which may lead to an underreporting of assaults as “domestic violence” and/or the misreporting of former intimate relationships as current intimate relationships. In 2010, same-sex couples could be spouses, common-law spouses, or classified as a “homosexual relationship” (depending on the state in which the incident occurred). Furthermore, in some states in 2010, violence in same-sex relationships was specifically excluded by law from being considered “domestic violence” (and was thus exempt from mandatory and pro-arrest policies). Thus, we excluded same-sex couples from the analysis. Future research focusing on the role of relationship status in the arrest decision and research on violence in same-sex relationships will need to account for these exclusions, inclusions, and changing definitions of marriage in their analyses.

Third, there are certainly other factors influencing officer decision making that are not included in the NIBRS, such as the victim’s preference for arrest and prior police contact (with either the victim or the offender). We do not claim that the variables included in these models are the only factors that officers use to determine whether or not to arrest an offender, or that officers rely on different factors when deciding to arrest in cases of physical IPV than in cases of sexual IPV and/or intimidation. What we emphasize is that observed differences in arrest rates between physical IPV and sexual IPV remain despite what, from a careful review of prior research, is likely to account for those differences. Furthermore, when we include incident, victim, offender, and relationship variables in the logistic regression model, a significant difference emerges in the likelihood of arrest in cases of sexual violence as compared with intimidation. The results of our analysis should serve as the foundation for future research using multiple methods and data sources to see what might account for these differences—not as the definitive answer as to what accounts for them. As Maltz (1994) notes, we need to “collect different kinds of data, using more relevant categories” and “extract data from the richer sources that are at our disposal” instead of simply relying on the results of regression models and significance testing (Maltz, 1994, p. 457).

By focusing on arrest rates for IPV, we are not implying that arrest is the best option in all cases of IPV-reported law enforcement. Evidence of the efficacy of mandatory arrest policies is mixed, with some studies showing that arrest decreases recidivism, and other studies showing that arrest actually increases recidivism of IPV (Maxwell et al., 2001). However, given the significant negative effects that sexual IPV has on victims, and the decreased rates of recidivism associated with reporting intimate partner sexual assault, we need to better understand police decision making in these types of cases. Future research should also explore the relationship between sexual IPV and other legal resources for victims of IPV, such as domestic violence civil protection orders. Given that the receipt of a protection order can reduce recidivism as effectively (if not more effectively) as an arrest (McFarlane et al., 2005), ensuring access to protection orders for survivors of sexual IPV is critically important.

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) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Alesha Durfee is an associate professor in the School of Social Transformation at Arizona State University. She has a PhD in sociology from the University of Washington. Her research interests include intimate partner violence (IPV), the use of the civil legal and criminal justice system by IPV survivors, and the construction of narratives of violence by IPV survivors and abusers.

Matthew D. Fetzer is an assistant professor of criminal justice at Shippensburg University. He has a PhD in criminal justice from the University at Albany. His research interests include juvenile justice, homicide, hate crime, and domestic violence.

References

Avakame

E. F.

Fyfe

J. J.

(2001). Differential police treatment of male-on-female spousal violence: Additional evidence on the leniency thesis. Violence Against Women, 7, 1, 22-45.

Avakame

E. F.

Fyfe

J. J.

McCoy

(1999). “Did you call the police? What did they do?” An empirical assessment of Black’s theory of mobilization of law. Justice Quarterly, 16, 765-792.

Bachman

Coker

A. L.

(1995). Police involvement in domestic violence: The interactive effects of victim injury, offender’s history of violence, and race. Violence and Victims, 10, 91-106.

Bonomi

A. E.

Holt

V. L.

Martin

D. P.

Thompson

R. S.

(2006). Severity of intimate partner violence and occurrence and frequency of police calls. Journal of Interpersonal Violence, 21, 1354-1364.

Bourg

Stock

H. V.

(1994). A review of domestic violence arrest statistics in a police department using a pro-arrest policy: Are pro-arrest policies enough? Journal of Family Violence, 9, 177-189.

Chen

Ullman

S. E.

(2010). Women’s reporting of sexual and physical assaults to police in the National Violence Against Women Survey. Violence Against Women, 16, 262-279.

Chesney-Lind

(2002). Criminalizing victimization: The unintended consequences of pro-arrest policies for girls and women. Criminology & Public Policy, 82, 81-90.

Coulter

Kuehnle

Byers

Alfonso

(1999). Police-reporting behavior and victim-police interactions as described by women in a domestic violence shelter. Journal of Interpersonal Violence, 14, 1290-1298.

DeJong

Burgess-Proctor

Elis

(2008). Police officer perceptions of intimate partner violence: An analysis of observational data. Violence and Victims, 23, 683-696.

10.

Durfee

(2012). Situational Ambiguity and Gendered Patterns of Arrest for Intimate Partner Violence. Violence Against Women, 18, 64-84.

11.

Eigenberg

McGuffee

Berry

Hall

W. H.

(2003). Protective order legislation: Trends in state statutes. Journal of Criminal Justice, 31, 411-422.

12.

Eitle

(2005). The influence of mandatory arrest policies, police organizational characteristics, and situational variables on the probability of arrest in domestic violence cases. Crime & Delinquency, 51, 573-597.

13.

Federal Bureau of Investigation. (2008). Crime in the United States 2007. Washington, DC: Government Printing Office.

14.

Federal Bureau of Investigation. (2013). Frequently asked questions about the change in the UCR definition of rape. Retrieved from http://www.fbi.gov/about-us/cjis/ucr/recent-program-updates/new-rape-definition-frequently-asked-questions

15.

Felson

R. B.

Ackerman

(2001). Arrests for domestic and other assaults. Criminology, 39, 655-676.

16.

Felson

R. B.

Pare

(2007). Does the criminal justice system treat domestic violence and sexual assault offenders leniently? Justice Quarterly, 24, 435-459.

17.

Ferraro

K. J.

(1989). Policing woman battering. Social Problems, 36, 61-74.

18.

Goodmark

(2004). Law is the answer? Do we know that for sure? Questioning the efficacy of legal interventions for battered women. Saint Louis University Public Law Review, 23, 7-48.

19.

Hirschel

(2008). Domestic violence cases: What research shows about arrest and dual arrest rates. Washington, DC: National Institute of Justice. Retrieved from http://www.nij.gov/publications/dv-dual-arrest-222679

20.

Hirschel

Buzawa

(2002). Understanding the context of dual arrest with directions for future research. Violence Against Women, 8, 1449-1473.

21.

Hirschel

Buzawa

Pattavina

Faggiani

(2007). Domestic violence and mandatory arrest laws: To what extent do they influence police arrest decisions? Journal of Criminal Law & Criminology, 98, 254-298.

22.

Hirschel

Hutchinson

I. W.

Shaw

(2010). The interrelationship between substance abuse and the likelihood of arrest, conviction, and re-offending in cases of intimate partner violence. Journal of Family Violence, 25, 81-90.

23.

Houry

Debra

Reddy

Sudha

Parramore

Constance

(2006). Characteristics of victims coarrested for intimate partner violence. Journal of Interpersonal Violence, 21, 1483-1492.

24.

Jones

D. A.

Belknap

(1999). Police responses to battering in a progressive pro-arrest jurisdiction. Justice Quarterly, 16, 249-273.

25.

Jones

J. S.

Alexander

Wynn

B. N.

Rossman

Dunnuck

(2009). Why women don’t report sexual assault to the police: The influence of psychosocial variables and traumatic injury. Journal of Emergency Medicine, 36, 417-424.

26.

Jordan

C. E.

(2004). Intimate partner violence and the justice system: An examination of the interface. Journal of Interpersonal Violence, 19, 1412-1434.

27.

Maltz

M. D.

(1994). Deviating from the mean: The declining significance of significance. Journal of Research in Crime & Delinquency, 31, 434-463.

28.

Martin

E. K.

Taft

C. T.

Resick

P. A.

(2007). A review of marital rape. Aggression and Violent Behavior, 12, 329-347.

29.

Maxwell

C. D.

Garner

J. H.

Fagan

J. A.

(2001). The effects of arrest on intimate partner violence: New Evidence from the Spouse Assault Replication Program (NCJ 188199). Washington, DC: National Institute of Justice Research in Brief.

30.

McFarlane

Maleche

Watson

Gist

Batten

Hall

Smith

(2005). Intimate partner sexual assault against women: Frequency, health consequences, and treatment outcomes. Obstetrics & Gynecology, 105, 99-108.

31.

Mignon

S. I.

Holmes

W. M.

(1995). Police response to mandatory arrest laws. Crime & Delinquency, 41, 430-442.

32.

Miller

(2004). Domestic violence: A review of state legislation defining police and prosecution duties and powers. Retrieved from http://www.ilj.org/publications/docs/Domestic_Violence_Legislation.pdf

33.

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.

34.

Nicholls

C. M.

(2012). Survivors’ stories: Personal experiences of sexual abuse and violence. London, England: NatCen Social Research.

35.

Pattavina

Buzawa

Hirschel

Faggiani

(2007). Policy, place, and perpetrators: Using NIBRS to explain arrest practices in intimate partner violence. Justice Research and Policy, 9, 31-51.

36.

Pattavina

Hirschel

Buzawa

Faggiani

Bentley

(2007). A comparison of the police response to heterosexual versus same-sex intimate partner violence. Violence Against Women, 13, 374-394.

37.

Rajan

McCloskey

K. A.

(2007). Victims of intimate partner violence: Arrest rates across recent studies. Journal of Aggression, Maltreatment & Trauma, 15, 27-52.

38.

Robinson

A. L.

Chandek

M. S.

(2000). The domestic violence arrest decision: Examining demographic, attitudinal, and situational variables. Crime & Delinquency, 46, 18-37.

39.

Simpson

S. S.

Bouffard

L. A.

Garner

Hickman

(2006). The influence of legal reform on the probability of arrest in domestic violence cases. Justice Quarterly, 23, 297-316.

40.

Temple

J. R.

Weston

Rodriguez

B. F.

Marshall

L. L.

(2007). Differing effects of partner and nonpartner sexual assault on women’s mental health. Violence Against Women, 13, 285-297.

41.

Tjaden

Thoennes

(2000). Extent, nature, and consequences of intimate partner violence. Washington, DC: National Institute of Justice and Centers for Disease Control.

42.

Ullman

S. E.

Starzynski

L. L.

Long

S. M.

Mason

G. E.

Long

L. M.

(2008). Exploring the relationships of women’s sexual assault disclosure. Journal of Interpersonal Violence, 23, 1235-1257.

43.

U.S. Department of Justice. (2000). National Incident-Based Reporting System, Volume 1: Data Collection Guidelines.

44.

Wolf

M. E.

Hobart

M. A.

Kernic

M. A.

(2003). Barriers to seeking police help for intimate partner violence. Journal of Family Violence, 18, 121-129.

45.

Zorza

(1993). Must we stop arresting batterers: Analysis and policy implications of new police domestic violence studies. New England Law Review, 28, 929.