What Makes Police Officers Act? Police Response to Instances of Intimate Partner Violence

Abstract

Using data from the 2009 General Social Survey (GSS Cycle 23: Victimization main file), this study assessed the relationship between individual and neighborhood-level factors and police response to spousal violence in Canada. A total of 890 participants in the GSS reported experiences of spousal violence within the previous 5 years, with approximately 22% of these victims reporting direct contact with the police because of the violence. Among this subsample of victims who had police contact, we evaluated the extent to which individual-level factors (sociodemographic and violence characteristics) and neighborhood-level factors (perceptions of social disorder and the presence of police facilities in one’s community) were related to distinct types of police response to reported spousal violence. Regression analyses suggest variations in police response based on sociodemographic, violence, and neighborhood characteristics; however, overall satisfaction with police actions taken were not found to be significantly different among victims based on these characteristics. Implications are relevant for policing practice as results suggest that different victims may have different safety needs and abilities to communicate these needs to the police.

Keywords

spousal violence intimate partner violence police contact police response law enforcement

Introduction

Despite substantial changes to the Canadian Criminal Justice System’s response to intimate partner violence (IPV), as well as major shifts in social intervention and public understanding and awareness of spousal violence over the last three decades (Burczycka, 2016; Sinha, 2012), IPV currently accounts for approximately 25% of violent crimes reported to the police in Canada annually (Sinha, 2013). Coupled with the finding that IPV and especially spousal violence is substantially underreported (Burczycka, 2016; Wolf, Ly, Hobart, & Kernic, 2003), it is evident that IPV represents a pervasive social and public health issue with major consequences not only for the direct victims, but also for families, communities, and society (Burczycka, 2016; Sinha, 2013; World Health Organization, 2013). In a detailed analysis examining the fiscal impact of spousal violence in Canada in 2009, Zhang, Hoddenbagh, McDonald, and Scrim (2012) estimated $7.4 billion CDN dollars in justice system, victim, and third party–related costs. As a result, the widespread impact of IPV in Canada cannot be underscored enough.

Spousal violence is a particular form of IPV that is defined specifically as physical and/or sexual violence that is perpetrated by a current or former spouse or common-law partner (Burczycka, 2016). Like other forms of IPV, spousal violence is a criminal offense subject to prosecution in Canada. While the physical acts of spousal violence are explicitly outlined in the Canadian Criminal Code, the Canadian Federal government, as well as many of the various provincial and territorial governments have taken further action in recognizing IPV and more specifically spousal violence as a unique form of violence. This has resulted in a series of specialized and targeted approaches designed to improve the criminal justice system’s response (Department of Justice Canada, 2003). As frontline officials of the criminal justice system, and those who are first to intervene, the police have an integral role when responding to reported instances of spousal violence (Stith, 1990).

While there is research exploring the criminal justice system’s response and specifically police response to IPV (Barrett, St. Pierre, & Vallaincourt, 2011; Buzawa & Austin, 1993; Dawson & Hotton, 2014; Leisenring, 2012; Millar & Brown, 2010; Robinson & Chandek, 2000a, 2000b; Wolf et al., 2003), research specifically exploring police response to spousal violence in a Canadian context is incomplete, in that the majority of studies focus solely on female victims of spousal violence (Barrett et al., 2011; Millar & Brown, 2010). In contributing to the research on spousal violence and police response in Canada, the current study sought to answer the primary research question: What factors are associated with police officers’ response to cases of spousal violence?

This study is novel in that it utilizes a nationally representative sample of male and female Canadians to simultaneously explore the relationship between individual, violence, and neighborhood characteristics and police response to spousal violence, as well as the relationship of these factors with victim’s satisfaction ratings of the police response.

Police Response to IPV

Although police officers are most equipped to provide immediate protection to victims (Apsler, Cummins, & Carl, 2003; Stith, 1990), and are the “gate keepers” to access resources and services for victims of IPV (Robinson & Chandek, 2000a; Russell & Light, 2006), spousal violence is grossly underreported to the police (Burczycka, 2016; Wolf et al., 2003). Underreporting of spousal violence has been tied to a series of both personal and structural barriers including previous negative experiences with the police (Wolf et al., 2003), fear of possible repercussions or retaliation by a partner (Tjaden & Thoennes, 2000; Wolf et al., 2003), and the perception or belief that the criminal justice system cannot help, that it is too minor of an incident, or that spousal violence is a personal matter (Wolf et al., 2003). On the contrary, research has reported that victims are more likely to contact the police if the violence involves the use of a weapon, they experience an injury, and/or they fear for their lives (Bachman & Coker, 1995; Barrett et al., 2011; Davies, Block, & Campbell, 2007; Felson, Messner, Hoskin, & Deane, 2002; Lee, Park, & Lightfoot, 2010), with the most commonly reported reason for contacting the police being to stop the violence or obtain protection (Barrett et al., 2011; Burczycka, 2016). Thus, despite low disclosure and official reporting rates, victims of spousal violence who come into contact with the police are often in need of immediate protection, further stressing the crucial role police have in cases of spousal violence.

Once police are notified of an incident of spousal violence, their response and whether they act or provide intervention affects not only a victim’s evaluations and satisfaction ratings of the police, but it also can influence a victim’s likelihood of reporting future instances of spousal violence to the police (Leisenring, 2012; Wolf et al., 2003). Canadian researchers Barrett et al. (2011) found that among female victims of spousal violence, more than half of victims with police contact indicated the police visited the scene, and/or took a report or made an investigation, while just less than one in three victims reported the police removed their partner. Less common interventions included issuing a warning, referring the victim to community resources, and taking no action at all (Barrett et al., 2011). Beyond these types of descriptive results suggesting variations in police response, and a seemingly relatively low frequency of police action, multivariate analyses suggest variations in police response to spousal violence in both the American and Canadian literature. More specifically, the small body of literature suggests that how police respond to cases of spousal violence will vary based on individual factors such as sociodemographic characteristics of victim, perpetrator, and/or responding police officer (Dawson & Hotton, 2014; Millar & Brown, 2010; Pattavina, Hirschel, Buzawa, Faggiani, & Bentley, 2007; Robinson & Chandek, 2000b); incident or situational factors such as violence characteristics (Bachman & Coker, 1995; Barrett et al., 2011; Belknap, 1995; Buzawa & Austin, 1993; Dawson & Hotton, 2014; Dichter, Marcus, Morabito, & Rhodes, 2011; Hall, 2005; Millar & Brown, 2010; Robinson & Chandek, 2000b); and neighborhood-level characteristics (Dichter et al., 2011).

The relationship between individual-level factors and police response to instances of spousal violence varies substantially in the literature. While some researchers report no relationship between victim and/or offender gender and the likelihood of arrest (Eitle, 2005; Robinson & Chandek, 2000b), others have reported that males were more likely than females to be arrested (Dawson & Hotton, 2014; Millar & Brown, 2010). Similarly, while some researchers have suggested that arrest is more likely if the victim is White (Eitle, 2005; Ferraro, 1989; Smith, 1987), others have found that police were more likely to remove the offending party if victims were visible minorities or Aboriginal, and to take a report and/or make an investigation if victims were born outside of Canada, or self-identified as Aboriginal (Barrett et al., 2011). Younger victims have been found to be less likely to report the police visited the scene, took a report/investigated, and issued a warning in cases of spousal violence (Barrett et al., 2011), with police less likely to arrest older perpetrators (Robinson & Chandek, 2000a). Other findings suggest a negative relationship between victims’ income and the likelihood of police removing the offending party (Barrett et al., 2011) and a positive relationship between police response of arrest and rural residence (Pattavina et al., 2007).

Incident- or situational-level factors such as violence characteristics have also been identified in the literature as related to police response to spousal violence. Among female victims, results suggest that while multiple incidents of violence were found to be related to a higher likelihood of police taking a report and/or investigating, multiple incidents of violence were negatively related to the police response of taking the partner away (Barrett et al., 2011). However, in a study of both male and female victims of spousal violence, repeat calls were more likely to result in the arrest of males, and/or a dual charge of both parties (Millar & Brown, 2010). Victim injury has been found to be related to a significantly higher likelihood of police action (Burczycka, 2016), including making an arrest (Bachman & Coker, 1995; Belknap, 1995; Buzawa & Austin, 1993; Dawson & Hotton, 2014; Eitle, 2005; Millar & Brown, 2010), and laying charges (Millar & Brown, 2010). If victims were fearful of their partners, police were significantly less likely to take a report and/or make an investigation, but were more likely to remove the offending party (Barrett et al., 2011). However, police were found to be less likely to make an arrest if victims were both injured and fearful of retaliation (Robinson & Chandek, 2000b), highlighting the complex interplay of factors associated with police response.

Other situational factors found to affect police response include a higher likelihood of arrest in cases where the victims called the police themselves, among partners who were cohabitating, and in cases where the perpetrator or witnesses were present (Hall, 2005; Robinson & Chandek, 2000b). Situational- and incident-level factors have been found to have a stronger impact than demographic factors on police response to spousal violence (Dichter et al., 2011; Robinson & Chandek, 2000b). Last, while the literature is extremely limited in terms of the relationship between neighborhood characteristics and police response to cases of spousal violence, we know that arrest is less likely in neighborhoods with lower socioeconomic status (Smith, 1987), and more likely in communities that are wealthier and suburban (Dichter et al., 2011). Moreover, prior research has suggested that neighborhood factors are associated with citizen’s attitudes toward the police, with assessments of one’s neighborhood as having little to no crime or disorder related to positive perceptions and evaluations of the police and policing services (Anderson, 1999; Payne & Gainey, 2007; Schafer, Huebner, & Bynum, 2003). Previous research has also identified positive relationships between observing or experiencing community policing activities and satisfaction with the police, trust in the police (Gill, Weisburd, Telep, Vitter, & Bennett, 2014; Skogan & Hartnett, 1997), as well as a negative relationship between community policing activities and perceptions of neighborhood disorder (Gill et al., 2014). These findings have led to the question of whether the presence of a local community policing station in one’s neighborhood may influence the response taken by police in cases of spousal violence, or victim’s reported levels of satisfaction with the police response.

Spousal Violence: Victims’ Satisfaction With Police Response

Not only are the ways police respond to instances of spousal violence important, but how victims evaluate the police response or intervention are crucial, especially considering that previous police interactions can affect future reporting among victims of spousal violence (Klein, 2009; Leisenring, 2012; Tutty et al., 2008; Wolf et al., 2003), while victim perceptions can provide insight into ways in which victim–police officer relations can be improved (Leisenring, 2012, p. 148). Research exploring victim satisfaction with the justice system and police response to domestic violence in Canada suggests a higher proportion of women reporting satisfaction compared with dissatisfaction with police response (Tutty et al., 2008). However, Leisenring’s (2012) study of female victims of IPV found that more than half the women in their sample were unsatisfied with the police response, citing that victims felt they were not taken seriously, that the police lacked sympathy and/or blamed them, and that they felt the police were especially critical of women who called the police multiple times and chose to stay with their partner after repeated instances of violence. Overall however, victims’ levels of satisfaction with police response rests largely on the congruency between their expectations of how they feel the police should respond and the actual response taken by the police (Buzawa & Austin, 1993; Buzawa, Austin, Bannon, & Jackson, 1992; Tutty et al., 2008), with a lack of police action identified as a key reason for dissatisfaction with the police among victims (Geller, 1991; MacLeod, 1987, as cited in Tutty et al., 2008, p. 57).

As this brief review of the literature suggests, although there is some research specifically exploring the various factors influencing police officer’s decisions and actions in cases of spousal violence, these studies are limited in generalizability, comprehensiveness, and breadth of police response, and thus vary in their findings (Dichter et al., 2011). Moreover, much of this research has been conducted in the United States within specific jurisdictions or states (Dichter et al., 2011; Jones & Belknap, 1999; Robinson & Chandek, 2000a, 2000b), thereby influencing the generalizability of the results beyond the specific sample utilized. Research tends to focus specifically on individual-level or incident-level factors in predicting police response (often measured as arrest or no arrest), and neglects to include community factors in the analysis of factors influencing police response to spousal violence (Dichter et al., 2011). Moreover, while it is encouraging that satisfaction with police has been increasing over the years among Canadian victims of IPV (Tutty et al., 2008), literature that explores factors that influence victims’ satisfaction with police response tends to focus specifically on the police response of arrest, failing to account for possible differences in satisfaction based on different police responses, and potential variations in victims’ opinions of whether or not there was any other action the police could have engaged in.

In research on spousal violence, the interrelationship between the key factors identified in the literature (individual-level/socio-demographic factors; situational-incident-level factors; neighborhood-level factors) and their combined association with police decisions in cases of spousal violence are not clearly understood, nor are these factors clearly understood in terms of their relationship to victim satisfaction with police response to instances of spousal violence. The following study attempts to address these gaps in the research by utilizing a nationally representative sample of both male and female victims in Canada to simultaneously explore the relationship between individual, violence, and neighborhood characteristics with a variety of different police responses in cases of spousal violence, and with victims’ subsequent levels of satisfaction with the police response. In other words, we explored the potential associations between individual-, incident-, and neighborhood-level factors and how police respond when contacted because of spousal violence in Canada.

The Current Study

Research Questions

The primary research questions investigated were as follows:

Research Question 1: Does police response to spousal violence vary by victims’ (a) sociodemographic characteristics, (b) violence characteristics, and/or (c) neighborhood characteristics?

Research Question 2: Does satisfaction with the police response vary based on victims’ (a) sociodemographic characteristics, (b) violence characteristics, and/or (c) neighborhood characteristics?

Method

Data

A secondary analysis was conducted using data from the 2009 Canadian General Social Survey (GSS) Cycle 23: The Victimization main file. The GSS is nationally representative annual telephone survey conducted by Statistics Canada to assess living conditions and indicators of social well-being among Canadians (Statistics Canada, 2010). While the annual topic of the GSS varies by year, every fifth year, the focus is on criminal victimization. As of January 1, 2016, the 2009 GSS was the most recently available microdata that evaluated experiences of spousal abuse, as the 2014 GSS was not yet released. While an in-depth description of the methods employed in the GSS have been previously published (Statistics Canada, 2010), a summary is provided.

The publicly available microdata file of the GSS suppresses variables involving sensitive subject matter (e.g., variables on spousal violence victimization), thus requiring researchers to apply to the Statistics Canada Research Data Centre (RDC) program for access to the restricted data. Data for this current study were accessed by researchers through this program, and as a result, all analysis occurred on site in an RDC location, following procedures set forth by Statistics Canada, including required weighting, rounding of raw numbers to a base of 0 or 5, and restrictions related to minimum cell counts (n > 10). To meet these data requirements, many variables were collapsed and recoded into dichotomous variables to maximize cell counts and emphasize variability for inclusion in the multivariate models.

Data Collection

The 2009 GSS data were collected between February and March 2009 throughout the 10 provinces in Canada (those residing in the territories were sampled separately and thus are not included in this analyses). All persons residing in Canada aged 15 years or older comprised the target population. Those who were full-time residents of institutions, and those who reside in households without a land line were not included in the sampling frame. Given that approximately 8% of Canadian households reported sole reliance on cell phones in the preceding year (i.e., 2008), survey weights were constructed to ensure the sample was reflective of those without land line telephones as well (Statistics Canada, 2010).

A geographically stratified random sample was obtained by dividing the 10 provinces of Canada into 27 geographic strata, and using random digit dialing (RDD) to randomly generate a list of phone numbers to selected households through simple random sampling. Within each household selected, one respondent above the age of 15 was randomly selected to participate in a computer-assisted telephone interview (CATI) in the official language (English or French) of their choice. The final sample size obtained was 19,422, representing a response rate of approximately 38% (Statistics Canada, 2010).

Measures

Spousal violence victimization

The spousal violence modules in the GSS focus on experiences related to abuse by a current or former spouse or common-law partner. These modules were therefore administered to respondents (both male and female) who reported having a current or former spouse/common-law partner within the previous 5 years. As a result, these modules captured the experience of a specific form of IPV, namely, the experience of spousal violence, and did not tap into violence experienced in casual, dating, or non-cohabitating relationships. While the first module assessed both emotional and financial forms of abuse, the second module focused specifically on physical and sexual forms of abuse. All items in the latter module are criminal acts under the Canadian Criminal Code that are under the jurisdiction of the police. This analysis focused on physical and/or sexual forms of abuse experienced within the context of a spousal/common-law relationship.

Physical/sexual abuse

This module began by asking respondents if their current or former partners (assessed separately) had engaged in the following abusive behaviors in the previous 5 years: (a) threatened to hit you with his or her fists or anything that could hurt you; (b) thrown anything at you that could have hurt you; (c), pushed, grabbed, or shoved you in a way that could have hurt you; (d) slapped you; (e) kicked you, bit you, or hit you with his or her fist; (f) hit you with something that could have hurt you; (g) beaten you; (h) choked you; (i) used or threatened to use a gun or knife on you; and (j) forced you into any unwanted sexual activity by threatening you, holding you down, or hurting you in some way. Original response options were “yes” and “no,” with individuals who reported any one of these 10 behaviors coded as a victim of physical/sexual spousal abuse. This way of operationalizing spousal violence victimization, by coding individuals as victims if they report experiencing any one instance of violence, is consistent with previous research (Brownridge, Tallieu, Tyler, Tiwari, & Chan, 2011; Brownridge et al., 2008). Based on this measure, a total of 890 respondents in the GSS reported experiencing physical/sexual spousal violence by a current or former intimate partner.

Police contact

To assess whether or not individuals who had experienced physical or sexual spousal violence had any contact with law enforcement as a result of the violence, participants who reported experiencing at least one of the 10 abuse items were asked “Did the police ever find out about the violence in any way?” (yes = 1, no = 0). Those who responded affirmatively to this statement were included in the sample for analysis. Thus, our sample consisted of all male and female victims of physical and/or sexual abuse who had direct contact with the police as a result of the abuse (n = 235).

Predictor variables

Sociodemographic characteristics

Five sociodemographic variables were included in this analysis. (a) Age was assessed as a continuous variable, while the other variables were dummy coded, and included (b) gender (female = 1, male = 0), (c) experiences of physical or mental limitations derived from four items which inquired, “Are your daily activities at home, work, school, or any other area limited by . . . (1) a physical condition; (2) a psychological, emotional, or mental health condition; (3) learning difficulties; or (4) any other health condition? Response options included no, sometimes, often, or always. Individuals who reported experiencing any of these four limitations sometimes, often, or always, were coded as having a physical or mental limitation (physical or mental limitations = 1, no reported limitations = 0), (d) income (less than or equal to Can$30,000 = 1, higher than Can$30,000 = 0), and (e) location of residence (rural = 1, urban = 0). Income was dichotomized at $30,000, as this corresponds roughly to the annual living wage in Canada for a full-time worker earning an hourly living wage of approximately $15 per hour (Living Wage Canada, 2013). The urban/rural dichotomy was included to account for potential differences in police response based on location of residence, and to help control for unobservable characteristics at the regional level.

Violence characteristics

Violence or incident characteristics were used to assess the severity of physical/sexual abuse experienced by victims. These three measures included (a) frequency of incidents of abuse, (b) physical injury as a result of the abuse, and (c) fearing for one’s life as a result of abuse. The frequency of physical/sexual abuse experienced was assessed with the following question: “You said yes to at least one of the previous 10 questions concerning physical and sexual violence. During the past five years, has he or she been violent to you on more than one occasion?” If respondents answered yes, they were subsequently asked, “How many times did these things happen during the past five years?” Responses options ranged from 2 to 10, and greater than 10 incidents. This variable was dummy coded to account for violence that is rare/occasional and violence that is more commonly experienced/frequent (1-2 instances of abuse = 0, greater than 2 instances of abuse = 1). Physical injuries were assessed with the following question: “You said that during the past five years your spouse/partner was violent. During this/any of these incidents were you ever physically injured in any way, for example, bruises, cuts, broken bones, etc.?” Original response options included yes and no (yes = 1, no = 0). Fearing for one’s life was appraised with the question, “During the past 5 years, did you ever fear that your life was in danger because of your spouse/partner’s violent or threatening behavior?” The original response options of yes or no were dummy coded (yes = 1, no = 0).

Neighborhood factors

Perceived social disorder was evaluated in the GSS with a module assessing perceptions of neighborhood crime or incivility. Respondents were asked to rate whether they felt a series of nine situations were “a very big problem,” “a fairly big problem,” or “not a problem at all” in their own neighborhood. These items included (a) noisy neighbors or loud parties; (b) people hanging around on the streets; (c) people sleeping on the streets or in other public places; (d) garbage or litter lying around; (e) vandalism, graffiti, and other deliberate damage to properties or vehicles; (f) people being attacked or harassed because of their skin color, ethnic origin, or religion; (g) people using or dealing drugs; (h) people being drunk or rowdy in public places; and (i) prostitution. Due to relatively low endorsements of these items, to maintain reasonable cell counts, social disorder was dummy coded as being present if the respondent indicated that any of the nine items were “a very big problem” or a “fairly big problem” in their neighborhoods (social disorder = 1, no social disorder = 0). Second, whether or not the respondent’s community has a police station was evaluated with the following question: In your neighborhood, do you have a police station or community policing center/Royal Canadian Mounted Police (RCMP) detachment?” Affirmative responses were coded 1, while those who indicated they did not have a policing station in their neighborhoods were coded 0.

Outcome variables

Police response/police intervention

If respondents reported police contact as a result of the violence, they were asked a series of follow-up questions inquiring about the type(s) of police intervention they received. More specifically, respondents were asked, “What action did the police take? Did they . . . ” (a) visit the scene, (b) make a report/conduct an investigation, (c) issue a warning to your spouse/partner, (d) take your spouse/partner away, (e) arrest or lay charges against your spouse/partner, (f) take any other action. Response options for each of these items were “yes” and “no.” A series of dummy variables were created for each of these six types of intervention (police took that action/intervention = 1, police did not undertake that form of action/intervention).

Satisfaction with police response/intervention

To assess the level of satisfaction victims had with the police response, respondents were asked, “How satisfied were you with the actions that the police took? Were you, very satisfied? somewhat satisfied? somewhat dissatisfied? or very dissatisfied?” Response options were collapsed and dummy coded with those reporting they were very satisfied and somewhat satisfied coded as “satisfied = 1,” while those who reported being somewhat dissatisfied or very dissatisfied were combined and coded as “not satisfied = 0.”

Sample

Just less than one quarter of victims reported police involvement as a result of the violence. Victims who reported having experienced spousal violence victimization, and who had contact with the police as a direct result of the violence comprised the sample for this study (n = 235). Table 1 provides a breakdown of the sample characteristics.

Table 1.

Sample Description of Spousal Violence Victims Who Had Contact With Police.

Variable	Weighted Percentage (%)
Age	40.66 years (SD = 11.0)
Gender (female)	69.68
Visible minority	13.8
Physical or mental limitations	54.3
Income < Can$30,000	49.8
Rural residence	22.9
No. of violence instances > 2	68.16
Injury	58.29
Fear	50.0
Presence of a community policing station	65.7
Neighborhood has social disorder	45.0

Note. We were unable to include visible minority in the regression analysis due to low cell counts.

In this sample, the proportion of female victims is slightly higher than the proportion of male victims, with female victims overwhelmingly more likely than male victims to report having contact with the police as a result of the spousal violence. Just over 1 in 10 victims of spousal violence who had contact with the police are visible minorities¹ (13%), and greater than half of victims report having physical and/or mental limitations, being fearful of their partners, having experienced injuries as a result of spousal violence, and having experienced multiple instances of violence. Moreover, 45% of victims indicate there is some social disorder in their neighborhoods, and more than two thirds of victims indicate there is a local/community policing station in their neighborhood.

Analysis

Research Question 1 was assessed with a series of seven binomial logistic regression equations predicting police response/action to instances of spousal violence. Three blocks of predictor variables were included in each of the regression equations. Block 1 included sociodemographic variables, Block 2 contained violence characteristics, and Block 3 included neighborhood characteristics. Similarly, binomial logistic regression was used to assess Research Question 2 predicting overall satisfaction with police response. Variables were entered in the same three blocks as in Research Question 1 (Block 1: sociodemographic variables, Block 2: violence characteristics, Block 3: neighborhood characteristics). These analyses were conducted in line with Statistics Canada guidelines for release of RDC data, using normalized weights, rounding raw frequency statistics to a base of 0 or 5, and only including variables which meet minimum cell count requirements of 10. Missing data were excluded from all analyses.

Results

Descriptive Statistics

Police response/intervention comprised the primary outcome variables in this analysis. Descriptive statistics show variation in police response (see Table 2). These descriptive statistics suggest that in more than seven out of 10 instances, victims reported the police visited the scene, made a report/conducted an investigation, and/or issued a warning. Taking the partner away, and arresting and charging the partner were the least frequently reported police responses. Variations in gender are also evident. Although the police were equally likely to visit the scene for a similar proportion of both male and female victims (82% compared with 85%), additional police action or intervention was more commonly reported among female victims of spousal violence. In terms of satisfaction with the police response, half of all male victims (50%) and two thirds of female victims (65%) report being satisfied with the actions the police took in response to their experience of spousal violence.

Table 2.

Frequency of Specific Police Response and Satisfaction With Police.

Police Response	Weighted Percentage (%)
Visited Scene	84
Made a report/investigated	81
Gave warning	70
Took away partner	44
Arrested or charged partner	37
Satisfied with police response	61

Multivariate Analysis

Research Question 1 sought to assess whether police response to spousal violence varies by victim’s sociodemographic characteristics, violence characteristics, and/or neighborhood-level factors. It should be noted that some sociodemographic variables (e.g., visible minority status) presented in Table 1 were not included in these multivariate models due to a failure to meet the minimum Statistics Canada cells counts.

Of the six binomial regression models predicting police response to spousal violence, visiting the scene, making a report/conducting an investigation, giving a warning, or responding with “other” action were not significant models (p > .05), suggesting that the sociodemographic characteristics, violence characteristics, and neighborhood characteristics included in our analyses were not significant predictors of these types of police intervention. However, models predicting police response of taking the partner away and arresting and/or charging the partner were significant (p < .05), with the former trending toward significant (p < .10). Table 3 summarizes the significant findings from these analyses and presents results in odds ratios.

Table 3.

Binomial Logistic Regression Models Predicting Police Response.

	Arrest/Charged Partner			Took Partner Away
	Block 1	Block 2	Block 3	Block 1	Block 2	Block 3
Individual
Age	0.997	0.993	0.995	1.000	0.998	0.998
Female	2.501	3.233^†	4.080^†	3.122*	7.557 **	9.048 **
Limitations	2.632^†	3.476 *	3.964 **	1.948	2.240	2.182
<30k income	1.557	1.820	1.645	1.185	1.086	0.925
Rural	3.093 *	3.271 *	2.800^†	1.647	1.993	1.556
Incident
No. of incidents > 2		0.272 *	0.248 *		0.928	0.996
Injury		1.039	1.191		0.780	0.918
Fear		0.952	1.282		0.268 *	0.314^†
Neighborhood
Police station			0.275 *			0.290 *
Social disorder			2.767^†			1.282
Model χ²	16.39 **	18.62 *	34.76 ***	9.10^†	12.98	17.43^†
Pseudo R²	.1207	.1633	.227	.077	.125	.164
Log pseudolikelihood	−89,135	−84,810	−78,312	−91,632	−86,853	−82,964
n	130	130	130	135	135	135

Note. Odds ratios are reported. Models predicting the police response of visiting the scene, making a report/investigating, giving a warning, and “other” action are not included in table (p > .05).

†

p < .1. *p < .05. **p < .01. ***p < .001.

Controlling for violence characteristics and neighborhood-level factors, compared with being a male victim, being a female victim significantly increased the odds by more than 9 times that the police took their partner away, while those with physical and/or mental limitations had 4 times higher odds than those without limitations to report the police arrested and/or charged their partner. No other sociodemographic characteristics were found to be significant predictors (p < .05) in any of the models; however, being a female and residing in a rural area both trended toward significance (p < .10) in the final model predicting the police response of arrest/charge, with female victims and victims residing in rural areas having higher odds than male victims and victims residing in urban areas to report the police arrested/charged their partners.

Violence characteristics were found to be significant only in the model predicting the police response of arresting and/or charging the partner, with a significant negative relationship between frequency of violence and likelihood of police arrest, while being physically injured or fearful of one’s partner were not significantly related to this police response/intervention. Holding sociodemographic characteristics and neighborhood characteristics constant, victims who reported multiple/more frequent instances of violence had significantly lower odds that the police arrested/charged their partner, compared with victims who reported their experiences of spousal violence to be less frequent (≤ 2).

In the model predicting the police response of taking one’s partner away, with the addition of the incident-level variables in Block 2, fear is significantly negatively related to the police taking one’s partner away; however, once neighborhood factors in Block 3 are included in the model, this relationship drops to borderline significant (p < .10), suggesting that the negative relationship between fear of one’s partner and the police response of taking one’s partner away can be somewhat explained via neighborhood factors.

Although social disorder was not significantly related to any of these police actions, a borderline significant (p < .10) positive relationship was found in the model predicting the police response of arrest/charge. More specifically, victims who resided in neighborhoods with social disorder had higher odds of reporting the police took the action of arresting/charging their partner, compared with victims who did not report any neighborhood social disorder. Counter to what we expected, the presence of a policing or community policing station in one’s neighborhood was found to be significantly associated with lower odds of police taking the partner away, and arresting and/or charging the partner.

In terms of model statistics, each block of predictors significantly added to the overall model’s predictive ability in accounting for the police response of arrest/charging one’s partner, providing support for including each block of predictors to account for the diverse factors associated with this police response. In the models assessing the police response of taking one’s partner away, Block 2 (incident-level characteristics) did not significantly improve the model; however, neighborhood characteristics (Block 3), and more specifically, the presence of a community policing station had a significant impact on the final model. Taken together, these results suggest that police response to instances of spousal violence varies by an interplay between victim sociodemographic characteristics, violence/incident characteristics, and neighborhood characteristics.

Research Question 2 sought to assess the potential influence of sociodemographic characteristics, violence characteristics, and neighborhood characteristics on victims’ level of satisfaction with the action(s) the police took. Results indicate that the model predicting satisfaction with police response was not significant (p > .05). Thus, we did not find support for the argument that individual characteristics, violence/incident characteristics, and neighborhood characteristics affect levels of satisfaction with the police by victims of spousal violence.

Discussion

The findings of this study suggest that in most of the cases of spousal violence, the police visited the scene (84%) and/or took a report/made an investigation (80%). These findings represent a higher rate of police action compared with earlier studies in Canada (Barrett et al., 2011), suggesting that policies mandating police respond and take action in cases of spousal violence are potentially having an impact on officer behavior. Responding to cases of spousal violence enables police officers to enact their role as protectors, to provide immediate protection, and to disrupt the violence, all of which are important considering that victims experiencing the most severe forms of violence are most likely to come into contact with the police (Bachman & Coker, 1995; Barrett et al., 2011; Davies et al., 2007; Felson et al., 2002; Lee et al., 2010).

Multivariate models predicting police response by three blocks of indicator variables were not significant (p > .05) in predicting the police response of visiting the scene, making a report and/or conducting an investigation, and issuing a warning. These findings suggest that the police response of taking each of these actions was not associated with individual-level characteristics such as gender, age, presence of victim’s physical/mental limitations and income of victims; violence characteristics such as frequency of violence, experience of injuries, and experience of fear; or neighborhood characteristics such as perceived social disorder or the presence of a policing/community policing station in one’s neighborhood. These nonsignificant findings are promising such that it appears police are taking action in most cases, and this action is unrelated to characteristics of the victim, the incident, or the neighborhood. However, it should be noted that these results could be due to the fact that sociodemographic characteristics such as ethnicity and sexual orientation could not be included in models assessing police response due to RDC minimum cell count requirements. It is quite probable that if these variables were included in the models, results may have been different.

Counter to previous research that reported the effects of age (Robinson & Chandek, 2000a) on police arrest, the only sociodemographic factor found to be significantly related to the police response of arrest/laying charges was whether or not victims had a physical and/or mental limitation, while gender and rural residency were borderline significant, but consistent with the direction of previous literature linking gender (Dawson & Hotton, 2014; Millar & Brown, 2010), and urban/rural residency (Pattavina et al., 2007) with the police response of arrest/charge. Victims who reported having a physical or mental limitation were significantly more likely than victims who did not report limitations to indicate the police made an arrest and/or laid charges. While the nature of our data prohibits further exploration as to why police officers are more likely to arrest perpetrators against victims with disabilities, this finding could be interpreted as promising in that it suggests police are recognizing some of the unique circumstances affecting victims with limitations. However, the finding may also reflect disparate treatment of victims with disabilities that may be, either consciously or unconsciously on the part of officers, reflective of common stereotypes about persons with disabilities. For example, disability scholars have pointed out that persons with disabilities are often socially constructed as helpless objects of pity (Shapiro, 1994). As such, persons with disabilities are often denied personal agency and are the recipients of paternalistic treatment grounded in the presumption that people with disabilities are in greater need of protection than other members of the public. Given these potential complexities, an important direction for future research is the further examination of police response specifically to cases of spousal violence reported by victims with disabilities to ensure that victims with disabilities are receiving equal protection and treatment from responding officers. A similar explanation could account for the borderline significant finding that female victims had higher odds than male victims to report the police arrested/charged their partners, and that female victims were significantly more likely than male victims to report the police took their partners away. It is possible that police officers view female victims as in greater need of protection than male victims, and thus may be more likely to act by removing/arresting/charging their partners in cases of spousal violence.

While violence characteristics were not predictive of the police response of taking the partner away, inconsistent with previous research (Millar & Brown, 2010), arrest was less likely in cases where victims reported multiple instances of violence, while victim injury and fear of one’s partner were not significantly related to the police officer response of making an arrest and/or laying charges. These findings run counter to what would be expected, as well as previous findings linking injuries to police response of making an arrest (Bachman & Coker, 1995; Belknap, 1995; Buzawa & Austin, 1993; Eitle, 2005; Millar & Brown, 2010), and laying charges (Dawson & Hotton, 2014; Millar & Brown, 2010). However, as indicated, female victims were significantly more likely than male victims to report the police took their partner away, and findings trended toward a significant positive relationship between being female and the police response of arrest/charging one’s partner. Previous research has discovered differences in the experiences of male and female victims of IPV, with female victims more likely to report being physically injured and to experience the most severe forms of IPV (Burczycka, 2016). Thus, females may be more likely to have their partners taken away and/or arrested because they have experienced the most severe forms of violence, and these violence characteristics may be suppressed by the effect of gender, as female victims were substantially more likely than male victims to report multiple instances of violence, to report having experienced an injury, and to report being fearful of their partners (Burczycka, 2016; Statistics Canada, 2005). Moreover, with each successive block of predictors added to models predicting the police response of arrest/charge, as well as the police response of taking one’s partner away, the odds ratio for gender increases, with female victims having substantially higher odds than male victims to report the police response to spousal violence. Thus, controlling for incident (Block 2) and neighborhood (Block 3) characteristics enhanced the magnitude of the odds ratio for females (see Table 3). These findings highlight how victims experiencing different dynamics of violence (e.g., “coercive controlling violence” vs. “situational couple violence”) may have different police interactions, and victims with ongoing abuse may be less able to communicate their need for safety to police due to ongoing threat of violence, or fear of retaliation by their partners. Thus, it is imperative that police respond to instances of spousal violence with training and experience in identifying the dynamics involved in the diverse types of violence.

Last, while self-reported levels of neighborhood social disorder were not significant across our models, it did trend toward significance (p < .10) in the model predicting the police response of arrest/charge, with social disorder being related to higher odds of the police taking this action in cases of spousal violence. The neighborhood factor found to be a significant predictor of police response was having a community policing station in one’s neighborhood. Previous literature has reported a positive relationship between witnessing community policing activities and positive evaluations of and trust in the police (Gill et al., 2014); however, in our analysis, the presence of a policing or community policing station was related to a significantly lower likelihood of police taking action, and was not significantly related to satisfaction with the police response following spousal violence, suggesting that police presence in the form of policing stations is not enough. Based on the results of this analysis, when a policing station is present in one’s community, police appear to be less likely to engage in the most formal types of intervention in cases of spousal violence (arrest/charge and taking the partner away). Multivariate models predicting the less severe types of police intervention (visiting the scene, taking a report, and providing a warning) were not significant, suggesting that neighborhood-level factors such as the presence of a policing station in one’s community are unrelated to these types of police response, and instead are only related to the most formal types of police response. However, it should be noted that this question specifically asked respondents, “In your neighborhood, do you have a police station or community policing center/RCMP detachment?” and given that this combines community policing centers, police stations, and RCMP detachments, the individual impact of each of these types of policing stations could not be assessed. It is very plausible that specifically focusing on community policing stations would yield different results; however, this was not possible to assess given the data available. Despite these ambiguous findings, they further emphasize the complex relationship of individual, incident, and neighborhood factors on police response to IPV, and more specifically, spousal violence. By utilizing a nationally representative sample of spousal violence victims in Canada, this study simultaneously explored the relationship between individual, violence/incident, and neighborhood characteristics on different police responses in cases of spousal violence, and with victims’ subsequent levels of satisfaction with the police response, these findings contribute to the body of literature on spousal violence and police response in Canada.

Limitations

While the findings of this study contribute to the literature on police response to IPV in Canada, this study is not without limitations. Given this research was conducted with Statistics Canada’s microdata file in the RDC, all raw analyses had to be rounded to the base of 0 or 5, and all cell counts had to meet minimum requirements for release (n = 10). As a result, this restricted our ability to utilize additional sociodemographic variables such as visible minority status, Aboriginal status, and sexual orientation, all of which have been identified in the literature as affecting police response to spousal violence, and thus represents a significant limitation to our study. As this present study was unable to assess variations among victims based on these characteristics, our findings do not necessarily reflect the experiences of racially and ethnically marginalized victims or those who are sexual minorities.

Conclusion

As frontline first responders to spousal violence, police officers play a critical role in supporting victims. Victims of spousal violence, however, are not a homogeneous group and may vary substantially in their needs and desired measures for protection from the police. While we know that underreporting of spousal violence is especially problematic (Burczycka, 2016; Wolf et al., 2003), with just more than one in five victims in the present study reporting police contact in the aftermath of violence, in greater than 80% of cases reported to the police, police officers visited the scene. This initial point of contact between victims and police highlights an avenue for police to strengthen their rapport with victims, especially because interactions with the police can influence victims’ future reporting practices (Leisenring, 2012; Wolf et al., 2003). In their meta-analysis of interventions designed to improve public confidence in perceptions of police legitimacy, Mazerolle, Bennett, Davis, Sargeant, and Manning (2013) found that the most crucial factor affecting public trust and confidence in the police was procedural justice dialogue. More specifically, researchers found that interventions in which police officers engaged in dialogue that encourages citizen participation and gives citizens a voice, that reflects respect, dignity, fairness, and trustworthy motives, as well as neutrality and fairness in decision making, were significantly related to police perceptions of police legitimacy and confidence (Mazerolle et al., 2013). Police policy and practice should maximize opportunities to strengthen their relationship with victims, by engaging in procedural justice dialogue during contact experiences.

The findings indicating that sociodemographic characteristics including gender and physical/mental limitations were some of the most salient factors related to police response emphasize the importance for policy to incorporate a gender-based plus analysis to account for the multiplicity of identity factors that intersect among individuals including, but not limited to, gender, race, ethnicity, religion, and mental or physical disability (Status of Women Canada, 2017). In doing so, this can help ensure that diverse needs of victims are being identified and met. Factors shaping women’s, men’s, and other gender diverse people’s experiences of IPV, as well as their interactions with the police are likely to be quite different. Understanding these diverse experiences is integral to encouraging positive police–citizen interactions. Findings from this study can contribute to the further development of evidence-based policing practices that support officers in carrying out their role in a manner that is equitable for all victims, and that recognizes that different victims may have different needs and abilities to communicate these needs.

Footnotes

Authors’ Note

Data for this project was accessed by the researchers through the Statistics Canada Research Data Centre program. The views and opinions expressed herein are solely those of the researchers and do not necessarily reflect those of Statistics Canada.

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 with funding from the Social Sciences and Humanities Research Council Insight Development Grant Program (Grant 430-2013-000465) and the University of Windsor Tri-Success Grant Program.

Notes

Author Biographies

Amy Peirone is a PhD candidate at the University in Windsor in the Department of Sociology, Anthropology and Criminology. Her research interests include intimate partner violence, policing and perceptions of the police, and quantitative research methodologies.

Betty Jo Barrett is an associate professor and director of Women’s and Gender Studies at the University of Windsor. Her research interests include the study of intimate partner violence in both heterosexual and LGBTQ (lesbian, gay, bisexual, transgender, questioning/queer) relationships, with a specific focus on survivors’ help seeking and their interactions with formal and informal sources of support in the aftermath of violence.

Chi Ho Cheung is a graduate of the doctoral program in the Department of Mathematics and Statistics at the University of Windsor.

Nazim Habibov is an associate professor in the School of Social Work at University of Windsor, Ontario, focusing on International Social Welfare. His research interests include the comparative social welfare policy, social security, institutions, and social trust.

References

Anderson

(1999). Code of the street: Decency, violence, and the moral life of the inner city. New York, NY: W.W. Norton.

Apsler

Cummins

M. R.

Carl

(2003). Perceptions of the police by female victims of domestic partner violence. Violence Against Women, 9, 1318-1335.

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.

Barrett

St. Pierre

Vallaincourt

. (2011). Police response to intimate partner violence in Canada: Do victim characteristics matter? Women & Criminal Justice, 21, 38-62.

Belknap

(1995). Law enforcement officers’ attitudes about the appropriate responses to woman battering. International Review of Victimology, 4, 47-62.

Brownridge

D. A.

Hiebert-Murphy

Ristock

Ling Chan

Tiwari

Tyler

K. A.

. . . Santos

S. C.

(2008). Violence against separated, divorced, and married women in Canada, 2004. Journal of Divorce & Remarriage, 49, 308-329.

Brownridge

D. A.

Tallieu

T. L.

Tyler

K. A.

Tiwari

Chan

K. L.

(2011). Pregnancy and intimate partner violence: Risk factors, severity, and health effects. Violence Against Women, 17, 858-881.

Burczycka

(2016). Family violence in Canada: A statistical profile, 2014. Section 1: Trends in self-reported spousal violence in Canada, 2014 (Juristat, 36(1) No. 85-002-X). Ottawa, Canada, ON: Centre for Justice Statistics.

Buzawa

E. S.

Austin

T. L.

(1993). Determining police response to domestic violence victims: The role of victim preference. The American Behavioral Scientist, 35, 610-623.

10.

Buzawa

E. S.

Austin

T. L.

Bannon

Jackson

(1992). Role of victim preference in determining police response to victims of domestic violence. In Buzawa

E. S.

Buzawa

C. G.

(Eds.), Domestic violence: The changing criminal justice response (pp. 255-269). Westport, CT: Auburn House.

11.

Davies

Block

C. R.

Campbell

(2007). Seeking help from the police: Battered women’s decisions and experiences. Criminal Justice Studies, 20, 15-41.

12.

Dawson

Hotton

(2014). Police charging practices for incidence of intimate partner violence in Canada. Journal of Research in Crime & Delinquency, 51, 655-683.

13.

Department of Justice Canada. (2003, April). Final report of the ad hoc federal-provincial-territorial working group reviewing: Spousal abuse policies and legislations. Retrieved from http://www.justice.gc.ca/eng/rp-pr/cj-jp/fv-vf/pol/spo_e-con_a.pdf

14.

Dichter

M. E.

Marcus

S. C.

Morabito

M. S.

Rhodes

(2011). Explaining the IPV arrest decision: Incident, agency, and community factors. Criminal Justice Review, 36, 22-39.

15.

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.

16.

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.

17.

Ferraro

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

18.

Gill

Weisburd

Telep

C. W.

Vitter

Bennett

(2014). Community-oriented policing to reduce crime, disorder and fear and increase satisfaction and legitimacy among citizens: A systematic review. Journal of Experimental Criminology, 10, 399-428.

19.

Hall

D. L.

(2005). Domestic violence arrest decision making: The role of suspect availability in the arrest decision. Criminal Justice and Behavior, 32, 390-411.

20.

Jones

D. A.

Belknap

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

21.

Klein

A. R.

(2009). Practical implications of current domestic violence research: For law enforcement, prosecutors, and judges. Washington, DC: National Institute of Justice, U.S. Department of Justice. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/225722.pdf

22.

Lee

H. Y.

Park

Lightfoot

(2010). When does a battered woman seek help from the police? The role of battered women’s functionality. Journal of Family Violence, 25, 195-204.

23.

Leisenring

(2012). Victims’ perceptions of police response to intimate partner violence. Journal of Police Crisis Negotiations, 12, 146-164.

24.

Living Wage Canada. (2013). Retrieved from http://www.livingwagecanada.ca/

25.

Mazerolle

Bennett

Davis

Sargeant

Manning

(2013). Legitimacy in policing: A systematic review. The Campbell Collaboration. Retrieved from http://library.college.police.uk/docs/Mazerolle-Legitimacy-Review-2013.pdf

26.

Millar

Brown

(2010). Explaining gender differences in police arresting and charging behavior in cases of spousal violence. Partner Abuse, 1, 314-331.

27.

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.

28.

Payne

B. K.

Gainey

R. R.

(2007). Attitudes about the police and neighborhood safety in disadvantaged neighborhoods: The influence of criminal victimization and perceptions of a drug problem. Criminal Justice Review, 32, 142-155.

29.

Robinson

A. L.

Chandek

M. S.

(2000a). Differential police response to black battered women. Women & Criminal Justice, 12(2/3), 29-61.

30.

Robinson

A. L.

Chandek

M. S.

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

31.

Russell

Light

(2006). Police and victim perspectives on empowerment of domestic violence victims. Police Quarterly, 9, 375-396.

32.

Schafer

J. A.

Huebner

B. M.

Bynum

T. S.

(2003). Citizen perceptions of police services: Race, neighbourhood context, and community policing. Policy Quarterly, 6, 440-468.

33.

Shapiro

J. P.

(1994). No pity: People with disabilities forging a civil rights movement. New York, NY: Three Rivers Press.

34.

Sinha

(2012). Family violence in Canada: A statistical profile, 2010 (Juristat, No. 85-002-X). Statistics Canada. Retrieved from http://www.statcan.gc.ca/pub/85-002-x/2012001/article/11643-eng.pdf

35.

Sinha

(2013). Family violence in Canada: A statistical profile, 2011 (Juristat, No. 85-002-X). Statistics Canada. Retrieved from http://www.statcan.gc.ca/pub/85-002-x/2013001/article/11805-eng.pdf

36.

Skogan

W. G.

Hartnett

S. M.

(1997). Community policing, Chicago style. New York, NY: Oxford University Press.

37.

Smith

D. A.

(1987). Police response to interpersonal violence: Defining the parameters of legal control. Social Forces, 65, 767-782.

38.

Statistics Canada. (2005). Family violence in Canada: A statistical profile 2005 (Catalogue number 85-224-XIE). Ottawa, Ontario: Canadian Centre for Justice Statistics.

39.

Statistics Canada. (2010). General Social Survey-Victimization (GSS): Detailed information for 2009 (Cycle 23). Retrieved from http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&Id=49195#a1

40.

Statistics Canada. (2015). Visible minority and population group reference guide, NHS, 2011 (Catalogue no. 99-010-XWE2011009). Retrieved from http://www12.statcan.gc.ca/nhs-enm/2011/ref/guides/99-010-x/99-010-x2011009-eng.cfm

41.

Status of Women Canada. (2017). Gender-based analysis plus. Government of Canada. Retrieved from http://www.swc-cfc.gc.ca/gba-acs/index-en.html

42.

Stith

S. M.

(1990). Police response to domestic violence: The influence of individual and familial factors. Violence and Victims, 5, 37-49.

43.

Tjaden

Thoennes

(2000). Full report of the prevalence, incidence and consequences of violence against women (No. NCJ183781). Washington, DC: U.S. Department of Justice. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/183781.pdf

44.

Tutty

Wyllie

Abbott

Mackenzie

Ursel

E. J.

Koshan

J. M.

(2008). The justice response to domestic violence: A literature review (RESOLVE). Calgary, Alberta, Canada: University of Calgary. Retrieved from http://www.ucalgary.ca/resolve-static/reports/2008/2008-01.pdf

45.

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.

46.

World Health Organization. (2013). Global and regional estimates of violence against women: Prevalence and health effects of intimate partner violence and non-partner sexual violence. Geneva, Switzerland: World Health Organization.

47.

Zhang

Hoddenbagh

McDonald

Scrim

(2012). An estimation of the economic impact of spousal violence in Canada, 2009. Department of Justice. Retrieved from http://www.justice.gc.ca/eng/rp-pr/cj-jp/fv-vf/rr12_7/rr12_7.pdf