Abstract
Limited research has investigated police service provision and advocate involvement in domestic violence (DV) incidents. This study used a stratified random sample of 368 cases from an urban police department to assess police officers’ decisions to provide service provision and to involve an advocate in formally-reported DV incidents. Multivariate binary logistic regression models revealed decreased suspect age and presence of physical abuse significantly increased service referral. Advocate involvement significantly decreased when the victim and suspect were married, when the suspect used alcohol and/or drugs, and when there was physical evidence present in the case. Alternatively, suspect weapon use, victim injury, and service referral by police increased later advocate involvement. Implications and future research directions are discussed.
Domestic violence (DV) 1 is a serious public concern, has affected millions of individuals within the United States (Smith et al., 2018; Truman & Oudekerk, 2019), and has produced acute and long-term adverse health sequelae (Campbell, 2002; Coker et al., 2000, 2002). DV survivors, for example, have suffered from immediate physical injuries, somatic complaints, and psychological trauma consequences (Campbell, 2002; Coker et al., 2000, 2002). Despite high prevalence and deleterious health outcomes, official statistics have suggested only half of victims formally report their abuse to law enforcement personnel (Morgan & Truman, 2018; Truman & Oudekerk, 2019).
Decades of research and policy debates have underscored limitations in the criminal justice response to DV. Indeed, a substantial portion of U.S. history has reflected law enforcement perceptions of partner violence as a “private,” family matter (Freedman, 2002; Lutze & Symons, 2003; Martin, 1975) where incidents were often dismissed by the criminal justice system (Lutze & Symons, 2003). As a result, DV victims were left vulnerable and unprotected from continued abuse. Extensive efforts by grassroots organizations, social justice advocates, and academicians resulted in policy changes. Notably, law enforcement agencies implemented mandatory and pro-arrest policies to improve the handling of DV calls for service and reduce re-victimization (Barner & Carney, 2011; Han, 2003; Lutze & Symons, 2003; Maxwell et al., 2002). That said, some police personnel have continued to respond negatively to DV through antagonistic and unsupportive behaviors, including hostility, stigma, and victim-blame (Alhusen et al., 2010; DeJong et al., 2008; Stewart et al., 2013). This has produced secondary victimization and exacerbated trauma (Campbell, 2002).
Existing efforts to understand shortcomings in police response to DV have predominantly assessed factors affecting formal decisions, such as arrest (Durfee & Fetzer, 2016; Franklin et al., 2019; Garza et al., 2020; Russell & Sturgeon, 2019). Comparatively less research has examined other informal decision-making processes, including victim service provision. Broadly, prior research has suggested that police officers may draw upon existing frameworks, or schema, to inform decision-making when responding to calls for service (CFS; Robinson, 2000; Stalans & Finn, 1995; Watson et al., 2014). Schema theory has posited that individuals interpret and respond to scenarios based on their frames of references that have been informed by a myriad of factors (Rumelhart, 1984; Rumelhart & Ortony, 1977; Taylor & Crocker, 1981). Within the context of policing, for example, officers’ schemata may have been shaped by their general socialization, academy and in-service training, and prior experiences in preforming their job (Watson et al., 2014). Moreover, police officers’ schema about a DV incident (e.g., that these calls are dangerous for police, that victims rarely leave abusive relationships, or are unwilling to cooperate with police efforts, etc.) can inform their decisions related to arrest or service provision. Of particular importance has been understanding service referral decisions by police officers because direct resource provision and advocate involvement have been effective tools in the trauma-informed response to DV (Department of Justice, 2015). Indeed, research has suggested victim- (Buzawa & Hotaling, 2006; Kernic & Bonomi, 2007; Sun, 2007), suspect- (Kernic & Bonomi, 2007), case- (Finn & Stalans, 1995; Stalans & Finn, 2000; Swerin et al., 2018), and officer characteristics (Belknap & McCall, 1994; Finn & Stalans, 1995; Goodson et al., 2020; Homat & Kennedy, 1985; Stalans & Finn, 1995, 2000; Sun, 2007) have predicted police-initiated victim service referral. While fruitful, this research has largely been atheoretical, limited in scope, and needs additional exploration. To that end, Schema Theory can be used as a broad theoretical framework for understanding factors that affect police decision-making in the provision of direct services and advocacy.
Using schema theory as the theoretical scaffolding to understand service provision, the current study employed a sample of 368 redacted DV case files drawn from an urban police department located in one of the five most populous and diverse U.S. cities. 2 Quantitative data were extracted and analyzed to assess: (1) the extent to which police officers provide service referrals to DV victims; (2) predictors of service referral; (3) the extent to which DV cases involved a victim advocate; and (4) predictors of victim advocate involvement.
Schema Theory and Police Response to Domestic Violence
While the criminal justice response to DV has improved over time, empirical studies have documented general deficits in how police officers handle DV incidents. For example, almost half of DV victims who formally reported violence to law enforcement have described adverse experiences with police, including disbelief, hostility, blame, and questions pertaining to credibility (Erez & Belknap, 1998; Stephens & Sinden, 2000). These negative reactions have produced secondary victimization, exacerbated trauma, and diminished cooperation among victims (Koss, 2000). As a result, victims may be vulnerable to continued partner abuse. Police response to victims characterized by insensitivity has produced case attrition, decreased suspect apprehension, and concerns surrounding public safety.
Feminist scholarship has suggested that adverse criminal justice responses to DV may be, in part, the result of entrenched cultural myths regarding sex roles within both the public and private spheres, including the ways that men and women should behave and interact in interpersonal relationships (Freedman, 2002; Fleming & Franklin, 2020; Koss et al., 1994). Negative sex scripts and stereotypes have promoted misconceptions surrounding unequal power dynamics and unrealistic expectations in relationships, particularly as it relates to women and femininity (see Franklin, 2013; Johnson, 1997). Police officers are not immune from these messages (DeJong et al., 2008; Fleming & Franklin, 2020). For example, durable sex stereotypes have suggested that femininity is manifested as passivity, emotionality, and dependence (Johnson, 1997) and masculinity is displayed as aggression, authority, independence, and decisiveness (Kilmartin, 2000). These gender norms have organized social life where men and masculinity are valued and women and femininity are viewed as less desirable. According to schema theory then, an individual’s interpretation of a given event is directly informed by their perceptual shorthand, which draws from these broader social messages and stereotypes. This schema can impact the way an individual would respond to an event (Rumelhart, 1984; Rumelhart & Ortony, 1977; Taylor & Crocker, 1981). Patriarchal socialization has promoted traditional gender norms within heterosexual relationships and excused partner violence (see Johnson, 1997; Martin, 1975). This has influenced schema surrounding DV and can impact the manner in which police respond (Franklin et al., 2019; Goodson et al., 2020), including their attributions of severity, seriousness, and culpability. Indeed, empirical studies have documented police officers’ dearth of understanding regarding the complex psychological dynamics of relationship abuse. This lack of insight has increased adherence to DV myths and frustration in responding to partner abuse (DeJong et al., 2008; Fleming & Franklin, 2020; McPhedran et al., 2017). Myth adherence and frustration can translate to insensitive police responses because their schema surrounding DV may suggest this offense is not worthy of criminal justice intervention.
Service Referral among Police Personnel
Current trends in effective criminal justice interventions to DV have involved victim-centered and trauma-informed strategies, especially as it relates to augmenting police response. Among the variety of tactics available to law enforcement, the importance of service provision information cannot be understated (Department of Justice, 2015). Police have a unique opportunity to provide referrals and involve victim advocates when responding to DV. These approaches can mitigate trauma consequences and encourage victim engagement in criminal justice case processing (Camacho & Alarid, 2008; Dawson & Dinovitzer, 2001); thereby, reducing case attrition and increasing public safety. Police agencies, however, are not homogenous and have employed different forms of victim service delivery. Wilson and Segrave (2011) identified three modes of victim service delivery used by police: (1) dedicated unit services, (2) a liaison officer, or (3) referral services. All three strategies are effective, though police agencies may adopt one of the three or create a hybrid model that best fits the organization.
Research has suggested police have been inconsistent in providing service referral information to DV victims who present for assistance (Corcoran et al., 2001; Finn & Stalans, 1995; Goodson et al., 2020; Jones & Belknap, 1999; Kernic & Bonomi, 2007; Oppenlander, 1982; Ruff, 2012; Stalans & Finn, 1995, 2000; Swerin et al., 2018). Early research reported DV referrals in 3% to 4% of calls for service (Ferraro, 1989; Oppenlander, 1982). Subsequent studies have reiterated inconsistencies in the frequency of service referral among police. Kernic and Bonomi (2007), for example, investigated police use of crisis intervention in 2,092 male-to-female perpetrated DV incidents. Findings demonstrated that responding officers activated crisis intervention services in only 19.8% of DV incidents.
Ruff (2012) assessed police response to domestic disputes following training on DV. Among these incidents, police informed victims of services available through a witness coordinator in 11% of cases (Ruff, 2012). Most recently, Goodson et al. (2020) examined self-reported frequency of service referral to IPV and sexual assault victims among 452 police officer participants. On average, police officers were relatively likely to provide victims with resource information (Goodson et al., 2020).
Predictors of DV Service Referral among Police Officers
Assessments have identified the significant relation between victim and suspect characteristics and police service referral decisions (Buzawa & Hotaling, 2006; Kernic & Bonomi, 2007; Sun, 2007). Police have been more likely to provide service referral when the victim is female compared to male (Sun, 2007). While prior studies have underscored the significance of victim age in service referral decisions, this relation has varied by age range. Buzawa and Hotaling (2006), for example, assessed police response in 320 domestic assaults; police were less likely to provide juvenile victims, compared to adult victims, with information about legal rights and protection orders. Further, Sun (2007) examined 273 DV incidents involving 623 citizens to investigate police supportive actions including providing information and physical help to victims, encouraging victims to sign a formal complaint and use the legal process, seek help from social agencies, and reassuring or comforting victims. Police officers were more likely to engage in supportive actions if victims were young (under 18 years of age) or elderly (60 and above) demonstrating a curvilinear relation between DV support and age. In addition to age, other demographic characteristics have been important. Kernic and Bonomi (2007) discovered police were more likely to activate crisis intervention services in cases of male-to-female DV when victims were married to the abusive partner, pregnant, or identified as Latina or Asian, compared to counterparts. Additionally, police were more likely to involve the victim crisis intervention team when the suspect used alcohol or drugs in the presenting incident compared to when substance consumption was not a factor. Finally, research has suggested that DV incidents involving affluent victims resulted in increased support from police when compared to less affluent victims (Sun, 2007).
Service provision by police has also been affected by case characteristics, including the presence of children, type of abuse, and severity and pattern of violence (Finn & Stalans, 1995; Kernic & Bonomi, 2007; Stalans & Finn, 1995, 2000; Sun, 2007; Swerin et al., 2018). Swerin et al. (2018) examined 345 DV reports from a municipal police agency in the Northwestern U.S. to assess how often children were present and the impact this had on police response. The presence of a child significantly increased the responding officer’s victim-directed interventions, comprised of service referrals, contact with a victim-witness coordinator, information about civil protection orders, and transport to a safe location. Further, victims of physical abuse, compared to psychological and other non-physical abuse, were more likely to receive crisis intervention services (Kernic & Bonomi, 2007). Additionally, abuse severity, including victim injury (Finn & Stalans, 1995; Kernic & Bonomi, 2007; Stalans & Finn, 2000; Sun, 2007) and weapon use, has increased service provision by law enforcement. Finally, patterns of violence have been a contributing factor for providing advocacy services. Victims with a disclosed history of DV have been more likely to receive information on available services by police compared to victims with no history of DV (Kernic & Bonomi, 2007).
Furthermore, officer demographics, occupational characteristics, and police attitudes have predicted service referral (Belknap & McCall, 1994; Finn & Stalans, 1995; Goodson et al., 2020; Homat & Kennedy, 1985; Stalans & Finn,1995, 2000; Sun, 2007). Women police personnel (Homat & Kennedy, 1985; Stalans & Finn, 2000; Sun, 2007) and older officers (Finn & Stalans, 1995) have been more likely to provide services to DV victims compared to men and younger police practitioners. Relevant occupational predictors have included educational attainment, officer rank, exposure to DV, and familiarity with services. Belknap and McCall (1994) found officers who reported higher educational attainment and increased rank were more likely to provide female victims with shelter referrals compared to less educated and lower rank personnel. Further, exposure to DV (Stalans & Finn, 2000) and increased familiarity with available community services (Goodson et al., 2020) have predicted higher levels of service referral. Cognitions surrounding DV survivors, suspects, and victim advocates have also influenced service provision. Police perceptions of DV victims as combative or argumentative (Stalans & Finn, 1995) and as lacking credibility (Stalans & Finn, 1995) have decreased referral to shelter services. Additionally, when police have perceived the suspect as more dangerous (Stalans & Finn, 1995) and increasingly responsible for the abuse (Finn & Stalans, 1995), shelter referrals have increased. Finally, police with more favorable perceptions of victim advocates have provided referrals to victims with greater frequently (Goodson et al., 2020).
Purpose of the Present Study
Despite the utility of service referral for DV victims, both in terms of trauma mitigation and increased engagement with investigations, little research has examined the police decision to provide referrals using case processing data. Further, given that police agencies may employ hybrid strategies for victim service delivery, it is fruitful to identify variation in officer decisions to provide a referral or involve a victim advocate. The purpose of the current study was to assess service referral decisions among police and subsequent advocate involvement in a sample of DV incident case files. Using a stratified random sample of 368 redacted DV incidents drawn from a large, urban police department located in one of the five most populous and diverse U.S. cities, the following four research questions were addressed:
Methods
Sample and Procedure
Data were derived from a larger federally-funded grant awarded by the Office on Violence Against Women (OVW) to assess the efficacy of trauma-informed training on police response to sexual and domestic violence. At the time of data collection, the police partner agency employed approximately 5,300 commissioned police personnel and 1,200 citizens, operating one central station and 14 substations. 3
The current sample consisted of redacted DV case files obtained through a multistage process in cooperation with law enforcement supervisors in the agency’s Special Victims’ Division. Sampling decisions were determined in collaboration with the Principle Investigator (PI) and the police partner agency. First, the population of 52,058 family violence case files 4 were generated from January 1, 2014 through August 31, 2016 5 by the police agency and were provided to the PI who created a stratified-random sample to include 125 cases in each category by disposition: (1) cleared—arrest, (2) cleared—other, (3) active, and (4) inactive. 6 In addition, the population of 119 unfounded family violence case files were added to the stratified-random sample for a total of 619 family violence incident numbers selected and provided to the police agency, who collected and redacted identifying victim, suspect, and officer information. Hard copies were provided to the PI and case files were coded. Quantifiable data were extracted and entered into an SPSS spreadsheet. Once data were cleaned, incidents involving a single, adult suspect and a single, adult victim were selected for inclusion in the analyses (n = 459). For the purpose of the present study’s focus on police service referral and advocate involvement in cases of DV, all incidents involving a current or former intimate partner were retained for analyses (n = 380). 7 While it is important to assess police service referral and advocate involvement in DV incidents among all relationship dyads, data in the present study included only 10 case files (2.6%) that involved same-sex partner violence and thus, were not retained for quantitative analysis. Among the remaining case files, two cases had missing data on victim and suspect race. The final sample was comprised of 368 DV cases.
Extraction of Quantitative Data
Data were entered into an SPSS 25.0 spreadsheet to capture victim, suspect, and case characteristics. Each case file allowed for the extraction of detailed, quantifiable information that included the responding officer’s initial incident report, victim, suspect, and witness information, accounts of the incident, whether or not the case was assigned to an investigator and subsequent reports, description of evidence, injury, victim cooperation, criminal history, and case disposition. Quantitative coding of each case file took 35 to 60 minutes, depending on the complexity of each incident and length of each case. Creation of variables relied upon affirmative, explicit mention of measures of interest (see Garza et al., 2020).
Dependent Variables
In line with Wilson and Segrave’s (2011) three typologies, the participating police partner operates a hybrid approach to victim service delivery. Specifically, the agency follows an informal policy that emphasizes the importance of front-line officers (patrol and responding officers) providing victims with detailed information about referral services. During the initial DV incident, responding officers are encouraged to provide a tangible paper card with information detailing crime victims’ rights, local crisis shelters, hotline numbers, and victim services housed within the police department. Police officers at this agency also rely on embedded victim advocates employed by law enforcement and cross-trained to respond to sex crimes and DV offenses. To capture the police partner’s hybrid approach, two dependent variables were created. First, a binary measure of service referral captured whether the responding officer or specialized DV investigator provided victims with a service referral card (No = 0, Yes = 1). Second, advocate involvement was a binary measure that captured whether an embedded advocate at the police agency was involved with the DV incident at any point from the initial call for service through final case disposition (No = 0, Yes = 1).
Independent Variables
Victim characteristics
Five variables were included to represent victim characteristics. Binary measures were used to capture victim sex (Male = 0, Female = 1) and victim race (White = 0, Person of Color = 1). Victim age was a continuous measure that captured the age of the victim at the time of the report. A binary variable captured whether the victim and suspect were married (No = 0, Yes = 1). Finally, victim general criminal history captured a formal criminal history check conducted by the responding officer and/or investigator. Any affirmative, explicit mention of this in the case file was captured through a binary variable (No = 0, Yes = 1). 8
Suspect characteristics
Four variables were included to represent suspect characteristics. Suspect age was a continuous measure that captured age at the time of the report. Suspect alcohol/drug use was a binary variable that captured explicit, affirmative mention by the responding officer, victim, suspect, or witnesses of any alcohol or drug use by the suspect in the presenting DV incident (No = 0, Yes = 1). Suspect history of DV captured whether the suspect previously: (1) exhibited any abusive behavior, (2) had prior arrests for partner abuse, or (3) was convicted of prior partner violence. This was captured through explicit, affirmative mention in the case file and was coded as a binary variable (No = 0, Yes = 1). Responding officers and investigator routinely conduct criminal background checks on suspects. Any explicit, affirmative suspect general criminal history mentioned in the case file was captured as a binary variable (No = 0, Yes = 1). 9
Case characteristics
Six variables were included to reflect case characteristics. Physical evidence was a binary variable that captured whether the responding officer or investigator collected any tangible, physical evidence such as biological evidence, photographs, weapons, sworn statements, electronic text messages/voicemails/calls, or destroyed personal property, among others (No = 0, Yes = 1). The presence of children was captured through explicit, affirmative mention in the case file and coded as a binary variable (No = 0, Yes = 1). Physical altercation captured incidents of DV that involved a physical attack as opposed to electronic or verbal threats/abuse and was coded as a binary variable (No = 0, Yes = 1). Suspect weapon use, victim injury, suspects use of threats to kill, and stalking are often used to reflect the severity of partner abuse (Dichter et al., 2011; Garza et al., 2020; Twis et al., 2018). Weapon use, which excluded bodily force, was captured through explicit, affirmative mention that the suspect threatened or used a weapon in the incident and was coded as a binary variable (No = 0, Yes = 1). Victim injury was a binary variable that captured any type of physical injury explicitly mentioned in the case file by the victim and/or police officer. Examples of physical injuries included bruising, cuts, swelling, bite marks, broken teeth, and bleeding, among others (No = 0, Yes = 1). Suspects threats to kill victim and/or children was captured through affirmative, explicit mention in the case file and coded as a binary variable (No = 0, Yes = 1). Finally, prior research has defined stalking as a pattern of unwanted pursuit contact behavior (Garza et al., 2020). Previous stalking behavior captured whether the suspect had “telephoned,” “waited outside or inside of places,” “watched from afar,” “followed,” “sent letters,” “emailed,” “sent gifts,” “showed up uninvited,” or engaged in “other” unique pursuit behaviors (Fisher et al., 2000, 2002) prior to the presenting DV incident. Coding was based on explicit, affirmative mention of at least two pursuit behaviors in the case file (No = 0, Yes = 1).
Control variables
DV cases that progress into the investigation phase of the criminal justice process may have more exposure to victim advocates and trained investigators with enhanced skills and knowledge. To control for this potential influence, a binary variable captured whether the case was transferred to an investigator (No = 0, Yes = 1).
Analytic Strategy
The analysis proceeded in three stages. First, univariate statistics were calculated to address RQs 1 and 3. Second, multivariate binary logistic regression models were estimated because of the dichotomous nature of the dependent variables. Model 1 regressed service referral on the independent and control variables (RQ2). Model 2, regressed advocate involvement on the independent and control variables on the (RQ4). IBM SPSS Statistics, version 26.0, was used to analyze the data.
Results
Descriptive Statistics
Table 1 presents the descriptive statistics for the variables addressing RQ1 and RQ2. Overall, police officers provided a service referral to victims in 70.7% of the presenting DV cases in this sample. Additionally, embedded advocates were involved in 37.2% of the DV cases in this sample. The majority of victims were female (86.1%) and averaged 32 years of age (SD = 9.84). Just over half of the victims were People of Color (56.5%) and a quarter of victims were married to the suspect (25.5%). Suspects averaged almost 40 years of age (SD = 10.07) and the majority of suspects did not use drugs or alcohol in the current incident (83.2%). More than one third of suspects had previously engaged in abusive behavior (38.6%) and had a general criminal history background (38.6%). Over half of the cases did not have any physical evidence (59.8%) or victim injuries (50.8%). A child was present in a quarter of DV incidents (25.0%). The majority of cases were physical altercations (73.6%) and only 20.4% involved a weapon. Finally, the majority of suspects did not threaten to kill the victim and/or children (73.4%) or engage in prior stalking behaviors (82.3%).
Descriptive Statistics (N = 368).
Binary Logistic Regression Models Assessing Service Referral and Advocate Involvement
Table 2 presents the results of the multivariate binary logistic regression model predicting service referral, addressing RQ3. The regression equation for Model 1, which estimated the effects of victim-, suspect-, and case-characteristics on service referral, was statistically significant, χ2(16) = 33.61, p = .006. Suspect age was significant, negative predictor of service referral, where a one-unit increase in suspect age was associated with a decrease in the odds of providing service referral (b = −0.04, OR = 0.96, p = .011). Physical altercation was a significant, positive predictor; DV cases involving a physical altercation, instead of a verbal argument, were associated with an increase in the odds of service referral by 3.3 times
Multivariate Binary Logistic Regression Predicting Service Referral (N = 368).
Note. *p < .05. **p < .01.
Model 2 in Table 3 presents the results of the multivariate binary logistic regression model estimating the effect of victim-, suspect-, and case-characteristics on advocate involvement, which addresses RQ4. The regression equation was statistically significant, χ2(18) = 78.74, p = .000. Victim marital status was a significant, negative predictor (b = −1.00, OR = 0.37, p = .003), where incidents involving married partners were associated with a decrease in the odds of embedded victim advocated involvement when compared to non-married partners. Suspect alcohol and/or drug use was a significant, negative predictor (b = −0.75, OR = 0.47, p = .041) of advocate involvement, where incidents involving suspect alcohol and/or drug use were associated with a decreased odds of embedded advocate involvement, compared to those incidents that did not involve substance consumption. Physical evidence was a significant, negative predictor (b = −0.84, OR = 0.43, p = .003), such that the presence of any physical evidence, compared to no physical evidence, was associated with a decrease in the odds of embedded advocate involvement. Weapon use was a significant, positive predictor (b = 0.72, OR = 2.06, p = .023) such that DV cases where suspects used a weapon were associated with an increase in the odds of advocate involvement by nearly two times as compared to DV incidents with no weapon use. Victim injury was also a significant, positive predictor (b = 0.64, OR = 1.89, p = .045), where any victim injury was associated with an increase in the odds of embedded advocate involvement by nearly two times as compared to incidents where the victim was not injured. Finally, service referral was a significant, positive predictor (b = 1.46, OR = 4.31, p = .000), such that cases involving an officer service referral, compared to no service referral, was associated with an increase in the odds of embedded advocate involvement by 4.31 times. The remaining variables were not significant. 10
Multivariate Binary Logistic Regression Predicting Advocate Involvement (N = 368).
Note. *p < .05. **p < .01.
Discussion
Trauma-informed and victim-centered approaches to partner violence have highlighted the ways police can enhance response to DV calls for service. Among these, research has noted the important role that service referral and advocate involvement can have on positive health outcomes for DV victims and on criminal justice case progression (Camacho & Alarid, 2008; Campbell, 2002; Dawson & Dinovitzer, 2001). Despite the importance of these approaches, few studies have assessed service referral decisions among police officers (but see Belknap & McCall, 1994; Buzawa & Hotaling, 2006; Finn & Stalans, 1995; Goodson et al., 2020; Homat & Kennedy, 1985; Kernic & Bonomi, 2007; Stalans & Finn, 1995, 2000; Sun, 2007; Swerin et al., 2018) or officer decisions to involve an advocate in reported DV cases using incident case file data. Importantly, Wilson and Segrave (2011) argued that police agencies may engage in different types of service delivery to include paper referrals or embedded victim advocates. Even so, empirical research has not assessed predictors of service provision among agencies using a hybrid model approach. The current study relied to schema theory as the overarching theoretical framework to assess (1) the extent to which police officers provided a service referral to DV victims, (2) predictors of service referral by police personnel, (3) the extent to which victim advocates were involved with DV cases, and (4) predictors of victim advocate involvement. Several findings are worthy of discussion.
First, the majority of police officers in this sample provided service referrals to victims of partner violence. Specifically, officers provided service referrals in nearly 71% of DV incidents, indicating that police are engaging in trauma-informed and victim-centered responses, at least in terms of the tangible provision of resources following victimization. It is important to note, that service referral still places the onus of contacting advocates on the victims of partner abuse. This may undermine service use because of the difficulty surrounding access, resources, and victim safety—leaving victims with limited alternatives but to remain in abusive relationships (LaViolette & Barnett, 2013). Alternatively, an embedded victim advocate was involved in just over one-third of DV incidents, despite research that has ranked advocates as the most important type of formal support (Erez & Belknap, 1998). Taken together, these findings are particularly interesting within the context of a hybrid approach to victim service delivery in this agency (Wilson & Segrave, 2011). Results have indicated an increased reliance on tangible service referral as compared to embedded advocacy. While we do not intend to suggest that embedded advocacy is superior to service referrals, it may be that officers perceive differences in the benefits and detriments associated with each mode of service delivery. This is an important point for future research to explore. From a schema theory perspective, it may be that existing beliefs about the role of victim advocates (e.g., due to organizational culture, academy training, and prior experiences) inform how officers perceive advocates, influencing their decision to involve them in DV investigations (Goodson et al., 2020).
Second, to further contextualize service referral among officers, the current study also considered the role of victim-, suspect-, and case- characteristics that predicted service referral decisions. Two factors emerged as significant for officer service referral decisions. An increase in the suspect’s age decreased the odds of service referral among officers. While not ideal, this could represent police perceptions of dangerousness in terms of how age reflects physical strength, stamina, and the ability to produce injury. Results demonstrated that when the DV incident involved a physical altercation between the victim and suspect, the odds of service referral increased by three times as compared to DV cases involving only emotional and/or psychological abuse, reiterating prior research (Kernic & Bonomi, 2007). This finding likely reflects police perceptions of urgency and severity, where physical DV is often perceived as more legitimate, credible, and serious when compared to abuse that does not produce demonstrable physical injury (Dichter et al., 2011; Eigenberg et al., 2012; El Sayed et al., 2020; Twis et al., 2018). El Sayed et al. (2020), for instance, recently found that nearly 500 officers from nine Southern police agencies indicated that evidence of physical trauma was one of the most important factors influencing perceptions of victim credibility. In the present study, it could be that officers may not have interpreted emotional and/or psychological DV as serious enough to warrant service referral. Given this finding, it would be important for future research to further identify and dismantle DV myths among officers. For example, if a substantial number of officers at this agency endorse the myth that emotional abuse is not serious because it does not produce harm, this could inform the content of training protocols designed to inform and better prepare police by focusing on the broader dynamics of partner abuse and the psychological and physiological trauma that this violence produces.
The present study was also concerned with identifying victim-, suspect-, and case- characteristics as predictors of victim advocate involvement. Cases involving married couples demonstrated a decrease in the odds of involving embedded victim advocates compared to cases involving non-married partners (e.g., cohabiting, dating, etc.). In these instances, perhaps DV victims were less likely to reach out for advocacy assistance because of their dependence on the perpetrator and their inability to leave the marital relationship, despite the dangers associated with staying. It could also be that police or specialized investigators, who are tasked with building a case against the DV perpetrator, adhered to schemata about married women in abusive relationships and their decreased cooperation with police efforts (e.g., El Sayed et al., 2020; Fleming & Franklin, 2020)—including embedded victim advocates. To that end, officers may be less likely to engage in efforts to involve victim advocates in cases involving married couples. Additionally, cases with a suspect who used alcohol or drugs in the presenting incident were less likely to involve a victim advocate compared to cases with suspects who were sober. Again, this may speak to long-held DV myths that violence in relationships only happens because of factors like alcohol that facilitate aggression and limit inhibitions among men who would otherwise not behave violently (Koss et al., 1994). These stereotypes have diminished perpetrator responsibility and affected attributions of severity and risk for would-be victims, potentially informing police schema surrounding partner abuse.
The presence of physical evidence also significantly decreased the odds of advocacy involvement. This finding was unexpected, given the particular relevance of physical evidence in formal criminal justice case processing (Hirschel & Hutchison, 2001). It could be that victim advocates were not involved in those cases with sufficient physical evidence for continued case processing. Prior research has suggested that police officers view victim advocates as an instrumental resource in facilitating victim cooperation, engagement, and participation with law enforcement and thus, advocacy assists with case progression (Goodson et al., 2020). If there was physical evidence collected in a DV investigation, however, officers may not view victim advocacy as a necessary tool in accelerating the formal progression of a DV case, despite the important ways that advocacy has encouraged victim safety and well-being. Moreover, both victim injury and weapon use increased the odds of victim advocate involvement. These findings demonstrate the parallel between offense seriousness and perceived notions of victim safety in terms of the utility of support and how a victim advocate can provide both care and service when partner abuse is perceived as severe.
Finally, the strongest predictor of victim advocate involvement in DV incidents was service referral. Results demonstrated initial police officer service provision increased the odds of advocate involvement by more than four times. Within the context of a hybrid service delivery approach, this finding has underscored the importance of consistently providing referrals to DV victims regardless of victim-, suspect-, and case- characteristics. Consistent service referral has the potential to augment advocacy involvement. In other words, victim service provision may be most effective when police officers rely on both modalities. In terms of policy implications, police agency administration and front-line supervisors should place emphasis on the importance of a multifaceted approach that addresses both resources for victims and advocacy to encourage cooperation and enhance victim well-being. Perhaps police response would benefit from agency administration who enact formal policies that require responding officers and specialized investigators to activate both forms of victim services.
While these results add to the broader knowledge on police decisions and service provision for DV victims, this study is not without limitations. Due to the nature of the case file data, the manner in which an advocate became involved in a specific DV incident could not be identified. This means that the use of official data from this agency precluded disentangling if the embedded advocate was contacted by the victim or became involved because a responding officer or investigator deemed certain cases with specific characteristics in need of advocacy. While the study found that police relied more heavily on service referral as compared to embedded advocacy, present findings cannot speak to why. Future research should examine police perceptions of different forms of victim services using qualitative methods to delineate why officers choose to involve advocates. Additionally, the DV cases used in this study were drawn from an urban police department located in a large and diverse U.S. city. It would be useful to identify if this level of service referral and advocate involvement is consistent in police agencies located in more rural and/or suburban locations, those serving more racially and ethnically homogenous populations, and those with fewer DV calls for service.
Police have an opportunity to interact with victims in a trauma-informed manner by providing information pertaining to available resources and advocacy. The importance of service provision and advocate involvement has enhanced victim well-being (Camacho & Alarid, 2008; Campbell, 2008). Findings from this study demonstrate that police officers largely engaged in service referral, but on average, advocates were less involved in this sample of DV incidents. The strongest predictor of advocate involvement was service provision. Collectively, results support the need to continue educating police about the importance of service referral and advocate involvement for DV investigations. This is particularly relevant in terms of police decision-making and victim safety. Xie and Lynch (2017), for example, used longitudinal data from the National Crime Victimization Survey from 1996 to 2012 to assess the effects of formal reporting, arrest, and victim advocacy services on the probability of DV revictimization. Findings suggested that advocacy involvement produced a 40% reduction in risk of revictimization, compared to a 34% reduction in risk of revictimization associated with formal reporting alone. Arrest was not associated with a reduction in risk of revictimization, emphasizing the role of advocacy on positive DV outcomes. Taken together, these findings highlight the importance of police efforts in providing service provision and involving victim advocates as a best-practice in trauma-informed and victim-centered responses to DV.
Footnotes
Acknowledgements
The authors would like to thank the police partner agency for providing access to redacted case files; Alex Updegrove, Jessica Fleming, and Matthew Bills for their assistance with data entry.
Authors’ Note
The opinions, findings, conclusions, and recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of the Department of Justice, Office on Violence Against Women.
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 project was supported by Grant No. 2016-SI-AX-0005 awarded by the Office on Violence Against Women, U.S. Department of Justice.
