Introduction

Since Martin’s (1976) seminal work publicizing the plight of battered women during the second wave of the feminist movement, intimate partner violence (IPV) has remained a serious public health issue with significant implications for social support and justice systems. Estimates have suggested that 37.3% of all US women have experienced lifetime IPV (Smith et al. 2018). While men are not immune from partner abuse, women have been more likely to be victimized, report heightened fear, and have tended to suffer more severe consequences, including adverse physical and mental health outcomes (Twis et al. 2018).

The Centers for Disease Control and Prevention (CDC) have suggested using intimate partner violence (Nicolaidis and Paranjape 2009) to describe behavior defined as, “physical violence, sexual violence, stalking, and psychological aggression (including coercive tactics) by a current or former intimate partner (i.e., spouse, boyfriend/girlfriend, dating partner, or ongoing sexual partner.)” (Breiding et al. 2015, p. 9).Footnote 1 For the purpose of the present study, the terms IPV and partner abuse are used to refer to those current or past relationships characterized by violence and abuse, coercion, control, and manipulation. While men can be both perpetrators and victims of IPV (Tarzia et al. 2020), this analysis focuses exclusively on myths surrounding Johnson’s (2008, p. 7) notion of “intimate terrorism”—or male-perpetrated violence against women.

IPV has remained significantly underreported to formal authorities (Decker et al. 2019; Holliday et al. 2019). Morgan and Oudekerk (2019), for example, reported that only 45% of IPV incidents were reported to police. Limited criminal justice intervention has been attributed to victim perceptions that police have been unable or unwilling to intervene on their behalf (Cho et al. 2019; Decker et al. 2019; Wolf et al. 2003). This has stemmed from a heteronormative, historical narrative that partner abuse was a private family matter to be addressed by male heads-of-household rather than a public issue necessitating formal case processing attention (Buzawa and Austin 1993; Ford 1983; Gracia et al. 2014). Traditionally, law enforcement were reluctant to get involved (Dicker 2008; Freedman 2002). While there have been significant strides in policy and response to IPV(see Lutze and Symons 2003), contemporary shortcomings have included limitations in training, agency policies that may not be sufficiently victim-centered, heightened rates of partner abuse among police personnel, and a general lack of knowledge surrounding the dynamics of these relationships or endorsement of stereotyped ideas surrounding partner violence and IPV victims (Eigenberg et al. 2012; Stinson Sr and Liederbach 2013; Dichter et al. 2011). Taken as a whole, these recent challenges have negatively influenced police response to IPV.

Indeed, endorsement of myths that stigmatize, invalidate, and re-victimize IPV survivors have serious implications for police response (Garza and Franklin 2020). Prior research has demonstrated that an officer’s formal case processing decisions about whether to arrest, charge, or jail a perpetrator have been influenced by attributions of culpability (Franklin et al. 2019b; Garza et al. 2020; Stewart and Madden 1997) and traditional gender role endorsement (Feder 1997). IPV myth endorsement has enhanced these culpability attributions (Erez and Belknap 1998; Goodson 2020; Logan et al. 2006; Twis et al. 2018) and aggravated case attrition though this program of research is dated and has shortcomings, including small sample sizes, poor response rates, and limited generalizability. The present study examines IPV endorsement among law enforcement personnel and advances the literature on police and IPV.

Cultural Myths and Law Enforcement Perceptions of IPV

Myths surrounding violence against women have been defined as widely held, false beliefs that draw from gender stereotypes and rationalize the systems that allow violence to continue (Koss et al. 1994; Lonsway and Fitzgerald 1994). Feminist theory has proffered that these myths have facilitated gender inequality, limited women’s movement and mobility, and justified violence against women. These myths are inextricably linked to a male-identified, male-dominated, and male-centered structure responsible for the systematic oppression of women (Johnson 2005). IPV myths place blame on the victim, exonerate the perpetrator, and minimize violence (Peters 2008), thus undermining the seriousness of women’s experiences and silencing their collective voice.

Cultural stereotypes surrounding gender violence have influenced the law enforcement response to IPV as police have not been immune from these social messages (Garza and Franklin 2020; O’Neal 2019). Police have been historically characterized as “biased, inconsistent, and inadequate” in terms of policies surrounding domestic violence and police enforcement (see Trujillo and Ross 2008, p. 455). These descriptors are concerning for both victims and perpetrators of IPV (Twis et al. 2018). Maladaptive attributions have also influenced IPV case attrition (Belknap 1995). Police perceptions of IPV as a “real crime” that requires formal criminal justice intervention have varied. Police may believe that arrest is not always the appropriate response (Sinden and Joyce Stephens 1999; Toon et al. 2005). In contrast, Friday et al. (1991), found support for mandatory arrest laws surrounding IPV among officers, but also revealed that officers viewed mandatory arrest as ineffective. Robinson and Chandek’s (2000) analysis of 471 case files demonstrated twice as many arrests when both the victim and the perpetrator were inebriated compared to counterparts. Arrest was more likely when witnesses were present and the incident involved a shared residence.

Furthermore, Johnson (2004) reported that many officers expressed frustration and disillusionment with IPV victims. Existing research has suggested officers struggle with cynicism and limited comprehension of victim behaviors (e.g., Lockwood and Prohaska 2015), particularly Walker’s (1977) cycle of violence and a victim’s inability to leave an abusive relationship. Lack of understanding regarding the complex challenges inherent in violent relationships have produced derogatory views and attribution of blame directed toward IPV victims (Goodson 2020). Frustration seems to have emerged from subtle victim-blaming myths, where officers believed that victims should leave when law enforcement is activated (Johnson 2004). Rigakos’ (1995) study of Canadian police officers found overt victim blaming among their sample, reporting perceptions of IPV victims as calculating and deceitful and perpetrators as, “victims of adverse circumstances.” Saunders’ (1995) study assessing police response to IPV in a sample of 111 officers in Wisconsin reported some adherence to IPV myths. Participant attributions included feeling uncomfortable interacting with IPV victims, that IPV is sometimes justified, and that victims who remained in abusive relationships did so because they enjoyed the abuse (Saunders 1995).

Additional analyses of police have reported myth adherence. Farris and Hollman (2015) used survey data from elected sheriffs and reported 46% of participants believed victims could easily leave their abusive relationship but opted not to. Gover et al. (2011) reported similar findings from a police at a large urban agency in a Western state. Nearly three-quarters of participants subscribed to victim-blaming myths. Using a sample of 315 officers from one midsized police department, Logan et al. (2006, p. 1368) reported police participants endorsed the belief that, “domestic violence is a crime that should be handled by treatment.” Logan et al.’s (2006) findings reiterated Belknap’s (1995) earlier conclusions where police indicated a preference for handling IPV with mediation instead of arrest. Conversely, Logan et al.’s (2006) participants reported formal sanctions as a more appropriate response to IPV perpetrators who abused substances compared to counterparts. Related, Farris and Hollman (2015) and Gover et al. (2011) found that officers subscribed to the fallacy that alcohol and drug abuse has caused IPV and therefore believed harsher punishments were fitting when alcohol was part of the presenting IPV incident.

Purpose of the Present Study

When responding to IPV, officer attributions of IPV myths can affect a victim’s experiences, formal investigation outcomes, and public safety. Early research established the concerning influence of IPV myth adherence among police. This research has been limited, however, in that much of this literature is dated, has used small samples, surveyed police personnel outside the US, or has reported measurement limitations in terms of operationalization of IPV myths using single item measures. Since cultural stereotypes surrounding IPV can influence police response and subsequent victim cooperation in the formal justice process, further assessment of IPV myth endorsement is fruitful for addressing these limitations. The current study advances the literature on police endorsement of IPV myths through the use of a survey administered to police personnel commissioned in a large, urban police department located in one of the five most populous and diverse US cities.

The present study assessed the following two research questions:

  • RQ1: To what degree do police officers endorse IPV myths?

  • RQ2: Do officer occupational characteristics (years of service, previous family violence response, prior specialized training, perceptions of preparedness in responding to IPV incidents), or attitudinal factors (trauma misperceptions) predict endorsement of IPV myths, controlling for officer demographics (race, sex, education)?

Methodology

Site Location and Participants

Data were collected from police officers in August 2016 as part of a larger federally-funded grant project awarded by the Office of Violence against Women (OVW). Survey administration took place in a metropolitan police agency located in one of the fifth most populous and diverse US cities.Footnote 2 At the time of data collection, the police partner employed 5300 commissioned officers and approximately 1200 civilians. A purposive sample of roll-call times were selected. Roll call times were held at 6:00 am-7:00 am, 2:00 pm-3:00 pm, and 10:00–11:00 pm. Data were collected during 55 roll call meetings at each of the 14 metropolitan police substations. Reminder announcements were made via email by police Lieutenants from the Special Victims’ Division prior to the scheduled survey administration date. Pencil-and-paper surveys were administered to commissioned officers who were present for roll call after reading an institutional review board (IRB) approved description of the study, highlighting the voluntary and anonymous nature of survey participation.Footnote 3 Completion took approximately 25 min, questions were presented in set order, and the survey was described as “Police Attitudes about Crime and Victimization.” Altogether, 694 surveys were administered and 633 surveys were returned, for a response rate of 91.2%. Of those returned, 523 surveys were retained. Cases excluded from the analyses were: 1) nine participants that either marked the same answer all the way through the survey or sections of the survey where variation was expected, or 2) responded to very few of the survey questions. In total, 101 surveys had missing data on pertinent variables for the present analysis and were excluded from the sample.

Participant Demographics

Table 1 presents the demographic characteristics of participants and descriptive statistics of study variables. The mean age of police participants was 38.12 years old. Participants averaged 11.82 years of service. Men represented the majority of the sample (n = 466, 89.1%), compared to women (n = 57, 10.9%). Participants were racially diverse, such that 38% (n = 199) were White, 27.5% (n = 144) were Latinx, 23.5% (n = 123) were Black, 8.0% (n = 42) were Asian/Pacific Islander, 0.4% (n = 2) were Native American/Alaska Native, and 2.5% (n = 13) identified as other. Approximately 40.1% (n = 209) of participants reported having a four-year degree, nearly 28.1% (n = 147) reported some college, and 12.0% (n = 63) reported completion of graduate school.

Table 1 Descriptive statistics of study variables (n = 523)
Full size table

A comparison of the sample to the agency population from which it was drawn revealed similarities across available demographic and occupational data reported by the police partner for 2016. Men represented the majority of commissioned police personnel (84.25%). While the agency population was racially and ethnically diverse, a direct comparison to the sample revealed some differences. White officers represented just under half of the commissioned police force (45.02%), followed by Latinx (26.74%), Black (21.45%), Asian (6.64%), and Native American (0.15%). Personnel averaged 18.05 years of service with a mean age of 47 years old. Personnel at the rank of “police officer” averaged 35 years old.

Measures

Dependent Variable

Endorsement of IPV myths was measured using Peters’ (2008) 18-item Domestic Violence Myth Acceptance Scale. Items were captured on a 6-point Likert-type scale from 0 (strongly disagree) to 5 (strongly agree). Items were subjected to exploratory factor analysis (EFA) with maximum likelihood estimation (MLE) and verimax rotation. Similar to Peters’ (2008, p. 15) conclusions, the present analysis found that “the factor structure of the instrument was highly unstable.” Next, a confirmatory factor analysis with MLE produced one primary factor. The scale was employed as a single 16-item index with loadings ranging from .675 to .351. Items were summed to create a scale that ranged from 0 to 80; increased values represented higher levels of IPV myth endorsement. Internal consistency reliability was excellent (α = .850).

Independent Variables

Family violenceFootnote 4 calls for service response to in the previous 12 months was measured with one item (Goodson et al. 2020; see also Garza and Franklin 2020), “How many family violence calls have you responded to in your current position in the last 12 months?” Responses were captured on an ordinal scale (None = 0, 1 to 5 = 1, 6 to 10 = 2, 11 to 20 = 3, 21 or more = 4).

Participant perceptions of preparedness in responding to family violence incidents was measured with one item (Garza and Franklin 2020; Goodson et al. 2020), “How prepared do you feel to respond effectively to calls for service for domestic violence involving intimate partners?” Responses were captured on a 6-point Likert-type scale from 0 (very unprepared) to 5 (very prepared); higher numbers represented perceptions of increased preparedness.

Prior specialized training was measured with three binary items (No =0, Yes =1) used in prior research (Garza and Franklin 2020). Three items included having received, “any specialized training on victim sensitivity?,” “any special training on the trauma of victimization?,” and “specialized training on crime victims’ reactions and behaviors in dealing with their victimization?” Three items were subjected to EFA with MLE, which produced one factor with an Eigenvalue greater than 1 that accounted for 68.86% of the variance. Factor loadings ranged from .769 to .885. Items were summed to create an index from 0 to 3; increased values represented increased participation in specialized training. Internal consistency reliability was excellent (α = .867).

Trauma misconceptions was measured using 9-items from Ask’ (2010) Attitudes Toward Crime Victims index that included statements measuring stereotypical trauma expectations such as hysteria, behavioral expressiveness, and emotionality as a sign of truth-telling. Responses were captured on a 6-point, Likert-type scale from 0 (strongly disagree) to 5 (strongly agree). Items were subjected to EFA with MLE and verimax rotation, which produced one primary factor with an Eigenvalue greater than 1 accounting for 32.49% of the variance. Factor loadings ranged from .547 to .671. Items were summed to create an index from 0 to 35; increased values represented increased trauma misconceptions (see also Franklin et al. 2019a). Mean values on individual items ranged from 2.08 to 2.91 and standard deviations ranged from 1.22 to 1.40 indicating adequate variability. Internal consistency reliability was excellent (α = .822).

Demographic Control Variables

Four measures were included as control variables. Participant sex was a dichotomous variable (Men = 0, Women = 1). Participant race/ethnicity was captured categorically and then recoded into four binary variables (White, Black, Latinx, Asian). Participant education was captured as an ordinal variable (High school diploma some college, two-year degree, four year degree, graduate school). Participant years of experience was a continuous variable.

Analytic Strategy

The analysis proceeded in three stages. First univariate statistics, means, and standard deviations were calculated for each of the study variables. Next, a bivariate correlation matrix presented statistically significant relations between the study variables. Finally, a multivariate Ordinary Least Squares (OLS) regression model was estimated to test the research questionsFootnote 5 (Tabachnick et al. 2007).

Results

Univariate Statistics and Bivariate Correlation Matrix

Participants reported low levels of IPV myth endorsement (RQ1), scoring well below the scale midpoint (M = 19.03, SD = 11.24; see Table 1). Pearson bivariate correlations are presented in Table 2. Several significant relations emerged between independent and dependent variables. There was positive, significant relation between trauma misconceptions and IPV myth endorsement, r(521) = .300, p = .001, and between family violence response and IPV myth endorsement, r(521) = .100, p = .005. Results also demonstrated an inverse, significant relation between IPV myth endorsement and officer sex (Men = 0, Women = 1), r(521) = −.110, p = .005. The remaining variables related to IPV myth endorsement were not statistically significant.

Table 2 Summary of bivariate correlations for study variables
Full size table

In evaluating relations between independent variables, several preliminary findings emerged. Prior general training was significant and positively related to perceptions of preparedness in responding to family violence involving intimates, r(521) = .088, p = .005, and to educational attainment, r(521) = .157, p = .001. Trauma misconceptions was statistically significant and inversely related to officer years of service, r(521) = −.108, p = .005. Perceptions of preparedness in responding to family violence was statistically significant and positively related to officer education, r(521) = .090, p = .005. Additionally, significant inverse relations emerged between officer years of service and officer sex, r(521) = −.114, p = .001 and between officer years of service and family violence response, r(521) = −.353, p = .001. Officer education was positive and significantly related to officer sex, r(521) = .135, p = .001, and officer education emerged as significant and inversely related to family violence response, r(521) = −.126, p = .001.

Multivariate OLS Regression Model

Table 3 presents the results of the multivariate OLS regression model predicting IPV myth endorsement, where independent and control variables were entered simultaneously (RQ2). IPV myth endorsement was regressed on occupational, attitudinal, and demographic characteristics. The regression equation was significant, R2 = .134, F(10, 512) = 7.95, p = .000, and explained 13.4% of the variance in IPV myth endorsement. Family violence response was a statistically significant, positive predictor of IPV myth endorsement, b = .802, t = 2.032, p = .043, suggesting increased experience in responding to family violence incidents was correlated with increased IPV myth endorsement. Perceptions of preparedness in responding to family violence was also a significant, inverse predictor of IPV myth endorsement, b = −1.320, t = −2.347, p = .019, where decreased perceptions of preparedness correlated with increased IPV myth endorsement. Trauma misconceptions was a statistically significant, positive predictor of IPV myth endorsement b = .496, t = 6.759, p = .000, such that increased trauma misconceptions correlated with increased IPV myth endorsement. Finally, officer sex (Men =0, Women = 1) was a significant, inverse predictor of IPV myth endorsement, b = −3.476, t = −2.283, p = .023, such that women reported significantly decreased endorsement of IPV myths compared to men.

Table 3 Multivariate OLS model predicting IPV myth endorsement
Full size table

Discussion

There has been a notable dearth of scholarship on the endorsement of IPV myths among law enforcement personnel and existing research has shortcomings. The present study advanced the program of research through the use of 523 paper-and pencil surveys administered to a purposive sample of police officers commissioned at an urban police department located in one of the fifth most populous and diverse US cities to examine occupational, attitudinal, and demographic predictors of IPV myth endorsement. Four findings are worthy of additional discussion.

First, results indicated relatively low mean IPV myth endorsement among police participants; a finding that supports recent research (Farris and Hollman 2015; Garza and Franklin 2020; Twis et al. 2018). Any endorsement of IPV myths is problematic for victims who formally present to police to report victimization, given the unique position of police officers as “gatekeepers” of the formal criminal justice process (Kerstetter 1990; LaFree 1989) and their ability to facilitate an investigation, validate a victim’s experience, and mitigate trauma through positive responses to formal disclosure (Eigenberg et al. 2012; Farris and Hollman 2015; Twis et al. 2018).

Second, participants with more experience in responding to family violence incidents reported increased IPV myth endorsement. This may be the result of cynicism from repeated exposure to partner abuse cases where victims do not behave according to expected norms in domestically violent relationships (e.g., LaViolette and Barnett 2014). There has been much confusion and misunderstanding among the general population regarding why women remain in abusive relationships, why women formally disclose to police and then opt not to cooperate with investigative efforts, or return to an abusive partner after successfully leaving. Additional efforts to better understand this relationship in the form of qualitative interviews or focus groups with police would be fruitful for clarifying the relation between increased experience in responding to family violence incidents and increased IPV myth endorsement.

Findings from the present study also revealed women reported significantly decreased endorsement of IPV myths compared to men. These findings underscore the need for targeted change directed toward cultural misconceptions surrounding IPV. Furthermore, this finding is consistent with existing research on police officer samples examining rape myth endorsement (Feild 1978; Garza and Franklin 2020; Page 2007, 2008; Rich and Seffrin 2012) and more generally among community and college samples (Hockett et al. 2016; Franklin and Garza 2018; Suarez and Gadalla 2010) where women are less likely to adhere to cultural stereotypes that neutralize violence against women. As sex was a significant predictor of IPV myth endorsement and women reported significantly decreased IPV endorsement, it would be beneficial to target the hiring of women across law enforcement agencies to dilute broader organizational cultures that harbor adverse myths and misconceptions about IPV. (Meier and Nicholson-Crotty 2006).

Fourth, results demonstrated that increased trauma misconceptions correlated with increased IPV myth endorsement (see also Garza and Franklin 2020; Franklin et al. 2019a, b). This finding illustrates that officers who endorsed trauma misconceptions also endorsed a suite of underlying misperceptions that may be particularly dangerous to victims who present to police for formal assistance centering around behavioral displays of hysteria and emotionality as evidence of truth-telling. Trauma misconceptions have had negative consequences for gender violence victims when victims present with flat affect or emotional numbing and, coupled with IPV myth endorsement, may be especially problematic for IPV victims disclosing to police who may view their disclosures as deceptive (see Franklin et al. 2019a, b for a review). Prior literature has noted that poor training or lack of training can facilitate endorsement of these stereotypes (Gover et al. 2011; Johnson 2004; Logan et al. 2006; Saunders 1995). Traditional educational programming that centers on dismantling cultural myths may not produce immediate or prolonged behavioral change among program completers, but may create long-term shifts in organizational culture through transmission of informal norms and values.

Sleath and Bull (2012) have noted the utility of implementing training protocols among agencies for the benefit of cultural and organizational transformation over time. Indeed, the organizational climate and informal norms of a law enforcement agency may take generations to evolve in ways more accepting of attitudes that reflect the reality of trauma and IPV. For this reason, training is not without value. Further, while educational programming in police settings has reported deficits, recent evaluation research has underscored the role of a 40-h comprehensive programming on rape and sexual assault among police personnel (Campbell et al. 2019), producing both short- and long-term differences in rape myth endorsement and trauma-related attitudes (see also Franklin et al. 2019a, b).

Finally, findings from this study revealed a relation between decreased perceptions of preparedness among police participants in responding to family violence response and increased IPV myth endorsement. It appears that the way police personnel perceive their role and accomplish their duty to “protect and serve” may correlate with their acceptance of myths that neutralize the seriousness of IPV, excuse the perpetrator, and blame the victim. This relation highlights the worth of programming designed to shift cultural and systemic mechanisms among law enforcement agencies over time.

Limitations and Future Directions

Despite the importance of the results presented here, findings are not without limitations. First, 17.4% of data were missing, however, this is the first study of its kind and provides an instructive starting point for future research. Women were under-represented in the present sample, though representation in this sample in consistent with the agency’s population demographics. Future research should employ stratified sampling techniques to encourage responses among women in police agencies. Additionally, data comprised responses from a purposive sample of police officers commissioned in a large, metropolitan police agency and results should be interpreted accordingly. Future studies should replicate this study among police from smaller and rural agencies with more homogenous populations to examine occupational, attitudinal, and demographic characteristics that may predict IPV myth endorsement. It is important to note the potential role of social desirability bias given the underreporting of undesirable viewpoints, particularly regarding sensitive questions (Tourangeau and Yan 2007), such as those pertaining to IPV myths. That said, the measures capturing IPV myth endorsement among this sample are a conservative estimate. Notably, for the purpose of this study, type of family violence response (intimate-partner, elder, child, etc.) was not specified in the survey instrument as this was a mechanism of the police partner agency offense codes. Furthermore, the IPV myth endorsement scale used in this study has relied on heteronormative language and a heterosexist narrative surrounding partner abuse. Future studies should consider IPV myth endorsement using inclusive and culturally-sensitive language that considers relationship violence among sexual minorities, transgender couples, and gender non-conforming couples. Related there may be additional factors unaccounted for in the present analysis, including personal experience with IPV. Findings should be interpreted with some caution.

Limitations notwithstanding, these results have important implications for practice and potential expansion for future research. Violence against women presents a challenge for law enforcement personnel, especially given the chronic and hidden nature of these crimes and the reluctance of victims to disclose IPV to law enforcement. While police response to gender violence has evolved, challenges still remain. Existing research has noted shortcomings in police response to IPV—which may be preceded by officer myths surrounding IPV incidents and IPV victims in particular (Buzawa and Buzawa 1992, 2003; Ferraro 1989; Hilton 1993; Johnson 2004; Sherman et al. 1992). The present study was the first of its kind to examine occupational, attitudinal, and demographic predictors of police officer IPV myth endorsement using myths from Peters’ (2008) scale. The results presented here demonstrated that when police officers endorse myths surrounding “appropriate” victim behaviors, officers may also endorse trauma misconceptions and may be ill-equipped to respond to partner violence incidents. Findings from the current study highlight the importance of IPV myth endorsement as a noteworthy area for continued agency attention.