Abstract
As research continues to describe negative experiences and high case attrition within sexual assault cases reported to the police, it is important to better understand the role of first-responding police officers. This study surveyed a sample of sworn police officers (N = 174) from one department in a midsized city in the Great Lakes region to examine the effect of individual police officer characteristics, rape myth acceptance (RMA), attributions of blame, and case characteristics from a hypothetical vignette, on officer perceptions of a “good” case and behavioral intentions. Results found that although RMA predicts both perceptions of the case and behavioral intentions to respond in ways indicative of perceived seriousness, RMA is mediated by attributions of suspect blame. Victim alcohol use was found to decrease an officer’s likelihood of responding more vigorously, showing less likelihood of calling a detective or arresting the suspect, if identified. Future research should begin to connect not only case characteristics but also individual police officer characteristics, attitudes, and attributions of blame, with real case outcomes, in addition to hypothetical scenarios.
Although research and advocacy has addressed sexual assault for decades, some research argues that police responses to sexual assault have not significantly improved over the past 30 years (Jordan, 2001; Temkin & Krahé, 2008). Research clearly indicates the negative and traumatizing experience of many sexual assault victim-survivors when they report the crime to the police (Campbell, 2006; Campbell, Wasco, Ahrens, Sefl, & Barnes, 2001; Chen & Ullman, 2010; Felson & Pare, 2008; Frazier & Haney, 1996; Frohmann, 2002; Kaukinen & DeMaris, 2009; Konradi & Burger, 2000; Larcombe, 2002; Martin & Powell, 1994; Monroe et al., 2005; Patterson, 2011; Rich & Seffrin, 2012; Ullman & Townsend, 2007). Ullman and Townsend’s (2007) study found victim-blaming statements by the police as the most significant challenge to the work of rape crisis center staff. Research points to high levels of secondary victimization among victim-survivors who decide to report a sexual assault to law enforcement (Alderden & Ullman, 2012; Campbell, 1998; Campbell & Raja, 1999; Campbell et al., 2001; Madigan & Gamble, 1991; Martin & Powell, 1994). Campbell and colleagues (2001) found that more than half of sexual assault victims viewed contact with the legal system as insensitive. Victims report that characteristics of their case, such as knowing the suspect, corresponded with police officers expressing disbelief, being lenient toward the suspect, and not taking the report seriously (Felson & Pare, 2008).
Police officer interactions with sexual assault victim-survivors have the potential to affect not only the victim-survivor but also the case outcome. Case attrition, linked to police and prosecutorial decision making, continues to be high within reported sexual assaults (Campbell, 1998; Campbell et al., 2001; Coates, 1996; Du Mont, Miller, & Myhr, 2003; Frazier & Haney, 1996; Patterson, 2011; Spohn, Beichner, & Davis-Frenzel, 2001; Spohn, White, & Tellis, 2014). Of reported sexual assaults, approximately 12% to 35% result in arrest and 9.7% result in felony charges (Alderden & Ullman, 2012; Spohn & Tellis, 2011). A recent study reported that approximately one third of female sexual assault reports in 12 New Hampshire communities stated that the victim decided not to pursue charges after the initial interview, although a clear rationale for this decision was not included in the police report (Murphy, Edwards, Bennett, Bibeau, & Sichelstiel, 2014).
Police officers have been labeled “gateways to justice” because of their discretion in case processing (Kerstetter, 1990). Discretion is present from the decision to found or unfound a case, to the decision to make an arrest of a suspect (Frazier & Haney, 1996; Kerstetter, 1990; Lord & Rassel, 2002). In small municipalities, patrol officers may exercise even more discretion because they may collect detailed information collected by detectives in other settings (Schwartz, 2010). One study focusing specifically on case outcomes looked at official records (N = 220) from a large Arizona city and found that both legal and extralegal factors influence police officer’s immediate response, specifically the identification and arrest of a suspect (Tasca, Rodriguez, Spohn, & Koss, 2013). Identification of a suspect was less likely when the assault was by a stranger, when multiple suspects were involved, and when officers exhibited some victim blame or did not believe the victim. Arrest of a suspect was more likely in stranger cases and cases with forensic evidence. Failure to arrest was predicted by late reporting and perceived lack of victim credibility (Tasca et al., 2013).
If research shows high attrition within the already low percentage of sexual assaults that are reported to the police, it is important to explore what happens early on in the reporting process. The present study sought to explain the role of a variety of factors—police rape myth acceptance (RMA), attributions of blame in a particular scenario, police experience and demographic characteristics, and case characteristics—in police officer perceptions of a hypothetical reported sexual assault, and officers’ hypothetical behavioral intentions to respond in a way indicative of perceived seriousness.
Literature Review
RMA
Acceptance of rape myths influences individual perceptions of sexual assault victim-survivors within the broader public and within law enforcement (Campbell & Johnson, 1997; Feldman-Summers & Palmer, 1980; Jordan, 2001; LeDoux & Hazelwood, 1985; Page, 2008; Ullman & Townsend, 2007; Venema, 2016). These biases may influence what is considered a “legitimate rape” and who is considered a “credible victim” (Brownmiller, 1975). Elements of RMA (e.g., sexual assaults only involve strangers, use of weapons, and extensive physical injury) influence police officer perceptions of victims and cases (Bachman, 1998; Chen & Ullman, 2010; Du Mont et al., 2003; Edward & Macleod, 1999; McGregor, Wiebe, Marion, & Livingstone, 2000; Starzynski, Ullman, Townsend, Long, & Long, 2007). However, studies show variability in law enforcement acceptance of rape myths (Campbell, 2005). Mennicke, Anderson, Oehme, and Kennedy (2014) argued that while some research points to a genuine shift in policing culture toward less victim blaming, research using open-ended questions about sexual assault still demonstrate the endorsement of rape myths. Although RMA among police officers varies, the extent to which RMA influences perceptions of cases and behavioral responses is unclear.
Attributions of Blame
Some research suggests that police officers attribute more blame to the victim in certain crimes, such as sexual assault. In comparing police perceptions of victims, Bieneck and Krahé (2011) found that more blame was attributed to the victim and less to the perpetrator in sexual assault compared with robbery. Police officers may consider whether they believe the victim caused or provoked the assault, which would lead to an assignment of responsibility on the victim (Shaver, 1975). Often attributions of responsibility are made on the basis of police officers’ perception of the degree to which the victim caused the event, had control over those causes, should be held responsible for the event, and should be blamed or not (Edward & Macleod, 1999). Factors influencing attributions of blame may include the victim-survivor’s attire or risk-taking behavior prior to the sexual assault (Cohn, Dupuis, & Brown, 2009; Stormo, Lang, & Stritzke, 1997). Research has shown that women who are deemed less respectable are blamed to a greater degree than those considered more respectable (Cohn et al., 2009; Feldman-Summers & Lindner, 1976; Jones & Aronson, 1973; Kanekar, Kolsawalla, & D’Souza, 1981). Sleath and Bull (2012) found that officers doubt victims who wore revealing clothing or were intoxicated when the sexual assault occurred, viewing more favorably those considered “ideal victims.” Jordan (2008) described an “ideal victim” as one whose characteristics would lead them to be regarded as a legitimate victim, someone who would not be viewed as “asking for rape.” For instance, someone who is not intoxicated and is sexually assaulted by a stranger with a weapon and unambiguous violence and injury would be considered an “ideal victim.”
Individual Police Officer Characteristics
Research shows mixed results on the influence of individual police officer demographic characteristics, such as gender, on perceptions and decision making within reported sexual assaults. Bieneck and Krahé (2011) found no gender differences in judgments about the legitimacy of sexual assault cases; however, they found that men were more inclined than women to blame the victim. Jordan (2002) found that RMA among female police officers was just as pervasive as that of male police officers. In a more recent study, Wentz and Archbold (2012) found qualitative support that female officers subscribe to rape myths and victim blaming more than male officers; however, no differences existed within quantitative results. In regard to actual decision-making outcomes rather than RMA or attributions of blame, Alderden’s (2008) study was among the first to include police officer demographic characteristics in the model testing extralegal and evidentiary characteristics in sexual assault case decision making. Alderden and Ullman (2012) concluded that sexual assault cases involving male detectives were more likely to result in arrest.
Research has shown a relationship between education, experience, and RMA; however, it is unclear the extent to which education and experience influence perceptions of actual or hypothetical cases, as opposed to general attitudes about sexual assault. Education and professional experience has also been associated with RMA (Campbell & Johnson, 1997; Page, 2007). Page’s (2007, 2008) study found a significant difference in RMA by educational level and experience responding to sexual assault cases. Sexual assault-related training has been associated with lower RMA (Ask, 2010; Campbell, 2005); however, very little research has demonstrated the effect of training on perceptions and behavior within hypothetical or real sexual assault cases (Rich & Seffrin, 2012). In a sample of 429 police officers, Rich and Seffrin (2012) found that sexual assault training was a significant predictor of interviewing skill; however, training was less important than attitudes.
Case Characteristics
In contrast to the effect of one’s individual RMA, a number of studies have examined factors related to the sexual assault itself that may influence perceptions and decision making. Often, these are considered “extralegal” factors because they relate to a broader cultural understanding of sexual assault that does not match current legal definitions; these may include victim and suspect characteristics, credibility factors, and characteristics of the sexual assault itself (Frazier & Haney, 1996). Cases that conform to stereotypes of “real rape” (Estrich, 1987) and “genuine victims” (LaFree, 1989) are ones in which a stranger commits an aggravated assault with a weapon. Much of the literature on sexual assault confirms that there are crimes that evoke substantively different responses from criminal justice officials and potential jurors (Bryden & Lengnick, 1997). Risk-taking behavior at the time of assault refers to behaviors such as using alcohol or illegal drugs, being in a bar alone at night, walking alone late at night, willingly accompanying the suspect to his or her residence, and inviting the suspect to his or her residence.
Substance use by the victim is one extralegal factor that maintains a strong influence on individual perceptions of sexual assault. Not only does alcohol use often correlate with the occurrence of sexual assault, but the presence of alcohol in real and hypothetical sexual assaults also influences perceptions of the crime, victim, and suspect, both among the lay population (Maurer & Robinson, 2008) and among criminal justice system actors (Campbell, 1998; Chandler & Torney, 1981; Frohmann, 1997; Kerstetter, 1990; Schuller & Stewart, 2000; Spears & Spohn, 1996). Negative evaluations, including victim blame and diminished perceptions of victim credibility, are more likely when the victim was intoxicated at the time of the assault (Bieneck & Krahé, 2011; Schuller & Stewart, 2000; Spears & Spohn, 1996). Alcohol and drug use by the victim also significantly deters case progression (Campbell, 1998; Chandler & Torney, 1981; Frohmann, 1997; Kerstetter, 1990; Spears & Spohn, 1996). Campbell (1998) found that victims who were drinking prior to the sexual assault were 4 times more likely to have their cases dropped during the beginning of the legal process.
Another important and “extralegal” sexual assault case characteristic that influences case processing focuses on the relationship between the victim and the suspect. Studies show that police often do not perceive nonstranger assaults as rape (Feldman-Summers & Palmer, 1980; LeDoux & Hazelwood, 1985). Bieneck and Krahé (2011) found that police attributed more blame to victims when the sexual assault did not involve a stranger. An Australian study found that individuals attribute more blame to acquaintance rape victims because they are judged to violate either gender or rape victim stereotypes (Masser, Lee, & McKimmie, 2010). Some studies show that sexual assault cases are less likely to progress through the criminal justice system if the victim and suspect had some relationship, whether being married, having a prior sexual relationship, or even being acquaintances (Buzawa, Austin, & Buzawa, 1995; Chandler & Torney, 1981; Kerstetter, 1990; LaFree, 1980). Other studies do not show this effect (Bachman, 1998; C. Spohn & Horney, 1993; Spears & Spohn, 1996, 1997). Addington’s (2008) findings add complexity to this research, with results that show higher odds of case clearance (or being solved through arrest) of sexual assault cases with known, nonintimate perpetrators than sexual assault cases with strangers. When looking at prosecutors, R. B. Spohn and colleagues (2001) found similar rates of prosecution for both stranger and acquaintance cases, noting that both types of cases involved victims engaging in “questionable” behavior (e.g., delayed reporting, engaging in prostitution, walking alone at night). Beichner and Spohn (2012) found that the victim’s prior sexual history with the suspect did not emerge as a significant predictor of charging decisions; however, they argued that prosecutors look “downstream” to predict how a judge or juror will view the victim, particularly the victim’s alcohol consumption, the willingness to be in the same room as the suspect, and the victim’s criminal record, all of which would damage the victim’s credibility.
Law Enforcement Peer Culture
Organizational culture or subjective norms within law enforcement, in addition to individual attitudes and characteristics, may also influence police officer perceptions, attributions, and decision making within sexual assault case processing. Normative beliefs about sexual assault may stem from socialization and the broader societal context and cultural definitions of what constitutes sexual assault or “real rape” (Estrich, 1987). Jordan (2002) argued that organizational culture, socialization, and peer pressure may influence personal attitudes, and in police work, for example, individual attitudes may be more reflective of work group attitudes than of other identity characteristics or group memberships (e.g., gender) of individuals.
Although focused on a population other than law enforcement officials, Bohner, Siebler, and Schmelcher’s (2006) study aimed to better understand the influence of perceived norms regarding what constitutes “real rape” on one’s own RMA and subsequent behavior. In this study, others’ perceived RMA acted as a social norm in that the perception of other’s endorsement of rape myths influenced the participant’s individual beliefs or RMA (Bohner et al., 2006).
Police officers and prosecutors build cases to meet probable cause, but more importantly, reasonable doubt standards (Martin, 2005). Frohmann (1997) has described this as the “downstream orientation” of decision makers, arguing that prosecutors consider the judge, jury, and defense attorney’s reaction to the case (p. 535). Spohn et al. (2014) extended this downstream orientation to the work of detectives who consider the prosecutor’s response when making decisions. One may wonder whether this downstream orientation exists within first-responding officers who interview the victim and write the report. Do they also consider the decision making of detectives and prosecutors when interviewing the victim-survivor and writing the report? Does this “downstream orientation” or law enforcement peer culture influence their perceptions and decision making in the case?
Current Study
Although studies are beginning to look at the influence of legal and extralegal factors found within police records, few studies examine the combination of individual police officer characteristics (e.g., gender, attitudes), case characteristics (e.g., victim-survivor alcohol use), and contextual measures (e.g., perception of peers), with officer perceptions and decision making. Do characteristics of the officer influence the ways in which the case is perceived and the subsequent response? Do officer RMA and attributions of blame influence perceptions? What is the connection between officer characteristics, attitudes, attributions of blame, and behavioral intentions? It is still unclear what function RMA and attributions of blame perform in police perceptions of the case, which may influence interactions with victims and the course of action taken. In addition, it is unclear the extent to which police officers consider the norms within the police department, or expectations of peers when responding to a case.
This study analyzes a variety of factors that explain police officer perceptions and behavioral intentions in reported sexual assault cases. These factors include officer characteristics (e.g., gender), officer experience (e.g., training), attitudes related to sexual assault (e.g., RMA), case characteristics (e.g., victim alcohol use, relationship between the victim and suspect), and attributions of blame in a hypothetical sexual assault case. Two dependent variables are analyzed: (a) perceptions of a “good” case and (b) behavioral intentions to respond in ways indicative of perceived seriousness. Perceptions of a “good” case include perceptions that the case is a legitimate sexual assault, involves a credible victim, and is a strong case. It should be noted that the phrase “good” case is used throughout this article in reference to the first dependent variable because it reflects language used by police officers in an earlier qualitative phase of this study (Venema, 2016). Cases that were perceived as a legitimate sexual assault, with a credible victim, and some strength of evidence were often called “good” cases by police officers. This in no way implies that any reported sexual assault is “good,” rather that police officers believe there is sufficient reason to pursue the case and believe the case will successfully move through the criminal justice system. The second dependent variable, behavioral intention, involves the intention to call out a detective immediately and arrest the suspect, if identified, indicating perceived seriousness. The following hypotheses were tested:
Method
Participants and Procedures
The participants include regularly sworn police officers recruited from one midsized police department in the Great Lakes region. The sampling frame included 294 officers, of which 174 completed the majority of the survey, a 59.2% response rate. Of the 174 participants, approximately 82.2% were male and 11.7% were female. The sample was predominantly White (85.9%), with 3.6% Hispanic, 2.4% African American, 1.2% Asian, and 6.1% who preferred not to answer. Ages varied, with 1.9% between 20 and 29, 33.5% between 30 and 39, 57.4% between 40 and 49, and 7.1% between 50 and 59. Educational level varied, with most respondents having a college degree as the highest degree (58.8%). An additional 13.9% had a graduate degree and 23.2% had a high school degree, an associate’s degree, or some college.
Participants volunteered to participate in the study, which involved data collection through the online survey software Qualtrics. Participants responded to a variety of measures and were then presented with a vignette. These vignettes mirrored hypothetical situations (albeit with much less detailed description and information) and incorporated important factors that might differentiate the way in which officers perceive and respond to the case (Alexander & Becker, 1978; Sleed, Durrheim, Solomon, & Baxter, 2002). Vignettes are particularly useful in situations that prompt “it depends” responses. Qualtrics randomization assigned one of four sexual assault report vignettes that varied victim alcohol use and prior relationship with the perpetrator. The vignette included the following language:
At 3:00 in the morning, a call came in from dispatch stating that a young woman had reported a sexual assault. You respond to the call by going to the alleged victim’s apartment. Upon arriving, the woman states that she had been at a party the night before for her friend’s birthday. While she was at the party she [had/had not] been drinking and [met a guy/ran into her ex-boyfriend]. [The guy she met/Her ex-boyfriend] said she should go over to his apartment to talk more and have coffee. The woman reports that she agreed to go to his apartment but that [the guy she met/her ex-boyfriend] had sexual intercourse with her even after she asked him to stop.
Measures
Dependent variables
Perceptions of a “good” case
Participants responded to three items based on a 7-point Likert-type scale of 1 (not at all) to 7 (very much). First, “To what extent do you consider this to be a legitimate sexual assault?” Second, “To what extent do you consider this to be a credible victim?” Third, “To what extent do you consider this to be a ‘strong’ case?” These three items were combined to form a scale called perceptions of a “good” case (three items, Cronbach’s α = .85, M = 13.98, SD = 3.6).
Behavioral intentions
Participants responded to two items regarding behavioral intentions suggestive of perceived seriousness of the case on a 7-point Likert-type scale of 1 (not at all) to 7 (very much). One item assessed the support for recommending that a detective respond immediately and the second item assessed the support for recommending arrest of the suspect, if identified. These two items, which were combined to form a scale called behavioral intentions (two items, Cronbach’s α = .64, M = 6.95, SD = 3.1), were frequently described by police officers in an earlier qualitative phase of the study (Venema, 2016) as actions that would be taken if a case were perceived as legitimate and serious.
Independent variables
RMA
The RMA scale is a global measure that assesses general attitudes and beliefs about what constitutes a sexual assault and who is considered a genuine victim. Participants completed a 25-item RMA scale assessed by a 7-point Likert-type scale from 1 (strongly disagree) to 7 (strongly agree), which was developed from both the Illinois Rape Myth Acceptance Scale–Short Form (IRMA-SF; Lonsway & Fitzgerald, 1994) and the Subtle Rape Myth Acceptance Scale (McMahon & Farmer, 2011). Because of concerns about the language in the IRMA-SF, McMahon and Farmer (2011) developed a slightly revised version of the IRMA-SF, which eliminates potentially biased or strong language (e.g., caught having an illicit affair) and replaces with more common vernacular (e.g., caught cheating). The RMA measure in this study used a combination of wording from both the IRMA-SF and the Subtle Rape Myth Acceptance Scale (see the appendix for full RMA scale). In the present study, scores on the RMA scale ranged from 28 to 114 (25 items, Cronbach’s α = .85, M = 61.48, SD = 17.8).
Relative suspect blame
Attribution of blame measures the judgments made in a particular scenario. Rather than a measure of generalized beliefs (e.g., RMA), relative suspect blame measures the judgment of suspect responsibility relative to victim responsibility in a particular scenario. Participants completed the Attribution of Blame Scale (Bieneck & Krahé, 2011), consisting of eight items that evaluate aspects of victim and suspect responsibility and blameworthiness. These items were assessed using a 7-point Likert-type scale from 1 (not at all) to 7 (very much). The Attribution of Blame scale (see the appendix for full scale) is divided into two subscales: Suspect Blame (four items, Cronbach’s α = .84) and Victim Blame (four items, Cronbach’s α = .73). An example of an item measuring victim blame includes the following: “Do you think the victim is to blame for the incident?” An example of an item measuring suspect blame includes the following: “Do you think the suspect should be held criminally liable for a sexual assault?” From these two scales, a measure of relative suspect blame was created by subtracting a participant’s victim blame score from the participant’s suspect blame score, which indicates the extent to which the participant blames the suspect more than the victim. Relative suspect blame ranged from 13 to 24 (M = 4.20, SD = 8.1).
Peer perceptions
This measure assessed participants’ expectations of an important referent group—other police officers. Participants responded to two items, “Most police officers I know would classify as a sexual assault” and “Most police officers I know would consider this a legitimate sexual assault.” Both items were assessed on a 7-point Likert-type scale from 1 (definitely not) to 7 (definitely will) and formed a scale ranging from 4 to 14 (two items, Cronbach’s α = .70, M = 10.64, SD = 2.3).
Control variables
Social desirability
The Marlowe–Crowne Social Desirability Scale–Short Form C (M-C Form C) was included as a control variable (Reynolds, 1982). The Marlowe–Crowne Scale is often used as an adjunct measure to assess the impact of social desirability on self-report measures. The M-C Form C consists of 13 items that are answered by selecting either “true” or “false.” An example is “No matter who I’m talking to, I’m always a good listener”; indicating “True” on this response would score as higher social desirability bias. Another example is “I sometimes feel resentful when I don’t get my way”; indicating “False” would score as higher social desirability bias. Necessary items were reversed and computed to form the scale which ranged from 13 to 24 (13 items, Cronbach’s α = .70, M = 19.49, SD = 2.5).
Police officer demographic and experience characteristics
Participants responded to items about general education level, years in law enforcement, and receiving training on sexual assault. Participants were asked whether they had training specifically on sexual assault (1 = yes), with 75.5% responding “yes.” Participants were also asked about their level of education, which was dichotomized as 1 (at least college graduate) or 0 (less than college graduate), with 73.6% responding with at least a college degree. Total years in law enforcement ranged from 0 to 35 (M = 17.53, SD = 6.1). Gender was dichotomized (1 = male) for analysis.
Case characteristics: Victim alcohol use and prior relationship
The variables victim alcohol use and prior relationship were dummy coded into two variables, victim alcohol use (1 = yes) and prior relationship (1 = yes).
Data Analysis
Analysis consisted of multivariate regression analysis on all predictor, control, and dependent variables. Regression analyses used hierarchical regression to test for mediation while controlling for other predictor variables. Cases with missing data on dependent variables were excluded from the final models. The mediated relation will be supported if regression analyses reveal that (a) the direct path from RMA to the dependent variables is significant, (b) the path from RMA to the proposed mediating variable, attributions of suspect blame, is significant, and (c) the path from the proposed mediating variable of suspect blame to the dependent variables is significant after the influence of RMA has been controlled.
Results
Predicting Perceptions of a “Good” Case
Table 1 evaluates the degree to which officer’s RMA is associated with officer’s perception of a “good” case, when controlling for officer characteristics, including gender, education level, years in law enforcement, training, and social desirability bias; and case characteristics, including victim alcohol use and prior relationship with the suspect. In Model 1, RMA is significant and negatively associated with perceptions of a “good” case, indicating that higher RMA is associated with lower perceptions of a “good” case. In Model 2, RMA remains significantly negatively associated with perceptions of a “good” case when controlling for individual characteristics and case characteristic variables. The third model adds relative suspect blame and the perception of peers. In Model 3, RMA is no longer significant, suggesting that the effect of RMA on officer perceptions of a “good” case is mediated by attributions of suspect blame. Interestingly, the extent to which officers believe that peers will also perceive the case as legitimate remains as a significant predictor of one’s own perceptions of the case as legitimate. Results suggest that, regardless of individual police characteristics or general attitudes (i.e. RMA), attributions of relative suspect blame within a case is the strongest predictor of perceptions, indicating that higher levels of suspect blame influences higher perceptions of a “good” case.
Regression of Perception of a “Good” Case.
Note. Reference categories are in parentheses. Intercepts are estimated but not shown. [1] = zero-order model; [2] = includes controls; [3] = full model.
p < .05. **p < .01. ***p < .001.
Predicting Behavioral Intentions
Table 2 evaluates the degree to which an officer’s RMA is associated with behavioral intentions to call out a detective immediately and arrest the suspect, if identified, both of which indicate perceived seriousness. In Model 1, RMA is significant and negatively associated with behavioral intentions indicative of perceived seriousness, indicating that higher RMA is associated with lower intentions to respond in ways indicative of case seriousness. In Model 2, RMA remains significant when controlling for demographic and case characteristic variables. Interesting, scores on the social desirability scale show a significant effect on this behavioral intention, indicating that higher social desirability is associated with higher intentions to respond with more immediacy. The third model adds relative suspect blame and the perception of peers. In Model 3, RMA is no longer significant, suggesting that the effect of RMA on officer intentions is mediated by attributions of relative suspect blame. Interestingly, the extent to which officers believe that peers will also perceive the case as legitimate remains as a significant predictor, as well as victim alcohol use, with victim alcohol use decreasing the intention to call out a detective immediately or arrest the suspect, if identified. Results suggest that, regardless of individual police characteristics or general attitudes (i.e. RMA), attributions of suspect blame within a case is the strongest predictor of behavioral intentions, with higher attributions of suspect blame influencing higher scores on intentions to call a detective or arrest the suspect, if identified.
Regression of Behavioral Intention (Case Seriousness).
Note. Reference categories are in parentheses. Intercepts are estimated but not shown. [1] = zero-order model; [2] = includes controls; [3] = full model.
p < .05. **p < .01. ***p < .001.
Discussion
Based on previous research, several hypotheses tested the relationship among a variety of characteristics, individual and case-specific, on police officer’s perceptions and behavioral intentions in a hypothetical sexual assault case. The results help to explain the role of RMA, or general attitudes about sexual assault, in officer perceptions of a “good” case and behavioral intentions that indicate perceived seriousness. Higher perceptions of a “good” case was initially predicted by lower RMA scores; however, when adding case-specific attributions of blame, general attitudes no longer have a significant effect, indicating a mediation effect. Behavioral intentions to call out a detective immediately, or arrest the suspect, is also initially predicted by lower RMA, but is mediated by attributions of suspect blame. In addition, victim alcohol use and the perception of peers remain significant predictors of behavioral intentions, when controlling for other variables.
In the current study, officers showed varying levels of RMA, but the effect was mediated through attributions of suspect blame, or the responsibility assigned to the suspect relative to the victim in a hypothetical scenario. This suggests that RMA (or general attitudes and beliefs about sexual assault) functions through the ways in which police officers attribute blame to the parties involved. Attitudes, such as RMA, help to explain perceptions and behavioral intentions, but a variety of other factors—some known and some unknown—have the strongest effect on attributions of blame, which in turn predicts officer perceptions and intentions. Edward and Macleod (1999) explained how attributions of responsibility are made on the basis of perceptions regarding the degree to which the victim caused the event, had control over those causes, should be held responsible for the event, and should be blamed or not.
It is no coincidence, then, that victim alcohol use remains a significant predictor of behavioral intentions; that is, officers are less likely to call out a detective immediately or arrest the suspect when the victim-survivor had been drinking, even when controlling for other variables. This result is consistent with previous research that shows negative evaluations, including victim blame and diminished perceptions of credibility, as more likely when the victim was intoxicated at the time of the assault (Bieneck & Krahé, 2011; Schuller & Stewart, 2000; Spears & Spohn, 1996). Research demonstrates that alcohol use by the victim also significantly deters case progression (Campbell, 1998; Chandler & Torney, 1981; Frohmann, 1997; Kerstetter, 1990; Spears & Spohn, 1996). The current study suggests a link between these attitudes and case outcomes; perceptions or attitudes of first-responding officers influence their behavioral intentions, which may have some effect on case progression or the lack thereof. Officers may negatively influence case progression when deciding not to involve detectives immediately when a case involves victim alcohol use. Further research should gain clarity on the places within case processing in which victim alcohol use influences behavior or decision making. In other words, to what extent does the police interview questioning about alcohol use influence victim-survivors to drop the case? To what extent does attribution of victim blame related to alcohol use get written into the police report? To what extent does this influence prosecutors, judges, and juries?
Research explains that attributions of blame may affect someone’s emotional reaction and willingness to help the victim. A study by Clarke and Lawson (2009) found that higher attributions of fault toward the victim were associated with greater negative affective reactions of anger, disgust, and decreased feelings of sympathy. These reactions included a decreased desire to help the victim. Alternatively, attributions of fault toward the perpetrator were associated with greater feelings of sympathy for the victim, lower expressions of negative affect, and an increased desire to help the victim (Clarke & Lawson, 2009). The current study’s findings along with previous research suggest that training and interventions for those likely to come into contact with victim-survivors should focus on addressing victim-blaming myths and emphasize that sexual assault is a crime for which the offender is to blame (Clarke & Lawson, 2009). The study’s results showed a consistent effect of attributions of suspect blame on perceptions of a “good” case and behavioral intentions. Training on the definition of sexual assault could lead to more inferences of responsibility and attributions of blame toward the suspect, even when the report or sexual assault scenario includes elements of what officers might have deemed as “red flags,” such as “improper” or risk-taking behavior on the part of the victim leading up to the sexual assault. Training that specifically addresses the role of alcohol in sexual assault should be included (Abbey, 2015).
Prior relationship between the victim-survivor and suspect did not show a significant effect on perceptions and behavioral intentions, even though studies show less case progression if the victim and alleged perpetrator had some relationship, whether being married, having prior sexual relationship, or even being acquaintances (Buzawa et al., 1995; Chandler & Torney, 1981; Kerstetter, 1990; LaFree, 1980). In the current study, vignettes either included an ex-boyfriend or an acquaintance, neither of which are part of the stereotypical “real rape” image that involves a stranger (Estrich, 1987; Venema, 2016). Because the alleged suspect in the vignette was always somewhat known, this may explain the lack of effect. In addition, because the victim willingly went to the alleged perpetrator’s apartment, this could be interpreted as blameworthy, even if the alleged suspect was not well-known.
The extent to which participants believed other officers would also consider the sexual assault legitimate and the victim credible influenced officer’s own perceptions and behavioral intentions. This is consistent with Frohmann’s (1997) “downstream orientation,” which argues that perceptions are often based on predictions about other party’s reactions to the case (p. 535). The current study extends the application of this “downstream orientation” to a more immediate referent group, other police officers, rather than going further downstream to prosecutors, for instance (Spohn et al., 2014). Thinking “downstream” and imagining that others will perceive the case similarly influences behavioral intentions of how to respond. When a case is perceived as legitimate by those “downstream,” this increases one’s own perception of legitimacy. However, if a case is perceived as weak, officers look “downstream” to how that will be interpreted by others (e.g., prosecutor), and are strengthened in their behavioral intention not to respond with immediacy. This suggests that police officers may be just as concerned about perceiving and reacting to cases in ways similar to peers within the police department, rather than using their own RMA or prior experience to guide decision making. This points to the need for regular training and a peer culture shift regarding sexual assault, as norms within the police department may determine response, more than individual attitudes or characteristics.
There is evidence of the exercise of discretion or differential responses by officer’s perceived seriousness of the report. When asked about support for calling out a detective immediately or arresting the suspect, if identified, there was much more variation in this response, with some officers expressing very little support and others expressing very much support. Because detectives work regular day shifts with on-call availability after hours in this department, officers exercise discretion in determining whether a case deserves immediate attention by a detective. It is possible that officers only call out detectives (who are on-call after their regular shift) when the report includes elements of “real rape.” Because officers may take on a strict crime fighter orientation, this may lead them to focus on more “serious, less ambiguous, criminal incidents (i.e. felonies)” rather than more ambiguous reported sexual assault cases (Paoline, 2003, p. 203). This may also be the case when the identity of the perpetrator is unknown and there is a need for an immediate search.
One specific suggestion for police practice focuses on the intentional selection of police officers to respond to sexual assault reports. This recommendation has also been suggested by Rich and Seffrin (2012), and is similar to the Crisis Intervention Team (CIT) model for police officers, which focuses on appropriate responses to persons with a mental illness (Compton, Bahora, Watson, & Oliva, 2005). Because the number of sexual assaults that are actually reported to the police is relatively small in proportion to the prevalence of the crime, it seems feasible to train a subset of officers to be trained as a Sexual Assault Police Response Team. Although by some measures, attitudes toward sexual assault may be changing among police officers, open-ended responses still indicate the acceptance of rape myths, which negatively affect victims and the progression of cases. Mennicke and colleagues (2014) argued that despite some evidence of a genuine shift in policing culture, research that uses open-ended questions about sexual assault still demonstrate the endorsement of rape myths. Rather than emphasizing the change in social norms (which are necessary both within police culture and the broader public), a more useful strategy is to train a smaller group of police officers who already show lower levels of RMA and are more sensitive to sexual assault victim-survivors. Instead of using the entire department as a referent group for decision making, officers in this specially-trained unit could look to the attitudes and behaviors of peers who are similarly trained in an understanding of sexual assault. A model of officer self-selection, similar to CIT, would potentially produce the most positive outcomes in the victim interview and subsequent case progression.
Training is important because RMA is malleable and strategically motivated and not intrinsic in nature (Chapleau & Oswald, 2013). Training should focus on both attitude and behavior. Although attitudes do not necessarily determine behavior, they are “proximal determinants of behavior and constitute the evaluative information that is necessary to initiate action (e.g., Ajzen, 1985; Fishbein & Ajzen, 1975)” (Elliott, Brewster, Thomson, Malcolm, & Rasmussen, 2014, p. 669). Rich and Seffrin (2012) provide a helpful discussion and review of literature on what training is most likely to succeed. They report that there are contradictory findings on whether attitudinal change is necessary to improve interviewing skills, or whether behavioral changes (i.e. interviewing skills improvement) may actually lead to attitudinal change (Jaccard & Blanton, 2005; Lonsway, Welch, & Fitzgerald, 2001). Although limited in scope, Currier and Carlson’s (2009) study found that explicit education with college students on violence against women had an effect in changing negative attitudes. Training should use real-case examples to demonstrate elements of an assault that make it a crime (e.g., lack of consent) or should be ignored (e.g., victim attire) rather than talking about sexual assault in the abstract.
Study Limitations
Despite the study’s contributions, the results are limited to one jurisdiction, which influences the generalizability of the results to other locations. Second, the sample size is relatively small, although it should be noted that it does represent a majority of officers within this one jurisdiction. Third, the use of a hypothetical vignette poses some limitations, even though some argue that vignettes, through contextualization and standardization, provide a more valid and reliable measure of opinions than simpler and more abstract questionnaires. In contrast, vignettes are only a snapshot of reality, and so they can never convey the real complexity in life, and they do not lend to results that demonstrate the impact of perceptions on actual decision making and behavior. Related, the measurement of behavioral intention is only a proxy for actual behavior. The extent to which attitudes really influence behavior is not tested within this study; future research should begin to link attitudes with actual decision making in real sexual assault cases.
Future Directions
Overall, this study reveals the important role of attributions of blame in police officer perceptions and behavioral intentions when responding to reported sexual assaults. Although much research has been done to understand decision making and case attrition, there is still a gap in understanding the ways in which individual attitudes function to influence perceptions of cases, which are likely to affect decisions and behaviors of police officers. Future research should begin to connect not only case characteristics but also individual police officer characteristics, attitudes, and attributions of blame, with real-case outcomes, instead of hypothetical scenarios. Tracking cases from initial report through the case progression, including a content analysis of police reports, would provide better insight into the role of various actors—particularly the first-responding police officer—in sexual assault case progression. Future research should evaluate the implementation of a specially trained Police Sexual Assault Response team to analyze changes in attitudes, knowledge, behavior, as well as perceptions of the reporting victim-survivor about sensitivity and thoroughness in the victim interview and overall responsiveness to the case.
It is important to continue uncovering the factors that influence such negative experiences of victim-survivors when reporting as well as high rates of case attrition. The extent to which first-responding officers play a role in these uniquely negative characteristics of sexual assault reporting and case progression needs to be studied further to better understand the criminal justice system response and provide better training or procedures that will improve that response.
Footnotes
Appendix
| RMA Scale |
| If a woman is raped while she is drunk, she is at least somewhat responsible for letting things get out of control. |
| Women who are caught cheating sometimes claim that it was rape. |
| A lot of times, women who say they were raped agreed to have sex and then regret it. |
| If a woman doesn’t physically fight back, you can’t really say it was rape. |
| When women go around wearing low-cut tops or short skirts, they are just asking for trouble. |
| Rape accusations are often used as a way of getting back at men. |
| A rape probably didn’t happen if the woman has no bruises or marks. |
| If a woman goes home with a man she doesn’t know, it is her own fault if she is raped. |
| If the accused rapist doesn’t have a weapon, you really can’t call it a rape. |
| When a woman is a sexual tease, eventually she is going to get into trouble. |
| A lot of times, women who say they were raped often led the man on and then had regrets. |
| A lot of times, women who claim they were raped just have emotional problems. |
| If a woman doesn’t physically resist sex—even if protesting verbally—it can’t be considered rape. |
| If a woman initiates kissing or hooking up, she should not be surprised if a guy assumes she wants to have sex. |
| When women are raped, it’s often because the way they said “no” was unclear. |
| A woman who dresses in skimpy clothes should not be surprised if a man tries to force her to have sex. |
| A woman who goes to the home or apartment of a man on the first date is implying that she wants to have sex. |
| If a woman claims to have been raped but has no bruises or scrapes, she probably shouldn’t be taken too seriously. |
| If a woman doesn’t say “no” she can’t claim rape. |
| When men rape, it is usually because of their strong desire for sex. |
| Men don’t usually intend to force sex on a girl, but sometimes they get too sexually carried away. |
| Rape happens when a man’s sex drive gets out of control. |
| If a man is drunk, he might rape someone unintentionally. |
| It shouldn’t be considered rape if a man is drunk and didn’t realize what he was doing. |
| If both people are drunk, it can’t be rape. |
| Attribution of Blame Scale |
| Do you think the suspect is responsible for the incident? |
| Do you think the suspect should be held criminally liable for rape? |
| How certain are you that the incident meets the legal definition of rape? |
| If you were a member of a jury, how certain are you that you would convict the suspect of rape? |
| Do you think the victim is to blame for the incident? |
| Do you think the victim could have avoided the incident? |
| Do you think the victim had control over the situation? |
| Do you feel sorry for the victim? |
Note. RMA = rape myth acceptance.
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 work was supported by the Alice Dan Dissertation Award, University of Illinois at Chicago.
