Addressing Violence Against Women: The Effect of Women’s Police Stations on Police Legitimacy

Abstract

With a focus on the implementation of women’s police stations (WPS), we posit that local policies that address violence against women can result in positive feedback effects on institutional legitimacy. We theorize that WPS increase police legitimacy among women by improving perceptions of personal safety and government responsiveness. To test our hypotheses, we rely on municipal and public opinion data from more than 100 municipalities in Brazil. The results of our multilevel analysis indicate that WPS produce positive feedback effects among women, resulting in higher trust in the police among women than men and closing the gender gap in perceptions of police effectiveness. Incorporating an instrumental variable in the analyses yields similar results, suggesting that these effects are not endogenous. Moreover, the results of our mediation models show that WPS’ positive effects on women’s views of police legitimacy are driven by improved perceptions of personal safety, and not perceptions of government responsiveness.

Keywords

institutional legitimacy gender equality policies policy feedback effect public opinion attitudes toward the police

Violence against women (VAW) is a prevalent problem across the world that has profound negative implications for women’s lives. VAW not only results in exorbitant physical and emotional costs for victims, but it is also detrimental for the advancement of women as a group. VAW is associated with a myriad of health problems for victims, including depression and posttraumatic stress disorder (e.g., Jordan, Campbell, & Follingstad, 2010), as well as generalized fear of crime among victims and non-victims (e.g., Ferraro, 1996). Among the many forms of violence targeted at women, intimate partner violence is one of the most common, alarmingly affecting at least one in three women in the world (United Nations, 2015). Yet, most incidents of VAW go underreported and consequently unpunished, which makes it difficult to deter further violence. Across the globe, less than 40% of female survivors seek help of any kind and, among those who do, less than 10% report their experiences to the police (United Nations, 2015). Low confidence in the police and negative views of police performance largely influence women’s decision to remain silent (e.g., Calton & Cattaneo, 2014).

To improve VAW services and increase reporting rates, several countries across the world have mandated the establishment of women’s police stations (WPS). WPS are specialized police units staffed primarily with female police officers designed to prevent, investigate, and punish VAW. Since Brazil inaugurated the first WPS in the world more than two decades ago, other countries have followed (Jubb et al., 2010), including at least eight countries in Latin America¹ and eight others outside the region.²

Drawing from literatures in the fields of public opinion, gender and politics, and criminology, we posit that the existence of a WPS in the place of residence improves women’s perceptions of police legitimacy. As such, this study addresses a largely understudied topic. Although extant research has theorized that the implementation of gender equality policies influences citizen views of political institutions (e.g., Barnes & Córdova, 2016; Kittilson, 2010; Schwindt-Bayer & Mishler, 2005), we still know little about the effects of policies specifically designed to deter VAW on public opinion. Besides examining whether WPS shape citizen opinions of police legitimacy, we contribute to the literature by identifying some of the mechanisms through which this effect might occur.

The bulk of the literature suggests that the adoption of gender equality policies results in positive evaluations of institutional legitimacy among women through its effect on perceptions of government responsiveness (e.g., Lawless, 2004; Schwindt-Bayer & Mishler, 2005). We complement this literature by theorizing that policies designed to prevent VAW, such as WPS, are also likely to generate positive attitudes toward political institutions by improving perceptions of personal safety. Public services that seek to protect victims and prevent VAW can lower fear of crime among the broader female population. In short, with a focus on WPS, we theorize that feedback effects resulting from VAW policies at the local level have the potential to increase police legitimacy through improvements in perceptions of personal safety, in addition to government responsiveness.

Our empirical analyses rely on public opinion and municipal-level data from more than 100 municipalities in Brazil. First, to evaluate the salience of WPS at the municipal level, we use data from a 2013 survey on VAW carried out by Data Popular and the Instituto Patricia Galvão. Second, we test our hypotheses using data from the 2014 Latin American Public Opinion Project (LAPOP) survey. The municipal-level data are from various sources, including the Brazilian Institute of Geography and Statistics (IBGE). Our multilevel analysis indicates that women who reside in municipalities with a WPS hold more positive views of police legitimacy, resulting in higher trust in the police among women than men and closing the gender gap in perceptions of police effectiveness. By contrast, although men are also aware of the existence of a WPS in their municipality, they do not turn to heuristic information drawn from WPS to form their opinions on police legitimacy.

The findings remain robust after we include an instrumental variable in our multilevel models, suggesting that the presence of a WPS at the municipal level causes observed effects on police legitimacy. Additional results from path analyses show that these effects are largely driven by improved perceptions of safety among women living in municipalities with a WPS. In municipalities with a WPS, men and women display similar rates of fear of crime, challenging the conventional wisdom that women are more fearful of crime than men. Moreover, contrary to what the literature suggests, the implementation of WPS does not affect police legitimacy through its effect on perceptions of government responsiveness.

Taken together, our findings carry with them an important policy implication. The results suggest that, by improving perceptions of police legitimacy among women, WPS can contribute to increasing women’s likelihood of reporting gender-based violence to the police.

Policy Feedback Effects and Gender Equality Policies

Extant literature has demonstrated that the adoption of public policy influences citizen attitudes toward political institutions (e.g., Campbell, 2012; Mettler, 2015; Mettler & Soss, 2004). According to this perspective, policies carry “informational content” that serves as cues to citizens to form their political opinions (Pierson, 1993, p. 619). As Zaller (1992) suggests, when asked about their opinion on an issue, citizens turn to policies that are “immediately salient to them” to construct an opinion (p. 1). The adoption of public policy per se generates heuristic information that aids citizens to emit a political view. Consequently, policy feedback effects on public opinion can take place even if individuals do not have complete information about the substantive impact of a given policy (e.g., Lü, 2014).

The adoption of gender equality policies then has the potential to shape men’s and women’s political views (Kittilson, 2010; McDonagh, 2010), independent of their factual knowledge on how a given policy operates in practice. For instance, Kittilson (2010) theorizes that, “the adoption and expansion of national policies on issues such as equal wages, child care provision, paid maternity leave, and VAW carry important messages to the electorate” (p. 217). Accordingly, scholars have focused their attention on the identification of mechanisms through which gender equality policies shape citizens’ political views, relying on Pitkin’s (1967) conceptualization and theorizing about the interconnectedness between substantive and symbolic representation.

Substantive representation is commonly understood as the adoption of laws or public policies that respond to citizen interests. Pitkin (1967) conceptualized the attitudinal responses to the representation process as symbolic representation. Scholars have theorized that substantive representation triggers a myriad of feelings and attitudes among the public, including perceptions of institutional legitimacy. For example, Schwindt-Bayer and Mishler (2005) posit that the adoption of gender equality policies increases the legitimacy of the legislature among women. The authors suggest that women’s perceptions of government responsiveness explain this effect. When political elites respond to women’s interests, women are more likely to perceive that they are “fairly and effectively represented” (Schwindt-Bayer & Mishler, 2005, p. 407).

The empirical literature examining the link between gender equality policies and institutional legitimacy, however, has yielded mixed findings. Results vary depending on the context and policies in consideration. For example, Schwindt-Bayer and Mishler (2005), using data for 31 democracies and a composite index of four gender equality policies, find no relationship between policy adoption and trust in the legislature. By contrast, Schwindt-Bayer (2010) concludes that, in the context of Latin America, the passage of gender equality policies, measured by the percentage of laws focused on women’s issues, leads to higher trust in government and the legislature among women. Empirical studies examining the relationship between the adoption of gender quota laws and legitimacy also remain inconclusive. Schwindt-Bayer (2010) finds no effect of gender quotas on trust in government and the legislature. Similarly, Zetterberg’s (2009) findings for Latin America show “no relationship whatsoever between quotas and a female citizen’s trust in politicians and political parties” (pp. 721-723).

In our view, the mixed findings are largely the result of variations in the salience and traceability of the gender policies investigated. Policies that are visible and traceable to government action are more likely to generate policy feedback effects (Campbell, 2012; Pierson, 1993), and consequently engender symbolic representation. We argue that gender policies implemented at the local level are particularly likely to be visible and traceable, and consequently more likely to be used by women as heuristic information to emit an opinion. Women are more likely to be aware of government efforts to provide local services for women, such as health centers and WPS, than about the provision of similar services in more distant and centralized locations. Moreover, gender policies implemented at the local level can facilitate women’s access to services, which can further influence their views of institutional legitimacy.

The extent to which gender equality policies shape views of political institutions also depends on the nature of benefits offered to women. Gender policies designed to benefit all women are more likely to produce generalized feedback effects. For example, daycare provision and equal wage laws have the potential to affect political legitimacy mainly among women with children and working women, respectively. By contrast, given the generalized nature of gendered violence, VAW policies are likely to influence most women’s opinions about political institutions. Indeed, women in general are at risk of gender-based violence, independent of their socioeconomic background and demographic characteristics (e.g., Jordan et al., 2010).

In this article, we examine the policy feedback effects of a visible and traceable VAW policy implemented at the local level that has the potential to benefit most women: WPS. Although the number of WPS and the quality of services they provide vary within and across countries that have implemented them, WPS share design features that make them particularly likely to generate feedback effects on police legitimacy. First, their localized nature increases access to and awareness of WPS’ services in the towns where they operate. Second, WPS require the creation of physical spaces for police officers and other staff to work, which increases WPS’ visibility at the local level. WPS’ offices typically have visible signs outside of their facilities with their proper name (e.g., Delegacia de Defensa da Mulher de Itapevi, São Paulo, Brazil). Third, WPS’ services are unambiguously traceable to government action. They are part of the police, a state institution that receives directives from government.

Finally, WPS share a unique design feature that is likely to reinforce policy feedback effects among women. By design, WPS are predominantly led by female police officers. Hence, WPS increase the “compositional similarity between representatives and the represented” (Schwindt-Bayer & Mishler, 2005, p. 407), namely descriptive representation (Pitkin, 1967). Previous studies suggest that, when a higher percentage of women are involved in the implementation of policies that affect women, citizens are more likely to perceive fairness in the decision-making process (Clayton, O’Brien, & Piscopo, 2019). Consequently, when a policy increases access to services that benefit all women and women are mainly responsible for service provision, policy feedback effects are likely to be maximized. In short, the establishment of WPS is likely to generate significant feedback effects on police legitimacy in part because it is a policy that offers services “by and for women” (Hautzinger, 2016).

Police Legitimacy and WPS

Given the central role of the police as an institution in charge of enforcing the rule of law in democratic societies, the determinants and consequences of police legitimacy have been the focus of extensive research, primarily in the field of criminology. Police legitimacy is a multidimensional concept that has been conventionally defined as “a quality possessed by [the police] that leads others to feel obligated to obey its decisions and directives voluntarily” (Tyler & Huo, 2002, p. 102). This definition, however, has been recently contested on the grounds that compliance with the police is an outcome and not a property of police legitimacy (Tankebe, 2013). Although there is no consensus on its formal definition, scholars have identified trust in the police and perceptions of police effectiveness as two important dimensions of police legitimacy (e.g., Johnson, Maguire, & Kuhns, 2014; Tankebe, 2013). With a focus on these two dimensions, we seek to contribute to the mass body of literature in criminology and political science that examines how institutional legitimacy can be promoted, and with it improve compliance with the rule of law in democratic societies. Indeed, more generally, political scientists identify trust and perceptions of effectiveness as key components of institutional legitimacy (e.g., Booth & Seligson, 2009; Norris, 1999; Weatherford, 1992).

Trust in the police measures diffuse support (Easton, 1975; Sharp & Johnson, 2009), or a deep-rooted attitude toward the police grounded on the belief that it is a proper institution for society. By contrast, perception of police effectiveness measures specific support, namely perceptions of performance at a given point in time. Both dimensions are strongly correlated with cooperation with the police (e.g., Reisig, Tankebe, & Mesko, 2012; Tankebe, 2009), including reporting crime. Given the relevance of police legitimacy, a large body of literature has examined its determinants, including the effect of police policy. Studies, for example, have explored the effect of community policing on trust in the police (Sharp & Johnson, 2009) and attitudinal effects of policies to reduce police corruption (Barnes, Beaulieu, & Saxton, 2018). Yet, the potential feedback effects of VAW policies implemented by the police have received little attention.

In general, police legitimacy has not been extensively analyzed from a gender perspective. This might be the case because empirical studies have consistently found that, on average, men’s and women’s views of police legitimacy are similar. Previous research in the context of the United States (e.g., Sharp & Johnson, 2009), Africa (e.g., Boateng, 2017), and Latin America (e.g., Corbacho, Pilipp, & Ruiz-Vega, 2012) have all found gender to be an insignificant predictor of police legitimacy. In this article, we posit that this average result masks important gendered effects at the subnational level depending on the implementation of localized police policies concerning women’s safety. More specifically, we argue that substantive representation in the form of specialized police stations to prevent and punish VAW generates policy feedback effects on women’s views of police legitimacy—a manifestation of symbolic representation. As explained above, the design features of WPS make this policy particularly likely to shape views of police legitimacy. In addition, we posit that this effect occurs mainly among women.

The literature indicates that direct policy beneficiaries are not only more likely to resort to heuristic policy information to form their political opinions but also more likely to express positive views toward institutions in charge of policy implementation (Campbell, 2012). This is because policies are more salient among those reaping their benefits, who in turn tend to express more favorable opinions toward policy implementers. Consequently, when asked to voice opinions about the trustworthiness and effectiveness of the police, women are more likely than men to resort to heuristic information drawn from WPS through a “top of the head” process (Zaller, 1992). Moreover, as direct beneficiaries, women living in municipalities with a WPS should be more inclined to express positive views toward the police than men. More specifically, we derive the following hypothesis:

Hypothesis 1 (H1): In municipalities with a WPS, women will express more positive evaluations of police legitimacy than men and their female counterparts in municipalities without a WPS.

Notice that, in this article, we are interested in determining the overall gendered effects of WPS on police legitimacy. However, we do not rule out the possibility that WPS produce heterogeneous feedback effects across female subpopulations. For example, previous studies indicate that members of minority groups tend to report negative experiences with the police (e.g., Peffley & Hurwitz, 2010), which suggests that the positive feedback effects derived from WPS might be less pronounced among women of color. Moreover, given that the quality of WPS services is typically far from ideal, survivors of gendered violence who have turned to WPS for help might hold less optimistic opinions of police legitimacy compared with women who have not used WPS services. However, since only a small percentage of women seek help from the police to denounce gendered violence, we expect this likely effect to be counterbalanced by the more positive views among non-users of WPS. In short, although WPS might generate heterogeneous feedback effects among women, we expect that WPS will improve views of police legitimacy among women on average. Next, we identify some of the mechanisms through which this average positive effect might occur.

As mentioned above, the gender literature identifies perceptions of government responsiveness as a likely driver of feedback effects associated with gender equality policies (e.g., Schwindt-Bayer & Mishler, 2005). Similarly, the policy feedback effect literature points to “beliefs about government responsiveness and one’s capacity to influence political outcomes” as an important underlying attitude explaining the effect of policy on public opinion (Mettler & Soss, 2004, p. 59). These insights suggest that, when citizens perceive they have benefited from policy outcomes, feelings of government responsiveness to their policy preferences ensue (Kölln, 2016). In turn, this has a positive effect on their political attitudes, including views of institutional legitimacy (Esaiasson & Wlezien, 2017). The link between perceptions of government responsiveness and legitimacy, as measured by political trust, finds wide empirical support (Catterberg & Moreno, 2006). This review of the literature suggests that the establishment of WPS can serve as heuristic information to women that the government responds to their interests, which can then lead to improved perceptions of police legitimacy. Indeed, extant research shows that perceptions of government responsiveness predict police legitimacy (Hu, Sun, & Wu, 2015). We draw the following hypothesis:

Hypothesis 2 (H2): The establishment of a WPS at the municipal level has a positive effect on women’s views of police legitimacy through perceptions of government responsiveness.

We posit, however, that an additional mechanism might be at work. WPS are likely to have a positive effect on police legitimacy by improving women’s perceptions of personal safety. Research in criminology has consistently shown that women are more fearful of being victims of all types of crime, despite their lower probability of becoming victims of non-sexual crimes than men. As Ferraro (1996) notes, this is largely explained by women’s higher risk of sexual assault and consequently higher probability that any kind of victimization involves sexual violence. Women’s fear of sexual violence drives their heightened fear of being victims of any type of crime. Empirical studies support this argument, finding, for example, that fear of rape is strongly associated with fear of burglary and robbery (e.g., Schafer, Huebner, & Bynum, 2006; Warr, 1984). As one of WPS’ objectives is to protect women against gendered violence, including sexual assault, the existence of a WPS in the place of residence has the potential to lower women’s fear of crime victimization in general.

Indeed, studies have found that police presence reduces fear of crime (e.g., Zhao, Scheider, & Thurman, 2002), suggesting that specialized police services for women can lower fear of crime among women in particular. The existence of a WPS in the municipality is likely to signal to women that they have access to police resources that can better protect them from crime and violence. Moreover, the fact that WPS are predominantly staffed by female police officers might make women perceive WPS as more effective to protect them against sexual violence and other types of crimes than regular police stations. The first WPS in Brazil was established with this presumption in mind. Policymakers and feminist groups hoped that gendering these specialized units would avoid some of the problems associated with male-dominated police stations, including the perception that VAW is a lesser crime and the re-victimization of female complainants (Hautzinger, 2007, 2016).³

In addition, women might perceive that the presence of a WPS by itself helps deter men from victimizing women, further reducing fear of crime among women. Hautzinger (2007), for example, documents that women in Bahia, Brazil, living in abusive relationships have been able to deter physical aggression by reminding male perpetrators that there is a WPS nearby. She describes the case of a woman who never set foot into the local WPS, but was able to stop her husband’s “hand in the air, preventing the slap he had intended when he raised it” (p. 137) by threatening to press charges against him at the local WPS. Hautzinger (2016) concludes that “the most powerful deterrent effect” of WPS is “the message their mere existence sends to the Brazilian population: Violent assaults against women are crimes that will not be tolerated” (p. 582). This discussion leads us to our last hypothesis:

Hypothesis 3 (H3): The establishment of a WPS at the municipal level has a positive effect on women’s views of police legitimacy through improved perceptions of safety.

Brazil as a Case Study

Brazil constitutes an ideal case study to test our hypotheses. First, although most municipalities in Brazil still lack a WPS, it is the country with the largest number of WPS in Latin America (United Nations Women, 2011). We exploit this variation and examine whether women in municipalities with a WPS have different views of the police compared with municipalities without one. Second, Brazil constitutes what in political methodology is termed a “critical case” (Eckstein, 1975) or a case where at first glance the probability of falsifying established hypotheses is likely to be high. Despite government efforts to lower VAW, Brazil remains one of the countries with the highest rates of VAW in the world (Waiselfisz, 2015), suggesting that WPS are unlikely to improve women’s views of police legitimacy. Consequently, if our empirical results support our hypotheses, it would provide strong evidence in favor of our theoretical insights. In this section, we provide background information on the establishment and functioning of WPS in Brazil and evidence supporting one of the core assumptions of our study. That is, WPS are highly visible at the local level and consequently likely to be a source of heuristic information for women when deciding how much legitimacy to attribute to the police.

WPS in Brazil

Feminist movements in Brazil have historically played an important role in pressing the government to address VAW in the country.⁴ Anti-violence feminist movements emerged in the 1970s from widespread protests that broke out in response to a series of lenient verdicts in which men who had murdered their wives were absolved on the grounds of “defense of honor” (Hautzinger, 2007). During the democratization process in the 1980s, feminist movements grew stronger and their activism focused on the problem of domestic violence (Htun, 2002). They pushed not only for the criminalization of domestic violence but also for public policies to prevent an issue that society had historically deemed as “private” (Santos, 2010).

In the early 1980s, feminist grassroots groups lobbied the state government of São Paulo to address VAW. In response, the state government proposed the creation of WPS (Delegacias Especializadas no Atendimento à Mulher), which would be staffed mainly by female police officers. Since the establishment of the first WPS in the city of São Paulo in 1985, the number of WPS has grown considerably across the country, reaching about 357 by 2016 (DataSenado, 2016). The enactment of the Maria da Penha law in 2006, which addresses intimate partner violence, further pushed for the expansion and standardization of WPS services across the country (Perova & Reynolds, 2017). The law marked a pivotal moment for the legal protection of Brazilian women by establishing a mandate for the federal, state, and municipal governments to collaborate on creating WPS (Gattegno, Wilkins, & Evans, 2016).

Although the government has not established specific criteria for WPS eligibility, its guidelines for WPS operation promote their enactment particularly in large municipalities, without specifying a population threshold (Secretaria de Políticas para as Mulheres, 2010). As a result, although there is wide variation in the population size of municipalities where WPS have been inaugurated, on average, municipalities with a WPS have a larger population size than those without one (Instituto Brasileiro de Geografia e Estatística [IBGE], 2014). This trend has resulted in the concentration of WPS in municipalities with higher development levels, at least on average. Despite the 2006 law’s mandate, WPS have only been established in about 6.4% of the 5,570 municipalities in the country (DataSenado, 2016). There is only one WPS per 15 municipalities. Yet, WPS represent the main institutional channel to address VAW in Brazil (Santos, 2010).

Survivors of VAW turn to WPS to press charges against their aggressors, file a complaint, and seek access to services, such as psychological attention and women’s shelters (Hautzinger, 2007). The majority of complaints tend to involve intimate partner violence, but WPS also investigate other gendered crimes. However, most WPS are understaffed, operate under limited hours, and officers rarely receive gender sensitization training (e.g., Perova & Reynolds, 2017). Despite these challenges, according to data from a 2013 survey conducted by Data Popular and the Instituto Patricia Galvão, 83% and 84% of men and women, respectively, identify WPS as the most effective government program to address domestic violence.

We argue that, independent of WPS’ service quality, the existence of a WPS by itself is likely to produce positive policy feedback effects on police legitimacy among women. Women residing in municipalities with a WPS are particularly likely to be aware of WPS services and use them as heuristic information to form opinions on the police. The 2013 survey, which was carried out across 99 municipalities in Brazil, allows us to examine how visible WPS are at the municipal level. The survey does not include a general question on knowledge about the existence of a WPS, but asks respondents a more specific question, whether they know the location of a WPS in their municipality. Our analysis of these data revealed that, over 50% of men and women in municipalities with a WPS know its location. These results indicate that WPS are highly visible among men and women in municipalities where they have been implemented. The policy feedback literature suggests, however, that men will be less likely than women to use “top of the head” information derived from WPS when asked to emit an opinion on police legitimacy. Since men are not the primary beneficiaries of the policy, other factors are expected to be more important for informing men’s opinions on the police (e.g., police corruption) than policies designed to benefit women.

Besides being highly visible at the local level, WPS are directly traceable to the police, which makes WPS likely to shape women’s views of police legitimacy. Policing in Brazil is divided between two agencies, the civil and military police. Both police branches play an important role in the operation of WPS. The military police are in charge of immediate crime response and preventive patrolling, and the civil police are mainly responsible for criminal investigation (Cano, 2006). Although WPS are part of the civil police, the military police work together with WPS’ civil police officers to process cases of VAW. The military police bring female victims and/or aggressors to WPS so that the civil police can then proceed with the criminal investigation (Hautzinger, 2007). Indeed, the guidelines established by the government for the standardization of WPS activities mandate the cooperation between the two police branches (Secretaria de Políticas para as Mulheres, 2010):

The civil and military police must act within their constitutional sphere of competence and seek synergy . . . This means that, when a military police officer attends an incident, he or she must refer the woman who has been victim of violence to the WPS closest to her home or the place of the act. (p. 31)

Given the involvement of both branches of the police in WPS’ operations, we expect WPS to shape citizens’ views toward the police in general and each branch of the police in particular. However, as we explain in the next section, the survey data we employ to test our hypotheses only allow us to explore general attitudes toward the local police (without specifying the branch) and the military police more specifically. The survey does not contain questions that make explicit reference to the civil police.

Data, Measurement, and Method

To test our hypotheses, we use data from the 2014 nationally representative LAPOP survey in Brazil and municipal-level data gathered from various sources.⁵ The total sample consists of 1,500 interviews carried out across 106 municipalities randomly selected using a probability proportional to size (PPS) procedure. With PPS, larger municipalities have a higher probability of selection, but small municipalities are also represented in the sample. Since WPS have been established primarily in large municipalities, the implementation of PPS conveniently results in a sample with a sufficient number of municipalities with a WPS, despite the low percentage of WPS nationwide. In our sample, 51.8% of the 106 municipalities have a WPS, and about 58% of the total number of interviews was carried out in these municipalities.⁶

Although LAPOP has gathered public opinion data in Brazil in other years, using the 2014 survey offers an important advantage. The survey was carried out right after the IBGE had collected data on municipal resources, including the existence or not of a WPS in each municipality. The timing of the LAPOP and IBGE fieldwork allows us to code with greater precision if a given municipality included in the survey had a WPS at the time citizens were interviewed. Since the government has been expanding the WPS program in Brazil, using WPS and public opinion data that were collected in sequence help us minimize the chances that a given municipality would be coded as not having a WPS when in fact it does have one.

Dependent Variables

Our two measures of police legitimacy, trust in the police and perceptions of police effectiveness, correspond to the following questions available in the LAPOP survey, respectively: (a) On a scale from 1 to 7, where 1 means not at all and 7 A lot, to what extent do you trust the military police? (b) Are you very satisfied (4), satisfied (3), dissatisfied (2), or very dissatisfied (1) with the performance of the police in your neighborhood? Each of these two survey items allows us to examine distinct dimensions of police legitimacy.⁷ Following previous literature, we posit that the first and second items measure diffuse and specific support, respectively. As can be noted, the first item asks about the military police, and the second one about the local police more generally, without mentioning the police branch, either the military or civil police. Given the strong coordination that takes place between police branches for the operation of WPS, we expect to find significant policy feedback effects when we use this measure of specific support even if the police branch is not explicitly mentioned.

On a 1-7 scale, the mean value of trust in the military police is 3.8 with a standard deviation of 2.0 points. The majority of Brazilians express a level of trust lower or equal to 4 points. Indeed, only 37.5% reported a level equal to or higher than 5 points, and only 12% chose the highest category (7). Opinions on police effectiveness are also highly divided. The mean value for perceptions of police effectiveness is 2.3 with a standard deviation of 0.75 points on a 1-4 scale. When asked about their level of satisfaction with police performance, about 41% of the population answered “dissatisfied,” and a similar percentage responded “satisfied.” Only 3.24% of the population said they were “very satisfied” with the performance of the police.

Independent Variables

At the individual level, our main independent variable is gender, coded 1 for female and 0 for male respondents. At the municipal level, our independent variable of focus is WPS, coded 1 if a WPS has been established in a municipality or 0 otherwise. To test our expectation that WPS have a differential effect on police legitimacy between men and women (H1), we add into our models an interaction term between gender and the WPS variable. In addition, since it is possible that WPS have been established in municipalities with higher levels of gendered violence, we control for femicide rates and interact this variable with gender to account for the possibility that women’s opinions of the police are influenced by their actual risk to be murdered in their place of residence. At the municipal level, our models also control for the existence of regular police stations, which are also constituted by the civil police. In 2014, about 76.8% of municipalities in Brazil had a regular police station (IBGE, 2014). In our sample, only five municipalities did not have one. We include this variable in our models to rule out the possibility that our results are driven by outlier cases.⁸ Our models also control for municipalities’ population size and level of economic development as measured by the Gross Domestic Product (GDP) per capita.

At the individual level, our models account for personal experiences with crime and police corruption, two factors found to be negatively associated with police legitimacy in the Latin American context (Malone, 2010). Since attitudes toward the incumbent government can influence citizens’ perceptions of political institutions, including the police, we also control for a survey item on presidential approval. In addition, we control for individuals’ socio-demographic characteristics, including education, age, urban residency, household wealth,⁹ marital status, number of children, and ethnic self-identification.

Method

To test H1, we estimate two-level multilevel models that account for the nested structure of the data (e.g., Snijders & Bosker, 2012).¹⁰ We take into account clustering effects at the municipal level in the estimation of standard errors. Since the two dependent variables of interest are ordinal, we estimate ordered logistic multilevel models with random effects for the intercept at the municipal level.¹¹ In addition, the models account for fixed effects at the state level. The models that test H1 take the following form:

\begin{matrix} Y_{i j} = β_{0} ​ β_{1} G e n d e r_{1 i j} + α_{1} W P S_{1 j} + γ_{1} W P S_{1 j} \times G e n d e r_{1 i j} \\ + α_{2} F e m i c i d e R a t e_{2 j} + γ_{2} F e m i c i d e R a t e_{2 j} \times G e n d e r_{1 i j} \\ + \dots + α_{n} X_{n j} + \dots β_{n} X_{n i j} + ε_{i j}, \end{matrix}

(1)

where $α_{n} X_{n j}$ are the municipal-level control variables, $β_{n} X_{n i j}$ are the individual-level control variables, and $ε_{i j}$ is an error term.

Empirical Findings

Table 1 presents the results of the multilevel models that test H1. Model 1 shows that, on average, gender does not have a statistically significant effect on trust in the police. However, women appear less likely to perceive police effectiveness than men (Model 3). Baseline Models 1 and 3 also show that the coefficient associated with the WPS variable is not statistically significant. Yet, Models 2 and 4 indicate that there is a statistically significant interaction effect between the WPS and gender variables. For each dependent variable, the sign of the interaction is positive and statistically significant (p < .05). To evaluate the effect of WPS on police legitimacy for men and women, all constituent terms associated with an interaction term need to be taken into account simultaneously. Consequently, we estimate predicted probabilities and present the results of the interactive models graphically in Figures 2 and 3. In Panel A of Figures 1 and 2, we present mean predicted probabilities separately for men and women when the WPS variable takes a value of 1 or 0.¹²

Table 1.

Gendered Effects of Women’s Police Stations (WPS) on Police Legitimacy.

	Trust in the police		Perception of police effectiveness
	Model 1	Model 2	Model 3	Model 4
Gender (1 = female; 0 = male)	0.103	−0.141	−0.219*	−0.322
	(0.097)	(0.204)	(0.110)	(0.235)
WPS (1 = yes; 0 = no)	0.324	0.115	0.332	0.108
	(0.222)	(0.241)	(0.334)	(0.353)
WPS × Gender		0.414*		0.453*
		(0.191)		(0.219)
Femicide rate	−0.003	−0.002	0.015	0.036
	(0.024)	(0.028)	(0.037)	(0.042)
Femicide rate × Gender		−0.001		−0.037
		(0.029)		(0.034)
Municipal-level controls
Regular police station (1 = yes; 0 = no)	−0.177	−0.198	0.644	0.673
	(0.452)	(0.451)	(0.651)	(0.652)
Population size	−0.121^†	−0.122^†	−0.190^†	−0.194^†
	(0.066)	(0.066)	(0.102)	(0.102)
Log (Gross Domestic Product per capita)	−0.200	−0.194	−0.236	−0.236
	(0.131)	(0.131)	(0.195)	(0.196)
Individual-level controls
Crime victimization (1 = yes; 0 = no)	−0.314*	−0.319*	−0.789***	−0.806***
	(0.131)	(0.132)	(0.149)	(0.149)
Presidential approval	0.364***	0.367***	0.360***	0.363***
	(0.051)	(0.051)	(0.057)	(0.057)
Police officer asked for a bribe	−1.066***	−1.058***	−0.624*	−0.618*
	(0.244)	(0.244)	(0.263)	(0.263)
Educational level	−0.037*	−0.036*	−0.024	−0.020
	(0.015)	(0.015)	(0.018)	(0.018)
Household wealth	0.027	0.023	0.009	0.003
	(0.042)	(0.042)	(0.049)	(0.049)
Age	0.007^†	0.007^†	0.005	0.005
	(0.004)	(0.004)	(0.005)	(0.005)
No. of children	0.022	0.025	0.078*	0.084*
	(0.031)	(0.031)	(0.037)	(0.037)
Married or with partner (= 1; 0 = single)	−0.202	−0.202	−0.133	−0.138
	(0.123)	(0.123)	(0.140)	(0.140)
Divorced, separated, or widow	−0.052	−0.059	0.245	0.239
	(0.201)	(0.202)	(0.235)	(0.235)
Urban (= 1; 0 = rural)	−0.246	−0.252	0.199	0.193
	(0.192)	(0.192)	(0.280)	(0.281)
Indigenous (= 1; 0 = White)	0.131	0.134	−0.455	−0.440
	(0.314)	(0.314)	(0.375)	(0.376)
Black (= 1; 0 = White)	0.060	0.058	0.038	0.036
	(0.168)	(0.168)	(0.192)	(0.192)
Mulato (Pardo)	−0.040	−0.034	−0.091	−0.089
	(0.121)	(0.121)	(0.140)	(0.141)
Other (= 1; 0 = White)	0.079	0.085	−0.412	−0.433
	(0.272)	(0.270)	(0.315)	(0.315)
No. of municipalities	106	106	106	106
No. of Observations	1,439	1,439	1,427	1,427

Two-level ordered logistic multilevel models. Fixed effects at the state level included.

†

p < .10. *p < .05. ***p < .001 (standard errors in parenthesis).

Figure 1.

Trust in the police.

Figure 2.

Perception of police effectiveness.

To establish the statistical significance of observed differences between men and women, a formal difference in means test must be performed (Mitchell, 2012, p. 204). Thus, we proceed to estimate the average gender gap at each value of the WPS variable (1 or 0) and determine its statistical significance (Panel B). The gap is the difference in the mean predicted probabilities between men and women, with negative values indicating lower trust in the police or more negative perceptions of police effectiveness among women than men.¹³ Panel B of Figures 1 and 2 presents the gender gap in municipalities with and without a WPS and its respective 95% confidence intervals.¹⁴ If confidence intervals cross the zero line, this indicates that the gender gap is not statistically significant.

Panel A of Figure 1 shows the mean predicted probability of having a high level of trust in the military police (i.e., scoring a value of 5 or higher on the 1-7 scale). As can be observed, the average predicted probability in municipalities without a WPS is low and similar between men and women. However, we observe striking gender effects in municipalities with a WPS. Women’s mean predicted probability of having high trust in the police increases from 33.1% to 44.1% as one moves from municipalities without a WPS to those that have one. This represents a statistically significant increase of 11 percentage points (p < .05). Although we observe a slight increase in men’s probability as well, we find that this effect is not statistically significant, suggesting that men do not rely on heuristic information derived from WPS. Panel B of Figure 1 indicates that the effects shown in Panel A result in a positive statistically significant gender gap, namely women are more likely to trust the police than men in municipalities with a WPS. More specifically, women are 7.5 percentage points more likely to express high trust in the police than men in municipalities with a WPS, while there is not a gender difference in municipalities without a WPS. These results give strong support to H1.

Panel A of Figure 2 presents the mean predicted probabilities for men’s and women’s perceptions of police effectiveness, corresponding to the two highest values of the survey item (3 and 4 on the 1-4 scale). In municipalities without a WPS, women are less likely to find the police effective compared with men, with a probability of 37.0% and 44.9%, respectively. The gender gap is about 8 points and statistically significant (see Panel B of Figure 2). However, we observe that, compared with municipalities without a WPS, women’s probability of perceiving police effectiveness is higher in municipalities with a WPS by 11.8%. This effect is strong enough to close the gender gap in perceptions of police effectiveness. By contrast, although we observe a slight increase in men’s probability of perceiving police effectiveness, this effect is not statistically significant.

In sum, the effect of WPS on views of police effectiveness is more nuanced than we predicted. The establishment of a WPS improves the views of police effectiveness among women, but does not result in a positive gender gap as in the case of trust in the police. This suggests that trust and perceptions of effectiveness are indeed two distinct dimensions of police legitimacy. As shown above, men and women in municipalities without a WPS have a similar probability of trusting the police. Yet, in these municipalities, women are more critical of police performance than men. In municipalities with a WPS, women are as likely as men to perceive police effectiveness but trust the police more. It is important to note that these results hold after accounting for other municipal and individual-level characteristics.

As shown in Table 1, none of the municipal-level control variables included in the models exert a strong effect. At the individual level, three variables consistently predict both dependent variables. Personal experiences with crime and police corruption exert a negative and statistically significant effect. Presidential approval results in a significant positive effect.

Robustness Tests

To evaluate the robustness of our results that tests H1, we conduct a series of additional analyses. To explore the possibility that the effect of WPS on police legitimacy is confounded by local government action on VAW, we control for a variable that indicates whether a given local government has adopted VAW legislation or not. It is possible that municipalities with a WPS have more progressive legislation on VAW, which in turn improves views toward local political institutions, including the police. If this were the case, our results could be explained by local governments’ action on VAW in general, and not by the presence of a WPS. We find that the results presented in Figures 1 and 2 remain similar after controlling for VAW legislation at the municipal level, indicating that WPS have an independent effect on women’s perceptions of police legitimacy.¹⁵

We also conduct additional tests to assess the consistency of our theoretical argument with the empirical findings. We argued that feedback effects derived from WPS are likely to be reinforced by female officers’ descriptive representation in these police stations. Our theoretical discussion on policy feedback effects suggests that in the absence of police policies that favor women, the descriptive representation of women in the police by itself would result in more tenuous effects on police legitimacy. As such, our argument is consistent with previous studies showing that when a given policy benefits women, the extent of women’s involvement in the policy-making process is less influential in citizens’ evaluations of the rightness of policy outcomes (Clayton et al., 2019). To evaluate the effect of descriptive representation, we estimate models that account for the percentage of female police officers at the state level (the lowest unit of analysis at which data on the gender composition of the police are available). We find that a higher descriptive representation of female police officers does not result in improved perceptions of police legitimacy among women at standard levels of statistical significance.¹⁶ Consistent with our theoretical discussion, this finding suggests that the substantive representation of women resulting from WPS is largely driving the results we presented above, and not the effect of female officers’ descriptive representation on its own. However, further research is needed to determine the extent to which female police officers’ representation reinforces feedback effects from WPS.

Furthermore, we evaluate whether WPS affect primarily the legitimacy of the policy implementer, as we argue, rather than resulting in trust toward state institutions not directly involved in the provision of WPS services, such as political parties, the military, and congress.¹⁷ Consistent with our argument, the results of a placebo test show that WPS do not affect women’s trust toward institutions more distant from WPS service provision. WPS affect the political legitimacy of the policy implementer, the police. We describe additional robustness tests in the online appendix.¹⁸

Uncovering Underlying Mechanisms

In this section, we examine whether the effects described above are mediated by perceptions of government responsiveness and safety, as H2 and H3 posit. More specifically, we split the sample by municipality type (with or without a WPS) and estimate separate mediation models for each dimension of police legitimacy. We use the following two items in the LAPOP survey as mediating variables, each measuring perceptions of government responsiveness and safety, respectively: (a) On a scale from 1-7, where 1 means strongly disagree and 7 strongly agree, how much do you agree or disagree with the following statement: Those who govern this country are interested in what people like you think; (b) Speaking of the neighborhood where you live and thinking of the possibility of being assaulted or robbed, do you feel very safe (4), somewhat safe (3), somewhat unsafe (2), or very unsafe (1)? The models test the effect of each of these mediation variables simultaneously while accounting for the nested nature of the data (individuals within municipalities) and controlling for municipal and individual-level variables. We estimate the models employing Bayesian mediation analysis (Wang & Preacher, 2015; Yuan & MacKinnon, 2009).¹⁹

Figure 3 summarizes the results. The solid and dashed lines represent the statistically significant and non-significant links, respectively. Consistent with the results presented above, we observe that, in municipalities without a WPS, there is no direct effect of gender on trust in the police (see Panel A). In municipalities with a WPS (Panel B), however, women are more likely than men to trust the police,²⁰ indicating that the substantive representation of women invokes symbolic representation. Similarly, the results for our second dependent variable mirror those presented above. Women are less likely to perceive police effectiveness than men in municipalities without a WPS (Panel C).²¹ This gap, however, is no longer statistically significant in municipalities with a WPS (see Panel D).

Figure 3.

Mediation analysis.

In addition, we find that, contrary to what the literature suggests, the implementation of WPS does not affect women’s views of police legitimacy through its effect on perceptions of government responsiveness. In all models, gender is not a statistically significant predictor of perceptions of government responsiveness. Instead, the results we observe are largely driven by improved perceptions of safety among women. In municipalities without a WPS, women are less likely to feel safe than men (Panels A and C).²² In municipalities with a WPS, being a woman is no longer associated with more negative perceptions of safety (Panels B and D). Women’s fear of crime is similar to that of men. Importantly, perceptions of safety predict police legitimacy, suggesting that women’s improved perceptions of safety increase their level of legitimacy in municipalities with a WPS. In brief, rather than priming feelings of government responsiveness, women’s substantive representation from WPS invokes symbolic representation in the form of police legitimacy by making women feel safer. Women perceive WPS as a resource that improves their safety, which lead them to grant a higher level of legitimacy to the police.

The results in Figure 4, which are based on a multilevel model, further confirm the effects uncovered in our mediation analysis.²³ We observe that, in municipalities without a WPS, men’s probability of feeling somewhat or very safe is about 11% higher than women’s (see Panel A of Figure 4). However, men and women are equally likely to feel somewhat or very safe in municipalities with a WPS. As Panel B shows, in municipalities without a WPS, there is a negative and statistically significant gender gap in perceptions of safety. This gap disappears in municipalities with a WPS (i.e., confidence intervals associated with the gap overlap with the line set at zero value). Importantly, the gap closes as a result of women’s improved perceptions of safety, not that of men. Although we observe a slight improvement in men’s perceptions of safety in municipalities with a WPS, this effect is not statistically significant.

Figure 4.

Predicted probability of feeling somewhat or very safe.

Causal Inference: An Instrumental Variable Approach

Our theoretical insights suggest that the establishment of a WPS at the municipal level causes the gender effects we observe. The robustness tests we presented above give us confidence that this is the case. However, since the assignment of WPS to municipalities did not occur at random, it is still possible that unobserved factors might be confounding our results. In an effort to examine empirically the causal effect of WPS on police legitimacy, we adopt an instrumental variable approach.

As mentioned above, although the government has not established a population size threshold, it has promoted the establishment of WPS in larger municipalities. In our analyses, we control for population size along with other municipal-level factors likely correlated with the establishment of WPS and police legitimacy. Yet, even after controlling for important variables such as population size, municipalities with and without a WPS might still differ along unobserved characteristics. Although, on average, municipalities with a WPS are larger, population size varies drastically across all municipalities with a WPS in the country.²⁴ This suggests that, besides population size, other factors might influence the decision to establish a WPS in a given municipality, which can confound the effect of WPS on women’s views of police legitimacy.

One possibility is that the government has also been more likely to inaugurate a WPS in municipalities where women have demanded more police presence and involvement in the fight of VAW. As a result, one concern is that women in those municipalities had more positive attitudes toward the police to start with, suggesting that women’s views of police legitimacy pushed for the establishment of WPS and not the other way around. However, we consider this possibility unlikely. Given Brazil’s overall low levels of trust in the police, it is unlikely to observe higher trust toward the police among women than men in the absence of police resources aimed at improving women’s safety. Still, to account for the effect of unobserved factors and examine the causal effect of WPS empirically, we employ an instrumental variable strategy.

To determine the impact of WPS on women’s views of police legitimacy, we need to identify a source of exogenous variation that helps explain the establishment of WPS at the municipal level (i.e., an instrumental variable). In addition, the “exclusion restriction” (Angrist & Pischke, 2009) requires that, conditional on control variables, the instrumental variable has no gender effects on police legitimacy other than through its effect on the probability of having a WPS. We identify the age of the municipality in 1985, the year when the first WPS was established in Brazil, as an exogenous variable predicting the existence of a WPS in a given municipality in 2014.

We argue that the probability that a municipality would follow São Paulo’s example and establish a WPS is largely determined by the age of the municipality. Municipalities that did not exist in 1985 or have only recently been created are less likely to have a WPS today, independent of population size and other municipal-level characteristics. This is because the age of the municipality in 1985 serves as a proxy measure of local government institutional capacity. As the country democratized, changes in the 1988 constitution led to the proliferation of municipalities in Brazil. A total of 5,570 municipalities exist today, whereas in 1985, there were only 4,040 (Samuels, 2000). The proliferation of municipalities was largely driven by politicians’ desire to secure jobs and political power in local government rather than respond to the needs of local populations (Samuels, 2000). New municipalities were created even when they had limited capacity to collect taxes and implement sound local policies. As a result, the proliferation of municipalities deepened inequalities in “municipal capacity for the provision of welfare and/or local development services” (Samuels, 2000, p. 95).

As such, younger municipal governments are less likely to have the capacity to facilitate the establishment of a WPS. Given their more limited capacity, newer municipalities are more likely to concentrate their efforts on conventional areas of development typically viewed as local governments’ primary responsibilities, such as trash removal or city beautification. Although the police are in charge of WPS’ daily operations, the creation of a WPS in a given municipality depends on the ability of the municipal government to administer and provide the physical space for the stations (Santos & Pasinato, 2008). In addition, to establish a WPS, municipal governments need to have a bureaucratic apparatus capable of coordinating efforts with the state and federal governments (Santos & Pasinato, 2008). Newer municipal governments often lack the human capital with the training and experience to coordinate policy efforts that involve different levels of government. In short, long-standing municipalities are more likely to have the institutional capacity to create the local conditions necessary to have a WPS operating under their jurisdiction.

We argue that the age of a municipality in 1985, a proxy measure of local government institutional capacity, satisfies the exclusion restriction. We do not have a reason to believe that this variable predicts directly the gender effects depicted in Figures 1 and 2. Why would a municipality’s age in 1985 result in a positive effect on police legitimacy in 2014 among women, but not among men? We argue that the effect of municipality age in 1985 only results in differential gender effects on police legitimacy through its effect on the probability that a WPS exist today in a given municipality. In our dataset, we find significant variation in the age of municipalities. For example, eight municipalities included in the sample did not exist in 1985 and four others had less than 5 years of being constituted, while seven of them had more than 200 years of existence, including São Paulo.

We adopt an instrumental variable approach to examine causality by estimating a structural equation model that specifies WPS as an endogenous variable. The model can be characterized as having dependencies between response variables (StataCorp, 2017). To estimate this model, we need to account for the nested nature of the data. Therefore, we estimate a multilevel structural equation model with random effects for the intercept at the municipal level.²⁵ We specify a multilevel structural equation model with a system of two equations. The first equation includes an interaction term between WPS and gender as a predictor of police legitimacy as in Equation 1 above. In the second equation, the WPS variable is a function of the age of the municipality in 1985 plus municipal-level controls, including population size, GDP per capita, and femicide rates:

\begin{matrix} W P S_{j} = α_{0} + α_{1} M u n i c i p a l i t y A g e 1985_{1 j} + α_{2} P o p u l a t i o n_{2 j} \\ + α_{3} G D P p e r c a p i t a_{3 j} + α_{4} F e m i c i d e R a t e_{4 j} + δ_{j} . \end{matrix}

(2)

The results presented in the online appendix show that, controlling for other municipal-level characteristics, the age of a municipality in 1985 has a positive and statistically significant effect on the probability that a WPS exist today (p < .001).²⁶ Population size and GDP per capita also have a positive and statistically significant effect. Femicide rates do not predict the establishment of a WPS. Moreover, the results indicate that after accounting for the instrumental variable, the results presented in Figures 1 and 2 above remain similar. WPS appear to cause the gender effects we observe on trust in the police and perceptions of police effectiveness. The results from a model that test the effect of WPS on perceptions of personal safety using the instrumental variable approach described above also suggest that WPS lead to improved perceptions of personal safety among women, as observed in Figure 4.²⁷

Conclusion

The spread of specialized WPS has the potential to increase police legitimacy in democracies across the world. We show that, in the context of Brazil, women are more likely to attribute legitimacy to the police in municipalities with specialized police stations to address gender-based violence. Women in these municipalities are more likely than men to resort to heuristic information from WPS to form opinions on the trustworthiness and effectiveness of the police. As such, our findings are consistent with the notion that the substantive representation of women results in symbolic representation. We posit that the design features of WPS makes this policy particularly likely to generate feedback effects on women’s perceptions of police legitimacy. Besides providing resources to a large proportion of the female population, WPS are highly visible and traceable to government action. Moreover, WPS are primarily staffed by female officers, increasing women’s descriptive representation in these police stations.

Our research identifies some of the underlying mechanisms that link the existence of a WPS with improved perceptions of police legitimacy among female citizens. The results show that WPS generate positive effects on women’s views of police legitimacy by improving perceptions of safety. We find that, when women have access to WPS, they are less fearful of crime, resulting in an equal probability of fearing crime between men and women. This result challenges a long-standing finding in criminology indicating that women are typically more fearful of crime than men. Moreover, contrary to what the literature suggests, we do not find evidence that the effect of WPS on police legitimacy is mediated by perceptions of government responsiveness. Consequently, our findings suggest that the mechanisms linking policy and institutional legitimacy depend on policy type. WPS improve police legitimacy by making women feel safer, and not by triggering feelings of government responsiveness to one’s policy preferences.

Taken together, our findings have an important implication for women’s likelihood of reporting gender-based violence to the police. Given the well-documented positive link between police legitimacy and citizen cooperation with the police, the establishment of WPS has the potential to increase reporting rates. However, if WPS lack the resources to investigate crime and survivors do not have access to high-quality services, the spread of WPS is unlikely to increase reporting rates significantly. In addition, in the context of Brazil and other countries in the region, the positive feedback effect of WPS on police legitimacy can be offset by the high levels of police corruption and excessive use of police force. For WPS to promote a self-sustaining virtuous cycle of police legitimacy and higher reporting rates, access to WPS has to come along with enhanced service quality and police reform.

Supplemental Material

CordovaandKras_Online_Appendix_Final – Supplemental material for Addressing Violence Against Women: The Effect of Women’s Police Stations on Police Legitimacy

Supplemental material, CordovaandKras_Online_Appendix_Final for Addressing Violence Against Women: The Effect of Women’s Police Stations on Police Legitimacy by Abby Córdova and Helen Kras in Comparative Political Studies

Supplemental Material

CordovaandKras_Replication_Files_Final – Supplemental material for Addressing Violence Against Women: The Effect of Women’s Police Stations on Police Legitimacy

Supplemental material, CordovaandKras_Replication_Files_Final for Addressing Violence Against Women: The Effect of Women’s Police Stations on Police Legitimacy by Abby Córdova and Helen Kras in Comparative Political Studies

Footnotes

Acknowledgements

This research was conducted with the support of a research assistantship from the Office for Policy Studies on Violence against Women (OPSVAW) at the University of Kentucky (UK). We thank OPSVAW’s director, Dr. Carol Jordan, for her support to our research. A previous version of this article was presented at 2018 Women Studying Violence workshop organized by Dr. Lucía Tiscornia, sponsored by the Kellogg Institute for International Studies and other programs at the University of Notre Dame. We are thankful for the excellent feedback we received from participants in this and other academic meetings, including the 2018 Latin American PolMeth and MidWest Political Science Association conferences. We are also thankful to Dr. Scott Hutson, director of the Latin American, Caribbean, and Latino Studies program (LACLS) at UK, for organizing a workshop to discuss our project. We are also appreciative of the excellent comments we received from the editors and three anonymous reviewers.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

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

ORCID iD

Abby Córdova

Supplemental Material

Supplemental material for this article is available online at the CPS website

Notes

Author Biographies

Abby Córdova is an associate professor in the Department of Political Science at the University of Kentucky and a visiting fellow at the Kellogg Institute for International Studies at the University of Notre Dame. Her research includes topics related to violence against women, criminal violence, inequality, and international migration.

Helen Kras is a PhD student in the Department of Political Science at the University of Kentucky and the inaugural recipient of the Ashley T. Judd Distinguished Graduate Fellowship from the Office for Policy Studies on Violence Against Women. Her research focuses on the effects of violence against women policies on political behavior.

References

Angrist

J. D.

Pischke

(2009). Mostly harmless econometrics: An empiricist’s companion. Princeton, NJ: Princeton University Press.

Barnes

T. D.

Beaulieu

Saxton

G. W.

(2018). Restoring trust in the police: Why female officers reduce suspicions of corruption. Governance, 31, 143-161.

Barnes

T. D.

Córdova

(2016). Making space for women: Explaining citizen support for legislative gender quotas in Latin America. Journal of Politics, 78, 670-686.

Boateng

F. D.

(2017). Police legitimacy in Africa: A multilevel multinational analysis. Policing and Society, 28, 1105-1120.

Booth

Seligson

(2009). The legitimacy puzzle in Latin America: Political support and democracy in eight nations. Cambridge, UK: Cambridge University Press.

Calton

Cattaneo

L. B.

(2014). The effect of procedural and distributive justice on intimate partner violence victims’ mental health and likelihood of future help seeking. American Journal of Orthopsychiatry, 84, 329-340.

Campbell

A. L.

(2012). Policy makes mass politics. Annual Review of Political Science, 15, 333-351.

Cano

(2006). Políticas de segurança pública no Brasil [Public Security Policies in Brazil]. Revista Internacional de Direitos Humanos [International Journal of Human Rights], 5. Retrieved from https://sur.conectas.org/politicas-de-seguranca-publica-no-brasil/

Catterberg

Moreno

(2006). The individual bases of political trust: Trends in new and established democracies. International Journal of Public Opinion Research, 18, 31-48.

10.

Clayton

O’Brien

Piscopo

(2019). All male panels? Representation and democratic legitimacy. American Journal of Political Science, 63, 113-129.

11.

Corbacho

Pilipp

Ruiz-Vega

(2012). Crime and erosion of trust: Evidence for Latin America (Inter-American Development Bank Working Paper Series). Retrieved from https://www.econstor.eu/bitstream/10419/89042/1/IDB-WP-344.pdf

12.

Córdova

. (2009). Methodological note: Measuring relative wealth using household asset indicators. AmericasBarometer Insights, 6. Retrieved from https://www.vanderbilt.edu/lapop/insights/I0806en.pdf

13.

DataSenado. (2016). Rede de enfrentamento à violência contra as mulheres (DEAMs): Resultado de pesquisa [The Network of Services Combating Violence Against Women (DEAMs): Research Results]. Retrieved from https://www12.senado.leg.br/institucional/datasenado/arquivos/rede-de-enfrentamento-a-violencia-contra-as-mulheres-deams

14.

Easton

(1975). A re-assessment of the concept of political support. British Journal of Political Science, 5, 435-457.

15.

Eckstein

(1975). Case study and theory in political science. In Greenstein

F. J.

Polsby

N. W.

(Eds.), Handbook of political science (Vol. 7, pp. 79-137). Reading, MA: Addison Wesley.

16.

Esaiasson

Wlezien

(2017). Responsiveness beyond policy satisfaction: Does it matter to citizens? Comparative Political Studies, 50, 739-765.

17.

Farris

E. M.

Holman

M. R.

(2015). Public officials and a private matter: Attitudes and polices in the county sheriff office regarding violence against women. Social Science Quarterly, 96, 1117-1135.

18.

Ferraro

K. F.

(1996). Women’s fear of victimization: Shadow of sexual assault? Social Forces, 75, 667-690.

19.

Gattegno

Wilkins

Evans

(2016). The relationship between the Maria da Penha law and intimate partner violence in two Brazilian states. International Journal for Equity in Health, 15, 2-9.

20.

Hautzinger

S. J.

(2007). Violence in the city of women: Police and batterers in Bahia, Brazil. Berkeley: University of California Press.

21.

Hautzinger

S. J.

(2016). Policing by and for women in and beyond Brazil. In Bradford

Jauregui

Loader

Steinberg

(Eds.), The SAGE handbook of global policing (pp. 573-593). London, England: Sage.

22.

Htun

(2002). Puzzles of women’s rights in Brazil. Social Research, 69, 733-751.

23.

Sun

I. Y.

(2015). Chinese trust in the police: The impact of political efficacy and participation. Social Science Quarterly, 96, 1012-1026.

24.

Instituto Brasileiro de Geografia e Estatística. (2014). Perfil dos Municípios [Brazilian Institute of Geography and Statistics. (2014) Profile of Municipalities]. Retrieved from https://ww2.ibge.gov.br/home/estatistica/pesquisas/pesquisa_resultados.php?id_pesquisa=89

25.

Johnson

Maguire

Kuhns

(2014). Public perceptions of the legitimacy of the law and legal authorities: Evidence from the Caribbean. Law & Society Review, 48, 947-978.

26.

Jordan

Campbell

Follingstad

(2010). Violence and women’s mental health: The impact of physical, sexual, and psychological aggression. Annual Review of Clinical Psychology, 6, 607-628.

27.

Jubb

Camacho

D’Angelo

Hernández

Macassi

Meléndez

. . . Yáñez

(2010). Comisaría de la mujer en América Latina: Una puerta para detener la violencia y acceder a la justicia [Women’s Police Stations in Latin America: The Entrance Point To Deter Violence and Access Justice]. Centro de Planificación y Estudios Sociales [Center for Planning and Social Studies]. Retrieved from http://www.americalatinagenera.org/es/boletin/documentos/201101-acceso-justicia-cm-regional-final.pdf

28.

Kittilson

M. C.

(2010). Comparing gender, institutions, and political behavior: Toward an integrated theoretical framework. Perspectives on Politics, 8, 217-222.

29.

Kölln

(2016). The virtuous circle of representation. Electoral Studies, 42, 126-134.

30.

Latin American Public Opinion Project. (2014). AmericasBarometer survey 2014. Nashville, TN: Vanderbilt University. Retrieved from http://www.vanderbilt.edu/lapop/

31.

Lawless

J. L.

(2004). Politics of presence? Congresswomen and symbolic representation. Political Research Quarterly, 57, 81-99.

32.

Lü

(2014). Social policy and regime legitimacy: The effects of education reform in China. American Political Science Review, 108, 423-437.

33.

Malone

(2010). The verdict is in: The impact of crime on public trust in central American justice systems. Journal of Politics in Latin America, 2(3), 99-128.

34.

McDonagh

(2010). It takes a state: Policy feedback model of women’s political representation. Perspective on Politics, 8, 69-91.

35.

Mettler

(2015). Twenty years on Paul Pierson’s dismantling the welfare state? PS: Political Science & Politics, 48, 270-273.

36.

Mettler

Soss

(2004). The consequences of public policy for democratic citizenship: Bridging policy studies and mass politics. Perspectives on Politics, 2, 55-73.

37.

Mitchell

M. N.

(2012). Interpreting and visualizing regression models using Stata. College Station, TX: Stata Press.

38.

Norris

(1999). Critical citizens: Global support for democratic government. New York, NY: Oxford University Press.

39.

Organization for Economic Cooperation and Development. (2008). Handbook on constructing composite indicators: User guide. Retrieved from https://www.oecd.org/sdd/42495745.pdf

40.

Peffley

Hurwitz

(2010). Justice in America: The separate realities of blacks and whites. New York, NY: Cambridge University Press.

41.

Perova

Reynolds

S. A.

(2017). Women’s police stations and intimate partner violence: Evidence from Brazil. Social Science & Medicine, 174, 188-196.

42.

Pierson

P. L.

(1993). When effect becomes cause: Policy feedback and political change. World Politics, 45, 595-618.

43.

Pitkin

(1967). The concept of representation. Berkeley: University of California Press.

44.

Reisig

M. D.

Tankebe

Mesko

(2012). Procedural justice, police legitimacy, and public cooperation with the police among young Slovene adults. Journal of Criminology and Security, 14, 147-164.

45.

Samuels

(2000). Reinventing local government? The evolution of Brazil’s municipalities. In Kingstone

Power

(Eds.), Democratic Brazil: Actors, institutions and processes (pp. 77-100). Pittsburgh, PA: University of Pittsburgh Press.

46.

Santos

C. M.

(2010). From women’s police stations to the Maria da Penha law: The state’s absorption/translation of women’s demands. Revista Crítica de Ciências Socias, 89, 153-170.

47.

Santos

C. M.

Pasinato

(2008). Mapeamento das delegacias da mulher no Brasil. Núcleo de estudos de gênero Pagu [Mapping Women’s Police Stations in Brazil. Department of Gender Studies], Universidade estadual de Campinas [Campinas State University], CEPLAES-IDRC. Retrieved from https://www12.senado.leg.br/institucional/omv/entenda-a-violencia/pdfs/mapeamento-das-delegacias-da-mulher-no-brasil

48.

Schafer

Huebner

Bynum

(2006). Fear of crime and criminal victimization: Gender-based contrasts. Journal of Criminal Justice, 34, 285-301.

49.

Schwindt-Bayer

L. A.

(2010). Political power and women’s representation in Latin America. New York, NY: Oxford University Press.

50.

Schwindt-Bayer

L. A.

Mishler

(2005). An integrated model of women’ representation. Journal of Politics, 67, 407-428.

51.

Secretaria de Políticas para as Mulheres. (2010). Norma técnica de padronização das delegacies especializadas de atendimento às mulheres (DEAMs) [Secretariat of Public Policies for Women (2010). Technical Note on the Standardization of Women’s Police Stations (DEAMs)]. Retrieved from http://www.biblioteca.presidencia.gov.br/presidencia/dilma-vana-rousseff/publicacoes/orgao-essenciais/secretaria-de-politica-para-mulheres/norma-tecnica-de-padronizacao-das-delegacias-especializadas-de-atendimento-a-mulheres-25-anos-de-conquista

52.

Sharp

Johnson

P. E.

(2009). Accounting for variation in distrust of local police. Justice Quarterly, 26, 157-182.

53.

Snijders

T. A.

Bosker

(2012). Multilevel analysis: An introduction to basic and advanced multilevel modelling (2nd ed.). London, England: Sage.

54.

StataCorp. (2017). Structural equation modeling manual. College Station, TX: Author. Retrieved from https://www.stata.com/manuals/sem.pdf

55.

Tankebe

(2009). Policing, procedural fairness and public behaviour: A review and critique. International Journal of Police Science & Management, 11, 8-19.

56.

Tankebe

(2013). Viewing things differently: The dimensions of public perceptions of police legitimacy. Criminology, 51, 103-135.

57.

Tyler

Huo

(2002). Trust in the law: Encouraging public cooperation with the police and courts. New York, NY: Russell Sage Foundation.

58.

United Nations. (2015). The world’s women 2015: Trends and statistics. Retrieved from https://unstats.un.org/unsd/gender/downloads/worldswomen2015_report.pdf

59.

United Nations Women. (2011). Women’s police stations in Latin America case study: An entry point for stopping violence and gaining access to justice (Brazil, Peru, Ecuador, and Nicaragua). Security Sector Module, 1-9. Retrieved from http://www.endvawnow.org/uploads/browser/files/security_wps_case_study.pdf

60.

Waiselfisz

(2015). Mapa da violência 2015: Homicidio de mulheres no Brasil [Map of Violence 2015: Homicides of Women in Brazil]. Retrieved from https://www.mapadaviolencia.org.br/pdf2015/MapaViolencia_2015_mulheres.pdf

61.

Wang

Preacher

K. J.

(2015). Moderated mediation analysis using Bayesian methods. Structural Equation Modeling: A Multidisciplinary Journal, 22, 249-269.

62.

Warr

(1984). Fear of victimization: Why are women and the elderly more afraid? Social Science Quarterly, 653, 681-702.

63.

Weatherford

(1992). Measuring political legitimacy. The American Political Science Review, 86, 149-166.

64.

Weldon

(2002). Protest, policy, and the problem of violence against women: A cross-national comparison. Pittsburgh, PA: University of Pittsburgh Press.

65.

Yuan

MacKinnon

D. P.

(2009). Bayesian mediation analysis. Psychological Methods, 14, 301-322.

66.

Zaller

(1992). The nature and origins of mass opinion. New York, NY: Cambridge University Press.

67.

Zetterberg

(2009). Do gender quotas foster women’s political engagement? Political Research Quarterly, 62, 715-730.

68.

Zhao

J. S.

Scheider

Thurman

(2002). The effect of police presence on public fear reduction and satisfaction: A review of the literature. The Justice Professional, 15, 273-299.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.32 MB

0.00 MB

0.57 MB