Debiasing Law Enforcement Officers: Evidence from an Expressive Arts Intervention in India

Authors: Sofia Amaral, Girija Borker, Nishith Prakash,Maria Nicaela Sviatschi

Abstract: Addressing institutional service-delivery related to gender-based violence is a major policy challenge. Despite its policy relevance, there exists considerable gaps in the literature as to how to address the issue. In this project we aim to address these gaps by focusing on the role policing quality may play in addressing GBV. We conducted a randomized control trial (RCT) in Bihar, India, to study the causal effect of a novel intervention that uses expressive arts pedagogy for police officers on technical and soft skills, norms and behaviors regarding GBV. For our study, we randomly assigned 419 police stations (PS) across 12 districts in the state to either receiving the training or a control group. Senior officers operating in treated stations are invited to attend the training. We test the effect of the intervention on new measures that aim to better understand how better interactions between citizens and officers may improve GBV survivors’ outcomes. We develop new tools to measure victim blaming, empathy and knowledge of GBV-specific laws among others. We also measure the effects of the intervention on different types of women interacting with officers, and through a decoy experiment.

Explaining Poor Police Predictions of Domestic Abuse

Authors: Jeffrey Grogger, Andrew Jordan, and Tom Kirchmaier

Sexual harassment and the leaky pipeline in academia

Authors: Rune Vesman, Jane Greve, Johanna Rickne, Olle Folke

Abstract: This project studies if and how sexual harassment affects careers. We focus on academia, a high-paying, male-dominated sector with clear career paths. We survey 10 cohorts of PhD students in Denmark and link these survey data to administrative data on career outcomes. The analysis shows a high prevalence of sexual harassment for women during their Ph.D., but not for men. A large share of this harassment comes from senior colleagues and takes place inside and outside the workplace, for example at evening events and conferences. Departments with a high prevalence of harassment against women contribute more to the leaky pipeline. The results also suggest pursuing an academic career puts ambitious women in harm’s way and makes sexual harassment a price that women pay for a continued career in academia.

Keeping the Peace while Getting Your Way: Information, Persuasion and Intimate Partner Violence

Authors: Dan Anderberg, Rachel Cassidy, Anaya Dam, Wendy Janssens, Karlijn Morsink and Anouk van Veldhoven

Abstract: We study the effect on intimate partner violence (IPV) of new information received by women only, men only, or both, relevant to a high-stakes household decision made under incomplete information. We model strategic communication of new information between spouses as Bayesian persuasion and develop a theoretical framework for IPV where abuse arises from spousal disagreements. Our framework offers the novel prediction that new information when received by only one spouse may lead to lower IPV incidence as compared to when received by both spouses as a privately informed spouse will seek to forge agreement through strategic communication (“persuasion”), reducing conflict. To test the model's predictions, we leverage an existing randomized controlled trial of an edutainment intervention addressing child marriage decisions in rural Pakistan, targeted at men, women, or both. Our empirical findings confirm the prediction that the likelihood of IPV is highest when men and women are jointly targeted. Due to systematic gender differences in preferences, our model further predicts that marriage delays are largest when targeting men alone or jointly with women and smallest when targeting women alone, predictions that are also confirmed in the data.

Firm responses to legislation on handling sexual harassment

Authors: Sonia Bhalotra, Medha Chatterjee, Kanika Mahajan, Daksh Walia, Fan Wang

Income & IPV: Signal of Strength or Struggle? - Evidence from Edutainment in Bangladesh

Authors: Nina Buchmann, Paula Lopez-Pena, Sakib Mahmood, Atonu Rabbani

Abstract: Intimate Partner Violence (IPV) remains widespread globally, and it disproportionally affects low-income households. In this paper, we study the role of mental resilience and signaling in explaining the prevalence of violence and the income gradient. We develop a simple signaling model to explain violence in an environment in which men are perceived to be either "strong" or "weak" and both income and a woman's behavior are perceived to be signals about the man's type. We show that men might use violence under two conditions: i) they lose control, and ii) they are naïve about the signaling value of violence, namely that violence and type are substitutes in explaining a woman's behavior. We experimentally test the model through a clustered-randomized control trial in rural Bangladesh in which we test the effectiveness of two "edutainment" interventions, soap operas with subtle messaging. The first intervention aims to correct men's beliefs about the signaling value of violence, namely that strong men don't use violent punishment ("signaling treatment"). The second intervention additionally aims to teach men mental resilience and anger management techniques ("signaling + resilience treatment"). To study the importance of self- and social image concerns, i.e., husband's first- and second-order beliefs about their type, we randomly divide the signaling treatment into two groups: a private arm, in which men are exposed to the edutainment intervention privately, and a public arm, in which men watch the intervention in a community setting. We measure changes in IPV prevalence and mental health status as compared to a placebo treatment without subtle messaging.

From Addiction to Aggression: The Spillover Effects of Opioid Policies on Intimate Partner Violence

Authors: Dhaval Dave, Bilge Erten, Pinar Keskin, Shuo Zhang

Abstract: Substance abuse is a major risk factor for intimate partner violence (IPV). We investigate how a key supply-side intervention – the abuse-deterrent reformulation of a widely-diverted opioid, OxyContin – affected IPV. Our results indicate counties with greater baseline rates of prescription opioid usage experienced relatively larger declines in IPV after OxyContin’s reformulation. The reformulation reduced IPV only in states with small illicit drug markets, while states with large illicit drug markets experienced increased heroin-involved IPV due to substitution towards illicit opioids. Our results underscore the importance of identifying populations at high risk of substitution to illicit opioids and moderating this risk with evidence-based policies.

Alcohol availability and violence against women

Authors: Susan Niknami, Arizo Karimi, Hanna Mühlrad, Petra Ornstein, Anna Sandberg

Abstract: We analyze a large-scale policy experiment in Sweden, initiated in 2000, where the government required all alcohol retail stores in selected areas to stay open on Saturdays. This experiment aimed to assess the potential social impacts of increased alcohol availability over weekends. Using rich individual level data, we show that this increase in alcohol availability raised both alcohol use and violence against women.

Overstretched: Financial Distress and Intimate Partner Violence in the U.S.

Authors: Chiara Santantonio, Olivia Masi

Abstract: This study examines the effect of financial distress within households on intimate partner violence in the United States. By leveraging the timing of bank closing days and fixed wage payment schedules, we identify months when households have to stretch their finances due to changes in their regular payment schedules. Using monthly records from the National Crime Victimization Survey, we find that these shocks increase the likelihood of women experiencing IPV. We further corroborate the assumption of worsened financial conditions during these periods by documenting changes in household behavior, in terms of expenditures and time use. Using Consumer Expenditure Survey (CEX) data, we show that households reduce their expenditures on food and leisure during periods of financial stretch. Additionally, using American Time Use Survey (ATUS) data, we find an increase in time devoted to consumption research. We also observe that spouses spend more time together, particularly with their children, during these periods.

Anti-Sexual Harassment Amenities at Work

Authors: Puneet Arora, Lori Beaman, Anisha Sharma, Karmini Sharma

Abstract: to follow

When Do People Condone Sexual Harassment? A Cross-Cultural Study of Trade-offs

Authors: Sonia Bhalotra, Matthew Ridley, Mateusz Stalinski

Abstract: Sexual harassment policies vary in their scope and precision, with the uncertainty reducing both deterrence and workplace collaboration between men and women. As a first step to mitigate this uncertainty, we study social acceptance for sexual harassment and how it depends on behaviour’s context. We recruit US adults in full employment for an online experiment where we display vignettes describing workplace interactions. We randomize the severity of harassment (e.g., asking out, sexual comments, touching) and the context, such as location (workstation vs. coffee shop), perpetrator’s seniority (boss vs. peer) or conversation topic (work-related vs. private). We measure people’s willingness to report perpetrators to HR. We also introduce a novel incentivized measure of the willingness to condemn, with participants able to make a financial sacrifice to reduce the payoff of the perpetrator – a person who admitted to behaviour consistent with the vignette in a separate survey. We report several preliminary results. First, we find that the willingness to condemn is highly responsive to the severity of sexual harassment. Second, we document that seniority, location, and conversation topic significantly affect the likelihood of HR reporting. However, the strongest moderator, with the effect size about three times as large, is whether harassment was repeated or a one-off. We also report that emotional encouragement and being in the ethnic in-group of the perpetrator have insignificant moderating effects. Lastly, we document important misperceptions, with people overestimating both the likelihood of strict sanctions and the extent to which exacerbating factors affect HR reporting.

Perceptions of workplace sexual harassment and support for policy action

Authors: Sonia Bhalotra, Matthew Ridley

Abstract: Workplace sexual harassment remains common and imposes substantial harms on women. We propose to study how perceptions of the prevalence of sexual harassment and its harms to victims affect support for (stronger) policy action against it in the UK. Using a survey experiment, we estimate quantitative perceptions of prevalence and harms and estimate the treatment effect of providing statistical information on each on policy support. We additionally test whether information on policy effectiveness can increase the efficacy of these treatments. Finally, we plan to investigate whether policymakers hold misperceptions that may influence sexual harassment policy.

Victim-blaming Norms and Violence Against Women

Authors: Sevinc Bermek, Konstantino Matakos, Asli Unan

Abstract: Intimate partner violence (IPV) entails huge socioeconomic and psychological costs, yet policy and behavior that combat it respond sluggishly. Are prevailing victim-blaming social norms responsible? What can change them to enact policy, and behavior change? We address these with a survey experiment on a representative sample of Turkish population (N = 4,000). We incentivize both the elicitation of social norms and subjects’ behavior. Leveraging a within-subject design, we provide information correcting any (mis)perceptions about victim-blaming norms. We find that attitudes and policy/behavior are influenced by different channels, respectively. While correcting misperceptions about norms (through information) causes a positive change in attitudes, it does not cause policy or behavior change. The former simply reflects conformity: initial perceptions about social norms were too “pessimistic”. In contrast, the process of eliciting higher-order beliefs about others’ views (norm) --absent any information provision-- forces subjects to engage in relative moral evaluations and causes a sizable policy and behavior response. The latter is driven by subjects who perceive themselves to be more virtuous (less victim-blaming) than average and donate to IPV prevention. Thus, moral comparisons triggered by introspection, raise the salience of norms generating `soft learning' that causes policy and behavior change. The policy implication is that, while changing societal attitudes is a long-run process, welfare-improving social change is possible in the short-run even without changing peoples' values.

Schools as Safety Nets: Breakdowns and Recovery in Reporting of Violence Against Children

Authors: Damian Clarke, Pilar Larroulet, Daniel Pailañir, Daniela Quintana,

Abstract: Schools are a key channel in formal reporting of violence against children, but this channel broke down with the onset of the COVID19 pandemic. We study how widespread such reporting declines are, and to what extent they were recovered once schools reopened. Examining the universe of all criminal reports of violence against children in Chile, we observe sharp declines in reporting of all types of violence (psychological, physical, and sexual), and that full recovery in reporting had not occurred, even nearly 2 years following initial school closures. Extending beyond the unexpected and long school closures during the pandemic, we find evidence of clear declines in violence reporting during regular school vacations, and in a period of student strikes at the secondary level suggesting a broader relevance of these results.

Gender-based violence in schools and girls’ education: Experimental evidence from Mozambique

Authors: Sofia Amaral, Aixa Garcia-Ramos, Selim Gulesci, Sarita Oré, Alejandra Ramos, Maria Micaela Sviatschi

Abstract: Gender-based violence (GBV) at schools is a pervasive problem that affects millions of adolescent girls worldwide. In partnership with the Ministry of Education in Mozambique, we developed an intervention to increase the capacity of key school personnel to address GBV and to improve students' awareness as well as proactive behaviors. To understand the role of GBV on girls' education, we randomized not only exposure to the intervention but also whether the student component was targeted to girls only, boys only, or both. Our findings indicate a reduction in sexual violence by teachers and school staff against girls, regardless of the targeted gender group, providing evidence of the role of improving the capacity of key school personnel to deter perpetrators. Using administrative records, we also find that in schools where the intervention encouraged proactive behavior by girls, there was an increase in their school enrollment, largely due to an increased propensity for GBV reporting by victims. Our findings suggest that effectively mitigating violence to improve girls’ schooling requires a dual approach: deterring potential perpetrators and fostering a proactive stance among victims, such as increased reporting.

Female empowerment and male backlash: Experimental evidence from India

Authors: Claire Cullen, Sarthak Joshi, Joseph Vecci, and Julia Talbot-Jones

Abstract: The unintended consequences of women’s empowerment are rarely measured and remain poorly understood. We study the impact of female empowerment on male backlash through a series of experiments involving 1,007 households in rural India. We find that men pay to punish empowered women at double the rate of women in an otherwise identical control group. We find that backlash occurs regardless of how women are empowered, with social image concerns being a key driver. Finally, we test several policies to reduce backlash and find that reframing empowerment programs to emphasize broader community benefits can help mitigate backlash.

Safe Travels: Transport Development and Women’s Safety in India

Author: Anisha Garg

Abstract: Beyond fostering inclusion and connectivity, does investing in mass transit impact crimes against women? I study the impact of the expansion of the metro rail system in Delhi, India, on reported crimes against women. Preparing a new and novel data set of the universe of crime reports at the police station-day level and leveraging the staggered opening of metro stations across these police jurisdictions between 2016 and 2019 in Delhi, I find a 29 percent decrease in reported incidents of sexual harassment against women in public spaces following the opening of the first metro station in the police jurisdiction. This reduction in reported sexual harassment is neither linked to an overall decrease in reported non-gender-specific crimes in public spaces, nor substituted by an increase in such crimes in metro stations or trains. Further analysis of crime reports from alternative public transport networks, combined with area safety data from a crowd-sourced mobile application, suggests that this reduction is driven by the enhanced safety provided by the metro network.

Sentencing Severity and Domestic Violence: Evidence from Brazil

Authors: Rui Costa, Olivia Masi, Beatriz Ribeiro, Matteo Sandi

Abstract: We estimate the impact of sentencing severity on the dynamics of domestic violence. The study uses ten years of merged individual-level administrative registers on domestic violence cases brought to the police and family linkages for Rio Grande do Sul (Brazil). Leveraging Brazil’s “Lei do Feminicidio”, which was implemented in March 2015 to include the crime of “femicide” in the Brazilian penal law, we find that sentencing severity significantly affects the behavior of both offenders and victims of domestic violence. While the policy change seemingly deterred potential offenders by reducing the incidence of domestic violence, victims of domestic violence became more likely to ask for protective measures and more reluctant to press charges against their abusive partners, as a framework of compensating mechanisms would predict. For a policymaker seeking to design effective sentences to combat domestic violence, the tension between these outcomes appears critical.