|
Field
Abstract
|
Before
Domestic violence (DV), defined as stalking, rape, or physical violence, is a global problem with 35% of women worldwide reporting experiencing DV (WHO 2017). More than 12 million people experience domestic violence in the U.S. each year (National DV Hotline, 2017), with severe implications for health (Coker et al., 2000), human capital (Borker, 2018), and child wellbeing (Aizer, 2011; Kavanaugh, Trako and Sviatschi, 2019). In addition, DV also has intergenerational consequences: children who experience DV are 7 times more likely to become perpetrators (St Jude House 2020). Furthermore, several studies have shown that unemployment (Harknett and McLanahan, 2016), social networks, and financial and emotional stress affect domestic violence (Brush, 2003; Benson and Fox, 2004).
Coronavirus (COVID-19), a highly infectious disease with no known cure or prevention till date (WHO, 2020), has caused one of the greatest macroeconomic shocks ever to hit the world. Coronavirus cases in the U.S. have now surpassed all other countries (CDC, 2020), and in an effort to prevent further spread, most businesses deemed non-essential have been temporarily closed. Unemployment claims in the final week of March rose from 282,000 to 6.65 million—more than four times the record in a single week (Bui and Wolfers, 2020). Beyond this, many still-employed Americans are being forced to work from home and asked to leave their homes only for necessities.
In the wake of COVID-19, countries from Australia to France to Brazil to the U.S. are experiencing unprecedented reports of domestic violence; the National Domestic Violence Hotline recently urged those at risk for DV to make safety plans as perpetrators began to use “the virus as scare tactic to threaten or isolate victims,” (Peterman et al., 2020). Many countries have already began implementing broad measures to curb violence in the wake of coronavirus; Australia has allocated $1bn for mental health services related to DV (The Guardian, 2020) and France has begun using hotels to keep victims safe as reports of DV have skyrocketed by 30% (Hume, 2020).
With DV rates rising sharply during COVID-19, an important question from a policy perspective is arising: What can be done to help DV survivors in the era of this pandemic? There are many barriers that keep DV victims from reporting abuse or leaving relationships; in this project, we will focus on four important barriers: lack of resources and information, low self-esteem and shame, self-blame, and incorrect beliefs about what abuse looks like.
We aim to test four interventions that are likely to determine pathways to aid victims of DV during the current pandemic and future mandated stay at home orders: our interventions target information constraints, belief constraints, and barriers to reporting such as victim self-blaming.
|
After
Rates of domestic violence (DV) are increasing rapidly in the era of the COVID-19 pandemic. What are the barriers that victims of DV face in reporting and leaving abusive relationships? How can policymakers eliminate the obstacles that victims face? This project aims to answer these questions through a novel intervention that targets information constraints, belief updating, and victim self-blaming. We use Amazon’s Mechanical Turk and Prolific to survey a set of 6,000 women and look at common barriers that women face when in an abusive relationship; we then provide treatments that give victims access to resources for reporting, information about the health status of their relationships, and videos that break down self-blame and encourage self-esteem building. Our research attempts to analyze what the most important barriers are for victims of domestic violence, and how they can reduce self-blaming, increase emotional and psychological wellbeing, and ultimately report or leave abusive relationships.
|
|
Field
Last Published
|
Before
June 23, 2020 11:12 AM
|
After
September 28, 2020 12:09 PM
|
|
Field
Intervention (Public)
|
Before
See Pre-Analysis Plan
|
After
We aim to test four interventions that are likely to determine pathways to aid victims of DV during the current pandemic and future mandated stay at home orders: our interventions target information constraints, belief constraints, and barriers to reporting such as victim self-blaming.
|
|
Field
Primary Outcomes (End Points)
|
Before
Primary outcomes for our project include information seeking, information sharing, and learning from our self-blame video.
● Information Seeking: We will ask respondents if they visited the National Domestic Violence Hotline site, as well as questions about the content of the site to check respondents’ truthfulness about visiting the site. We will also ask respondents if they used the National Domestic Violence Hotline anonymous chat. In this way, we will measure the take up and use of the resources that we provide in our treatments.
● Information Sharing: We will ask respondents if they expect to share information about the health of their relationship with friends or family; we will also ask about their willingness to share information with a counselor. In this way, we measure their willingness to share and speak about their relationship and its health.
|
After
Primary outcomes for our project include information seeking, information sharing, and relationship beliefs.
● Information Seeking: We will ask respondents if they visited the National Domestic Violence Hotline site, as well as questions about the content of the site to check respondents’ truthfulness about visiting the site. We will also ask respondents if they used the National Domestic Violence Hotline anonymous chat. In this way, we will measure the take up and use of the resources that we provide in our treatments.
● Information Sharing: We will ask respondents if they expect to share information about the health of their relationship with friends or family; we will also ask about their willingness to share information with a counselor. In this way, we measure their willingness to share and speak about their relationship and its health.
● Relationship Beliefs: We ask respondents in our baseline their beliefs about a variety of behaviors and if these behaviors qualify as abusive. In our baseline, large gaps are evident: many women rate physical violence as abuse but do not rate behaviors such as controlling what a partner wears or isolating a partner from family as abusive. We will measure these beliefs again at endline in order to see if women update their beliefs about what qualifies as abusive behavior. In this way, we measure women’s understanding of forms of abuse such as psychological and emotional abuse.
|
|
Field
Experimental Design (Public)
|
Before
See Pre-Analysis Plan
|
After
We will randomize our 6000 subjects into the five treatment arms (including one control arm). We will restrict respondents on demographic elements; namely, we will only sample women who are currently partnered and living with their partners. From this sample, we will randomize the 6000 women into the five arms at the individual level.
|
|
Field
Randomization Method
|
Before
See Pre-Analysis Plan
|
After
Randomization done by a computer
|
|
Field
Randomization Unit
|
Before
See Pre-Analysis Plan
|
After
Individual
|
|
Field
Planned Number of Clusters
|
Before
8000
|
After
6,000
|
|
Field
Planned Number of Observations
|
Before
8000
|
After
6,000
|
|
Field
Secondary Outcomes (End Points)
|
Before
The secondary outcomes that we will look at are relationship conditions, relationship expectations, psychological wellbeing, and attendance to a One Love Foundation workshop.
● Relationship Conditions: We will ask respondents a series of questions about the health of their relationship, including questions that measure physical, emotional, and mental abuse.
● Relationship Expectations: Our survey asks respondents their satisfaction with their relationship and their beliefs about the future of their relationship. We will measure how these answers change when women are updated about the health of their relationship, given information about resources available to them if they are in an unhealthy relationship, and given information to help break cycles of self-blame and embarrassment. Beliefs about relationships will be elicited in both endlines to measure how sensitive these effects are to fade out.
● Relationship Beliefs: We ask respondents hypothetical questions about what they believe abuse is. For instance, we ask them if they think it is completely healthy, healthy, somewhat unhealthy, unhealthy, or abusive for a partner to check another partner’s texts and calls. In this way, we measure beliefs about abuse in general, that are not specific to a woman’s relationship. We hope that our interventions will increase awareness around what abuse is and looks like.
● Psychological Wellbeing: To measure psychological wellbeing, we use questions from the GHQ-12 scale of depression and anxiety, as well as from the “Psychological Well Being Index.” Both indices allow us to look at levels of stress, depression, anxiety, and self-worth in respondents. We also will use a measure of self-esteem, risk aversion, and self-worth from Gallup; we use questions from the Cognitive Emotion Regulation Questionnaire to measure changes in behavior in dealing with adverse events—for instance, we look at whether respondents are more likely to blame themselves when in an adverse situation, or to try and positively refocus their attention.
● Locus of Control: We will use questions to determine respondent’s “locus of control,” or how strongly they believe they have control over situations that happen to them. People with an external locus of control are more likely to have higher anxiety and attribute bad situations to fate; they don’t believe they have control over their lives. We will test for changes in locus of control.
● Self-blame Learning: In our fourth treatment arm, we ask respondents to summarize the contents of a video that covers domestic violence and victim self-blaming. We look at their summaries as well as their willingness to blame victims in DV incidents. We measure their response to statements such as “Domestic violence victims are often as responsible for the situation as perpetrators are.” Beyond this, we also look at updates in respondents blaming themselves with questions from Gallup’s polls about self-blame, self-image, and self-esteem.
● One Love Workshop Attendance: At the end of our baseline survey, we will ask participants if they are interested in attending a workshop on unhealthy and healthy relationships put on by our partner, One Love Foundation. We measure attendance to this workshop as a secondary and exploratory outcome.
|
After
The secondary outcomes that we will look at are: relationship conditions, relationship expectations, economic expectations, psychological wellbeing, and attendance to a One Love Foundation workshop.
● Relationship Conditions: We ask respondents a series of questions about the health of their relationship, including questions that measure physical, emotional, and mental abuse. In this way, we check to see if there are any changes in the conditions of women’s relationships before and after treatment.
● Relationship Expectations: Our survey asks respondents their satisfaction with their relationship and their beliefs about the future of their relationship. We look at whether women plan to have children with their current partner and how long they expect to be in their current relationship. We will measure how these answers change when women are updated about the health of their relationship, given information about resources available to them if they are in an unhealthy relationship, and given information to help break cycles of self-blame and embarrassment. Beliefs about relationships will be elicited in both endlines to measure how sensitive these effects are to fade out.
● Economic Expectations: Our surveys ask respondents about their future expectations to make economic decisions with a partner. We ask about large purchases such as homes and car, joint savings and education accounts, and willingness to open private savings accounts without a partner’s knowledge.
● Psychological Wellbeing: To measure psychological wellbeing, we use questions from the GHQ-12 scale of depression and anxiety. This index allows us to look at levels of stress, depression, anxiety, and self-worth in respondents. We also will use a measure of self-esteem, risk aversion, and self-worth from Gallup.
● Locus of Control: We will use questions to determine respondent’s “locus of control,” or how strongly they believe they have control over situations that happen to them. People with an external locus of control are more likely to have higher anxiety and attribute bad situations to fate; they don’t believe they have control over their lives. We will test for changes in locus of control.
● Self-blame Learning: In our fourth treatment arm, we ask respondents to summarize the contents of a video that covers domestic violence and victim self-blaming. We look at their summaries as well as their willingness to blame victims in DV incidents. We measure their response to statements such as “Domestic violence victims are often as responsible for the situation as perpetrators are.” Beyond this, we also look at updates in respondents blaming themselves with questions from Gallup’s polls about self-blame, self-image, and self-esteem.
● One Love Workshop Attendance: At the end of our baseline survey, we will ask participants if they are interested in attending a workshop on unhealthy and healthy relationships put on by our partner, One Love Foundation. We measure attendance to this workshop as a secondary and exploratory outcome.
|