Introduction
The coronavirus disease (COVID-19) pandemic and subsequent social and physical restrictions put in place to address this crisis have caused personal and economic stressors and uncertainty about the future on a national level. One of the main concerns during disaster conditions is that pressures put on households may increase domestic conflict, including verbal and physical fights between partners. Reference Gearhart, Perez-Patron and Hammond1 To understand the impact of COVID-19 on domestic conflict, we assessed reports of increased verbal and physical fights in a nationally representative sample of the U.S. adult population. In the wake of disaster, rates of domestic conflict often increase in the weeks/months following the event. Reference Gearhart, Perez-Patron and Hammond1,Reference Rung, Gaston and Oral2 Economic disasters, such as periods of recession, are also related to increases in domestic violence (e.g., 2007–2009) and have been linked to underlying conditions, like prevalence of economic uncertainty. Reference Schneider, Harknett and McLanahan3
Evidence suggests that risk for domestic conflict is not evenly distributed. Socially vulnerable populations (e.g., women, low socioeconomic status, people of color, low education level) experience disproportionate losses and high levels of stress following natural and economic disasters, and increases in domestic violence. Reference Schneider, Harknett and McLanahan3,Reference Enarson4 Such findings suggest that economically vulnerable groups may be at increased risk for domestic conflict in the early days of the COVID-19 pandemic. Additionally, a longer duration of disaster condition exposure is positively associated with the number of domestic assaults. Reference Gearhart, Perez-Patron and Hammond1 The effect of duration of disaster on domestic violence outcomes would also suggest that certain groups need continued assistance during prolonged disaster conditions and in the months that follow.
Measuring domestic conflict is often challenging, especially in the midst of a disaster, like COVID-19. In the United States, domestic violence hotlines have reported that calls have been brief but intense, perhaps due to fewer private moments and opportunities to call for help. Reference Li and Schwartzapfel5 In late spring of 2020, just a few months into the COVID-19 pandemic, several major US city police departments reported a drop in reported domestic violence rates as the prevalence of COVID-19 and stay-at-home orders continued nationwide, which may have been due to difficulty in reporting problems and getting help during these times. Reference Li and Schwartzapfel5
While such findings provide specific examples of changes in domestic violence reporting during COVID-19, the samples are not representative of the American population. Thus, the use of a nationally representative sample to directly assess nationwide personal reports of verbal or physical altercations at home during the early days of the COVID-19 pandemic is invaluable for investigating increases in domestic conflict during this unique crisis. Our goal is to directly assess increases in domestic conflict in the 2 months after the onset of the COVID-19 outbreak and stay-at-home orders in the United States in a nationally representative sample to identify risk factors that can help target public health interventions to reduce domestic conflict during crisis conditions. Findings may help shed light on particularly vulnerable groups that need attention from policy-makers soon after disaster occurs and groups that should be considered when thinking about the continued consequences of COVID-19, as well as other potential disasters, in the months to come.
Methods
Sample
A survey related to the impact of COVID-19 was fielded between May 1 and May 6 in the RAND American Life Panel (ALP). The ALP is a nationally representative sample of US adults who have been recruited almost entirely through random digit dialing. Respondents periodically complete surveys by computer or phone and are compensated for completed surveys. Respondents lacking Internet service are provided with an Internet subscription and a netbook computer to ensure representativeness. Reference Carman and Nataraj6 A total of 2622 panel members were invited to participate in the survey. The survey was closed after 2052 panel members participated, leading to a 78% participation rate. Using the ALP weighting methodology, Reference Carman and Nataraj6 the sample was weighted to account for non-response and to match the characteristics of the March 2019 Current Population Survey with respect to age, gender, race/ethnicity, household income, household size, and education. Characteristics of the sample are described in an open access report. Reference Breslau, Finucane and Locker7
Measures
Domestic conflict questions were asked of respondents who indicated they had a partner (n = 1196). The questions asked whether the (1) number or (2) intensity of physical or verbal arguments with their partner increased, decreased, or stayed the same since the beginning of the COVID-19 outbreak began. Respondents were also asked whether household income changed or whether they were worried their household income would be negatively impacted by COVID-19. Those with a decrease in household income and those who were very/somewhat worried that their income would be negatively impacted were considered to have financial worry.
Statistical Analysis
Chi-square tests were used to assess increases in domestic conflict by demographic characteristic and by financial worry. A logistic regression model controlling for sex, age, race/ethnicity, education, income, and financial worry was used to investigate risk of increases in either the number or intensity of domestic conflict versus no change or decrease. Analyses were conducted using SAS software (SAS Institute, Inc., Cary, NC). All data analyses used sampling weights and adjusted for the complex sample design.
Institutional Review Board Approval
The American Life Panel survey was approved by the RAND Human Subjects Protection Committee (HSPC). The HSPC is RAND’s Institutional Review Board to review research involving human subjects, as required by federal regulations. RAND’s “Federalwide Assurance for the Protection of Human Subjects” (FWA00003425, effective through June 22, 2023) serves as our assurance of compliance with the regulations of 16 federal departments and agencies.
Results
An increase in either the number or the intensity of domestic conflicts since the beginning of the COVID-19 pandemic was reported by 10.6% of the sample; 9.0% reported an increase in number, and 7.6% reported an increase in intensity. Table 1 displays the proportion reporting an increase in either number or intensity of domestic conflict by demographic characteristics prior to the pandemic. In these unadjusted comparisons, increase in domestic conflict is significantly associated with age and education, but not with gender, race/ethnicity, or income. Increase in domestic conflict is higher in younger respondents and in those with some high school education.
Table 1. Proportion reporting an increase in domestic conflict during the first 2 months of the COVID-19 pandemic
* p-value <0.05.
Significant associations identified in unadjusted comparisons were sustained in adjusted regression models with pre-COVID-19 demographic characteristics as predictors (Table 2, Model 1). Respondents of ages 20–39 years had 2.58 times higher odds (95% CI: 1.19-5.6) than respondents of age 60 or over. Compared with respondents with a high school degree, odds were higher both in people with lower levels of educational attainment. These associations are sustained when financial worry during the pandemic is included as a predictor (Model 2). Financial worry is associated with higher odds of increase in domestic conflict (OR = 2.07, 95% CI: 1.09-3.92).
Table 2. Associations of demographic characteristics and financial worry with increase in domestic conflict during the first 2 months of the COVID-19 pandemic
Note: Model 1 and Model 2 are logistic regressions of increase in domestic conflict on the listed predictors. Model 2 adds financial worry (vs no financial worry) as a predictor to the list of predictors in Model 1.
Discussion
Our findings suggest that, during the first 2 months of the COVID-19 pandemic, about 1 in 10 (10.6%) American adults in relationships experienced an increase in verbal or physical domestic conflict. Both younger individuals and individuals with lower levels of education, in particular, reported a significant increase in domestic conflict. These findings are in line with past domestic conflict reports that show that low income, lower education level, and younger age are all risk factors for intimate partner violence. Reference Benavides, Berry and Mangus8 In periods of economic downturns, younger and lower educated groups of individuals often report increases in stress following job loss. One potential reason for this outcome may be the lack of immediately available resources to support unexpected needs during times of crisis, which may result in higher stress and more tumultuous home conditions in the first few weeks to months following onset of a disaster. In the case of the COVID-19 pandemic, economic and societal changes have forced lifestyle changes for many. For younger and lower educated groups, there may be limited options for ways to maintain normal daily life in the short time period following disaster, and this may contribute to more stress and subsequent increases in domestic conflict in these households.
With the increase in unemployment rates during COVID-19, it is not surprising that financial worry is a significant stressor for many individuals. Past studies have shown that economic uncertainty is related to increases in intimate partner violence during economic recession, more than household unemployment or material hardships. Reference Rung, Gaston and Oral2 Our findings support this idea as increases in domestic conflict were higher in groups with financial stress than those without during COVID-19. Another recent survey from April 2020 found that COVID-19 caused financial stress for almost 9 in 10 Americans. 9 A reduction of income is not the only type of financial stressor, but other stressors include uncertainty and stress around the potential for loss of income, including things like job security (39%) and market volatility (25%). 9 Such findings are similar to RAND’s larger comprehensive assessment of COVID-19 outcomes that revealed that respondents were somewhat worried about their finances (39%) due to potential for income reduction during COVID-19. Reference Carman and Nataraj6 While results from this study report on financial stressors in the first months of COVID-19, further investigation of the persistence of financial stressors over time is necessary to understand whether such stressors continue and how they contribute to or exacerbate rates of domestic conflict.
Additional findings from the RAND survey have revealed that COVID-19 has affected low-income households in much different ways than middle- or high-income households. The former study reports that individuals in low-income households may resort to selling possessions, borrowing money, or not being able to meet expenses if they have major income changes during COVID-19. Reference Carman and Nataraj6 While the Federal Pandemic Unemployment Compensation has been instrumental in helping lower income households meet immediate financial needs, this limited financial support may not be enough to reduce financial stress on a large scale or in a persistent way. European countries have offered business loans for those harmed by COVID-19 restrictions and shutdowns, Canada has provided monthly assistance to citizens who have lost income due to the pandemic, and the Netherlands pledged to replace up to 90% of lost wages, due to the pandemic, for eligible companies. The point-in-time assistance programs that the United States are using are large and widespread and do not target those groups who may be more at risk/more in need of aid because of risk of income loss during COVID-19. More research should be done to understand long-term effects of US assistance approaches compared with other countries’ approaches and what are the best next steps to support individuals who are experiencing continued uncertainty surrounding the ability to pay for household items or the backlog of financial issues that accumulate over the course of a disaster.
The nationwide stay-at-home-orders and physical distancing mandates associated with COVID-19 have also amplified feelings of concern. COVID-19 has generated a constant fear of illness, uncertainty surrounding safe places to go, concern over availability of food/resources, and questions regarding how long COVID-19 and social distancing restrictions will last. Additional findings from our group have revealed that COVID-19 has negatively impacted mental health outcomes on a national level. Reference Breslau, Finucane and Locker7 Because of high levels of financial stress that higher risk groups are experiencing, such groups may not have the funds to seek mental health assistance, despite the heightened need for it in the wake of increasing mental health issues. Government programs that increase availability of counseling or provide financial assistance for mental health services during the COVID-19 pandemic and additional crises may be one way to help prevent increases in domestic conflict in higher risk groups. Because longer durations of disaster conditions are associated with increases in domestic assaults, Reference Gearhart, Perez-Patron and Hammond1 the fact that our study shows increases in financial worry and domestic conflict early on in the outbreak of COVID-19 may suggest continued stress in the months after the COVID-19 onset and may amplify increases in domestic conflict over time. Thus, finding ways to provide/enhance a variety of resources for vulnerable populations who may be at higher risk for domestic conflict, early on after a disaster strikes, may be helpful in preventing domestic conflict increases following disasters.
Conclusion
Findings from this study may offer support for policy-makers pushing for continued programs that provide direct support to high-risk groups. Such changes could help mitigate increases in domestic conflict that may increase as additional COVID-19-related or other disaster conditions occur.
Author Contributions
AL wrote and revised the manuscript. MF and JB assisted in writing the manuscript and made major edits and additions. ER performed the data analysis. KC designed the survey and organized and executed data collection.
Funding Statement
Support for this research was provided by gifts from RAND supporters and income from operations, as well as NIMHD Grant #RO1010274 and NSF Grant #2028091.
Conflict(s) of Interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this paper.
Open access
