Domestic Violence in the Netherlands During the First 15 Weeks of the COVID-19 Pandemic: A Longitudinal Study

The study investigates whether and how the COVID-19 pandemic and associated restrictions affected domestic violence in the Netherlands. Early media and scholarly concerns suggested lockdowns could increase domestic violence due to stress, isolation, financial strain, and reduced opportunities to escape or seek help. Empirical evidence internationally has been mixed, and Dutch administrative indicators had not shown clear increases. This study leverages pre-pandemic and early-pandemic police-handled domestic violence cases in a large Dutch city to examine: (RQ1) changes in levels of reported domestic violence; (RQ2) changes in the nature and characteristics of incidents (type, location, reciprocity, injuries, suspect/victim profiles, relationships, case handling); and (RQ3) changes in the levels and nature of minors’ involvement (as suspects, victims, witnesses). The goal is to provide evidence-based insight into short-term impacts of early COVID-19 measures on domestic violence.

Prior research on natural disasters (earthquakes, hurricanes, tsunamis, floods) suggests possible increases in domestic violence due to disruption, stress, and mental health impacts, though findings are mixed and often limited by design and data (e.g., Keenan et al., 2004; Schumacher et al., 2010; Rao, 2020; Cerna-Turoff et al., 2019). Reviews indicate inconsistent associations, with some studies finding increases and others no change or decreases (Bell & Folkerth, 2016; Rao, 2020). Early COVID-19 literature documented spikes in police calls and reports in many countries and meta-analytic evidence of an average 8% increase, especially in the U.S. (Piquero et al., 2021; UN Women, 2020; Boserup et al., 2020), while other contexts showed no change or declines, including some Dutch sources and studies focused on youth or child abuse reports (Kruisbergen et al., 2020; Steketee et al., 2020; Baglivio et al., 2022; Kourti et al., 2021). In the Netherlands specifically, police and Safe at Home (Veilig Thuis) reporting volumes in 2020 were similar to or slightly below 2019, though professionals noted some increased severity and a shift toward neighbor reports (Kruisbergen et al., 2020; Van Gelder et al., 2021; Coomans et al., 2021). Overall, evidence on short-term impacts of COVID-19 measures on domestic violence is mixed across settings.

Design and context: A quasi-longitudinal comparison using police-handled domestic violence incidents processed via the Dutch Public Prosecution Service’s rapid case handling stream (ZSM). ZSM handles frequently occurring, less severe cases suitable for out-of-court settlement or rapid prosecution; more severe cases are typically excluded from ZSM. Data sources and periods: All domestic violence incidents handled by a ZSM location in a large Dutch city (~500,000 inhabitants) during two 15-week windows: pre-pandemic (Aug 16–Nov 30, 2019) and early pandemic (Mar 16–Jun 30, 2020). Due to lockdown constraints, the intervening period (Dec 2019–mid-Mar 2020) could not be collected. Sample: N=434 incidents (206 pre-pandemic; 228 pandemic). Definition: Domestic violence as per Dutch Public Prosecution Service—violence committed by someone in the victim’s domestic circle (partners, ex-partners, family, close friends); forms include physical, sexual, psychological (property destruction also recorded). Measures extracted from criminal files using a structured topic list: date; type of violence (physical, psychological, destruction, combination); reciprocity (mutual vs. not); injuries (none, light, severe); location (home of suspect/victim, public area, other, multiple); reporter to police (victim via emergency number, witness via emergency number, victim at station, victim via social worker); suspect and victim characteristics (gender, age); suspect–victim relationship ((ex-)partners, parent/child in both directions, other family, friends, roommates); minors’ involvement (<18 years: as suspect, victim, witness; number of minors involved, ages); case processing outcomes (public prosecutor vs. court; settlement types and court decisions). Follow-up: Each case followed through mid-February 2021 to capture dispositions. Analytic approach: Descriptive statistics and group comparisons between periods using t-tests for mean weekly incidents and actor ages; chi-square tests for categorical distributions (incident types, reciprocity, injuries, locations, reporters, relationships, suspect/victim gender, settlement contents and distributions, court decisions); day-of-week comparisons via t-tests. Ethical approval: Authorized by the Dutch Public Prosecution Service Board of Prosecutors General; approved by the Ethics Committee, Department of Law and Criminology, Vrije Universiteit Amsterdam.

Levels over time: 206 reports in 15 weeks pre-pandemic vs 228 in 15 weeks during the pandemic; mean weekly incidents did not differ (M=13.7 vs 15.2; t(28) = −0.90, p = .38). No significant differences by matched day-of-week. A peak in weeks 20–21 (May 11–25, 2020) coincided with primary school reopening but was not tied to changes in minors’ involvement. Nature and characteristics of incidents: • Type of violence: Distribution did not differ (χ2(3)=2.33, p=.51). Physical violence predominated (pre: 75%; pandemic: 72%); combinations increased modestly (pre: 11%; pandemic: 16%). • Reciprocity: Among cases with data (~half), reciprocal violence decreased during the pandemic (pre: 42% vs pandemic: 18%; χ2(1)=12.94, p<.01). • Injuries: No differences in none/light/severe (χ2(2)=0.12, p=.94); severe injuries were rare (2% both periods). • Location: Shift toward homes during the pandemic; distribution differed (χ2(5)=27.07, p<.001). More incidents at home(s) of suspect/victim; fewer in other locations. • Reporter to police: Distribution differed (χ2(3)=9.01, p<.05). During the pandemic, incidents were more often reported by witnesses via emergency number and less often by victims at the police station. Suspects and victims: • Suspect gender: Higher proportion male during the pandemic (92% vs 85%; χ2(1)=4.79, p<.05). • Victim gender: No change (χ2(1)=0.10, p=.75). • Ages: No differences (suspects: ~36 years; victims: ~37–38 years). • Relationship: Similar distribution (χ2(6)=8.69, p=.19). About three-quarters involved (ex-)partners; child-to-parent in ~12–15%; parent-to-child ~4%. Case processing: • Share handled by prosecutor vs court unchanged (≈70% prosecutor, 30% court; χ2(1)=0.07, p=.80). • Prosecutor settlement content differed (χ2(5)=15.48, p<.05): more conditional dismissals (pandemic 35% vs pre 21%), fewer unconditional dismissals (pandemic 47% vs pre 67%). • Court decisions did not differ (χ2(8)=13.68, p=.09). Involvement of minors: • Pre-pandemic: minors involved in 34% (71/206); pandemic: 43% (98/228); increase not statistically significant (χ2(1)=3.30, p=.07). • Role distribution unchanged (χ2(2)=1.89, p=.39): minors mostly witnesses (83–84%); smaller shares as main suspect (pre 8%; pandemic 4%) or main victim (pre 8%; pandemic 12%). • Ages: Witnesses younger than minor suspects and victims; age patterns similar across periods.

Findings indicate that, in this Dutch city and for ZSM-handled cases, overall police-reported domestic violence volumes did not increase during the first 15 weeks of COVID-19 restrictions compared to a pre-pandemic period. This contrasts with increases documented in many international settings but aligns with other Dutch indicators. While overall levels were stable, situational characteristics shifted in ways consistent with lockdown conditions: more incidents occurred at home, and reporting patterns shifted toward witness-initiated calls and fewer in-person victim reports at stations. Violence was less often reciprocal during the pandemic, and suspects were more frequently male; the severity as inferred from injuries did not change. Case processing volumes were stable, though prosecutors used conditional dismissals more often and unconditional dismissals less often. Minors’ involvement appeared somewhat higher in percentage terms but was not statistically significant; their roles (primarily witnesses) and age patterns remained similar. These results suggest that early-pandemic restrictions altered the context and reporting of domestic violence more than they changed overall incident counts in this setting. The study contributes to nuanced understanding that pandemic impacts on domestic violence are context-dependent and may not uniformly elevate incident volumes, even while changing incident dynamics.

This study leverages comparable pre- and early-pandemic police/ZSM data from a large Dutch city to show that domestic violence levels remained stable during the first 15 weeks of COVID-19 restrictions, while incident context shifted toward homes, reciprocal violence declined, suspects were more often male, and reporting by witnesses increased. The severity of injuries and relationship patterns did not change, and minors’ involvement rose nonsignificantly, with roles remaining predominantly as witnesses. These findings temper concerns of drastic short-term surges in domestic violence in the Netherlands, while highlighting shifts in dynamics and reporting consistent with lockdown conditions. Future research should: (1) monitor longer-term trajectories as measures persist or evolve; (2) incorporate non-ZSM and more severe cases; (3) address potential seasonal effects and include continuous time coverage; (4) integrate self-report and service-use data to assess reporting biases; and (5) examine heterogeneity across socio-demographic groups most affected by pandemic stressors.

• Causality cannot be inferred; no control group unaffected by restrictions. • Temporal gap between pre- and pandemic samples (Dec 2019–mid-Mar 2020 missing) limits continuous time analyses and raises potential seasonal confounding. • ZSM scope excludes more severe domestic violence cases; findings may not generalize to all severity levels. • Reliance on police-reported incidents may be affected by changes in willingness/ability to report (e.g., victims’ reduced mobility vs. increased neighbor presence at home). • Limited information on sample diversity and socio-economic heterogeneity; differential impacts on disadvantaged groups cannot be evaluated. • Short observation window focuses on early, short-term effects; longer-term impacts remain unknown.