Problem definition: The U.S. nursing home industry has suffered decades-long understaffing and high worker turnover, with 99% of nursing homes surveyed in 2021 reporting being short-staffed. Although it is known that understaffing can reduce the quality of nursing home services and care, we examine the under-recognized possibility that understaffing may harm patients by incentivizing unethical behavior. Methodology/results: We empirically study the causal relationship between nursing home staffing and “chemical restraints,” i.e., the use of antipsychotic medications to improperly sedate difficult patients with dementia. Using nationwide data obtained from the U.S. Centers for Medicare & Medicaid Services through Freedom of Information Act requests, our causal analyses exploit local market wages for nurses and changes in states’ minimum-staffing regulations as sources of exogenous variation in nursing homes’ staffing levels. We find that, on average, every 15 minutes of increased daily nurse staffing hour per resident reduces a nursing home’s usage of antipsychotic drugs by 1.51 residents or approximately 1.54% of its resident population. Managerial implications: Our findings demonstrate a robust causal relationship between staffing and the widespread use of potentially abusive and harmful practices, raising the alarm for managers and policymakers alike. In particular, our estimates predict that the new federal minimum staffing standards for nursing homes introduced in April 2024 will reduce the nationwide usage of antipsychotic drugs by 24,148 nursing home residents, or 9.6%.