Sample-selection in highway safety data exists because riskier drivers are over-represented in crash data bases.
Sample-selection is studied by considering crash data before and during the COVID-19 pandemic.
Riskier drivers are found to make up a larger proportion of vehicle miles traveled during the pandemic than before.
Data analyses show changes in driver behavior are largely responsible for increasing injury severities.
Islam, M., Alogaili, A., Mannering, F., Maness, M., 2023. Evidence of sample selectivity in highway injury-severity models: The case of risky driving during COVID-19. Analytic Methods in Accident Research 38, 100263.