Temporal stability analysis of crash injury severity in school zones: A mixed logit modeling approach

School zones present critical challenges where vehicular traffic intersects with vulnerable pedestrians, especially children. While previous studies have examined school zone crash risk and countermeasures, few have investigated how injury severity risk factors change over time especially before and during the COVID-19 pandemic leaving a gap in understanding the temporal stability of these influences. This study analyzes 3638 police-reported crashes in Louisianna school zones from 2018 to 2021 to identify factors influencing injury severity. Using a Random Parameters Logit Model with Heterogeneity in Means and Variances, the analysis accounts for unobserved heterogeneity and temporal changes, comparing pre-pandemic (2018–2019) and pandemic (2020−2021) periods. Results show that crashes near state highways significantly increase the likelihood of fatal, severe, or moderate injuries. In contrast, lower speed limits (≤ 25 mph), improved lighting conditions, and seasons such as summer and autumn correlate with reduced injury severity. Driver behaviors like failure to yield and weekday crashes increase possible injury risks. Temporal instability in risk factors highlights the need for regionally and temporally calibrated interventions. These findings support targeted safety measures such as traffic calming near high-speed roads, enhanced enforcement, and educational campaigns to reduce injury severity and improve safety for all road users in school zones. • The factors influencing crash severity in school zones are explored. • The study examines the stability of risk factors before and after the COVID-19 pandemic. • A random parameters logit model with heterogeneity in means and variances is employed. • Temporal instability of factors is identified via a likelihood ratio test. • Failing to yield, weekdays, lack of traffic control, and daylight conditions consistently increase severity in school zones.