Traffic fatalities in 50 U.S. States and District of Columbia from 1970-2024

Traffic fatalities in 50 U.S. States and District of Columbia were analyzed from 1970-2024 using a 1970 reference. Benchmarks were defined for groups of States with High, Mid, and Low traffic fatality ratios (smallest to largest fatality reductions). Relationships were determined for fatality ratios based on State population, registered vehicles, vehicle mileage (VMT), and gross domestic product (GDP).

The fatality ratio (e.g., D₂₄/D₇₀) was calculated by the number of traffic fatalities each year (e.g., D₂₄) divided by the fatalities in 1970 (D₇₀) for each State. The fatality reduction was 1.00 - D₂₄/D₇₀. The ratios and States were sorted from the highest to the lowest (smallest to largest fatality reductions) at nine times. Benchmarks for the High, Mid, and Low fatality ratios were determined using the average ± standard deviation for the three groups, resulting in generally non-overlapping benchmarks from 1970 to 2024. Characteristics were determined for population (P), registered vehicles (N), VTM (V), and GDP (G). Ratios were determined for the characteristics (e.g., P₂₄/P₇₀). Linear, multi-linear, power, and natural logarithm regression models were determined for the fatality ratio (D₂₄/D₇₀) based on State characteristic ratios. Categorical analyses identified three strata for State fatality ratios.

In 2024, the High fatality ratio group (D₂₄/D₇₀) of States averaged 1.229 ± 0.386, the Mid 0.682 ± 0.115, and the Low 0.400 ± 0.079. The High group increased traffic fatalities 22.9 ± 38.6% since 1970. The Low group reduced fatalities by 60.0 ± 7.9%. Benchmarks for the three groups were developed from 1970 to 2024. The strongest regression models depended on ratios. A predictive model was D₂₄/D₇₀ = 0.349 + 0.749ln(P₂₄/P₇₀), R² = 0.769, p < 0.0001. Regression of raw counts for D₂₄ resulted in negative coefficients for key characteristics, rendering them impractical. The analyses identified Alaska, Utah, Idaho, Minnesota, New Hampshire, and Massachusetts had fatality ratios below predicted by regression and categorical analyses (i.e., fewer traffic deaths). Florida, Arizona, Oklahoma, Nevada, Arkansas, Connecticut, Kentucky, West Virginia, and Pennsylvania had fatality ratios above predicted (i.e., more deaths).

Benchmarks were determined for the High, Mid, and Low fatality ratio States (smallest to largest fatality reductions) from 1970 to 2024. The Low ratio States reduced fatalities 60.0 ± 7.9% by 2024. The lowest fatality ratios for ten countries were statistically lower at 0.142 ± 0.017 (t = 6.68, p < 0.0001), an 85.8 ± 1.7% reduction since 1970. Regression analyses found the natural logarithm of population (P₂₄/P₇₀) significantly correlated with State fatality ratios (D₂₄/D₇₀). The number of registered vehicles (N) and GDP (G) were not significant predictors. VMT (V) had a 2^nd order influence on traffic fatalities.