Traffic fatalities in 40 countries and 50 U.S. states plus DC in 1970 and 2024

This study completes a series on traffic fatalities in U.S. states and international countries. It compares traffic deaths, population, fatality rates, and fatality ratios, between states and countries, and it develops regression equations for traffic deaths based on population changes since 1970.

Traffic fatalities and population were determined for 40 countries (including the U.S.) and 50 U.S. states plus DC (n = 91) in 1970 and 2024. Fatality rates (D₂₄/P₂₄), ratios (D₂₄/D₇₀) and other factors were determined using deaths in 2024 (D24), population in 2024 (P24), and deaths in 1970 (D70) for the states and countries. Welch's t-test compared differences between states and countries (excluding the U.S.). Fatality ratios (D₂₄/D₇₀) were clustered from lowest (safest) in strata 1, mid in strata 2, and largest (least safe) in strata 3. 3 × 2 contingency analysis determined the association between the number of states and countries in strata 1, 2, and 3. Regression analyses determined D₂₄/D₇₀ based on P₂₄/P₇₀ in 91 states and countries using linear, power, and natural logarithms. Categorical regression separated the relationship between D₂₄/D₇₀ and ln(P₂₄/P₇₀) into three strata with the same slope and different D₂₄/D₇₀ offsets to cluster the states and countries. A regression model was determined for only the countries (excluding the U.S.). The U.S. data was fed into the model to determine the expected D₂₄ if the states acted like the countries.

T-tests showed the countries were significantly safer with: 1) 4.19 ± 2.19 deaths/100,000 population compared to 12.34 ± 4.35 deaths/100,000 in states (t = 10.6, p < 0.00001), and 2) 0.387 ± 0.307 fatality ratios (D₂₄/D₇₀) compared to 0.734 ± 0.330 in states (t = 5.140, p < 0.0001). Fatality ratios were grouped in three statistically different strata. Strata 1 had D₂₄/D₇₀ = 0.286 ± 0.126, strata 2 had 0.744 ± 0.154 (t_1-2 = 16.6, p < 0.0001), and strata 3 had 1.721 ± 0.175 (t_2-3 = 12.7, p < 0.0001). Strata 1 had 70.0% countries. Strata 2-3 had 76.5% states. The difference was significant (χ² = 19.7, p < 0.0001). Regression found D₂₄/D₇₀ depended on population changes with D₂₄/D₇₀ = 0.305 + 0.668ln(P₂₄/P₇₀), R² = 0.501, p < 0.0001. The regression equation was separated into 3 strata with different intercepts (R² = 0.918) to differentiate the countries and states. Analysis showed that if U.S. states acted like countries, U.S. fatalities would be 11,552 ± 1,708 in 2024. The actual deaths were 40,242.

The 39 countries (excluding the U.S.) and 50 states plus DC were similar in size, GDP, and fatalities in 2024, but they differed significantly in traffic safety. Traffic fatalities dropped 26.3% in U.S. states from 54,633 in 1970 to 40,242 in 2024. They dropped 69.6% in countries from 130,533 to 39,746; the difference was significant (χ² = 1.05*10⁴, p < 0.00001). If the U.S. states improved traffic safety like the countries the past 50 years, there would be 28,690 ± 1,708 fewer traffic deaths in the U.S. in 2024.