Safety-belt usage has increased significantly in the US since the introduction of mandatory safety-belt usage laws in the 1980s. This paper analyzes the impact of these laws on increasing safety-belt usage while controlling for other state-specific variables. A fixed effects cross-sectional time-series analyses shows the relative significance of various state-level attributes in explaining safety-belt usage, including whether or not primary or secondary safety-belt laws have been passed. To further explore these relationships we employ spatial analyses techniques and find spatial autocorrelation in the data. Spatial correlation also exhibits a clear east-west direction. When the analyses is further corrected for temporal autocorrelation we find that the spatial autocorrelation is greatly diminished and that many variables lose their statistical significance, though safety-belt laws are still statistically significant. Results suggest that for this data, it is critical to control for temporal autocorrelation while spatial autocorrelation is less important. We also find that our spatial analyses does provide interesting information on similarities between various regions on the effectiveness of safety-belt laws. (A) "Reprinted with permission from Elsevier".
Abstract