1370196 - Developing an Automated Methodological Procedure for Improving the Identification of Curve-Related Crashes in the Texas Crash Records Information System (CRIS)

Horizontal curves play a vital role in improving the safety and comfort of a ride for drivers and passengers by preventing a sharp turn from one direction to another. However, the safety impacts of curves on vehicle crashes in Texas are underestimated due to the substantial number of curve-related crash misclassifications in the Crash Records Information System (CRIS). Therefore, there is a need to improve the identification of curve-related crashes in CRIS. This study conducted a comprehensive analysis to check data consistency in CRIS. It was found that approximately 77.3 percent of 2017-2020 crash records have consistent curve-related information, whereas 22.7 percent of crash records contain internally inconsistent curve attributes, showing in one data field that the crash is curve-related but in the other data field that the crash did not occur on a horizontal curve. Based on this finding, a systematic methodological procedure was developed and tested to be effective in improving the identification of curve-related crashes in CRIS using the Highway Curves GIS layer provided by Texas Department of Transportation (TxDOT). Furthermore, this study categorized curve-related crash misclassifications into six types. The results showed that, on average, 27.1 percent of crashes in CRIS 2017–2020 were misclassified in terms of whether they were curve-related. Finally, using the Texas Peace Officer’s Crash Reports (CR-3) of misclassified crashes, the study investigated potential causes for curve-related crash misclassifications in CRIS.