Start date: 05/01/06
End date: 08/31/06
Other authors: Tom Stout
About the research
With the quickening pace of crash reporting, the statistical editing of data on a weekly basis, and the ability to push working databases to CTRE/Iowa Traffic Safety Data Service, the University of Iowa, and Iowa DOT users, we have a situation where databases that would be considered incomplete by past standards of static data files are in "public use" even as the dynamic nature of the central DOT database allows changes to be made in both the aggregate of data and in the individual crashes already reported. The "definitive" analyses of serious crashes will, by their nature, lag seriously behind preliminary files. Even after these anlyses, the dynamic nature of the mainfram data file means that crash numbers can continue to change long after the incident year.
The Iowa DOT, its Office of Driver Services (the data owner), and institutional data users/distributors must establish data use, distribution, and labeling protocols to deal with the new, dynamic nature of data. In order to set these protocols, data must be collected concerning the magnitude of difference between database records and crash narratives and diagrams. This study determines the difference between database records and crash narratives for the Iowa Department of Transportations Office of Traffic and Safety crash database and the impacts of this difference.
Report: Crash data validation: An Iowa case study (390 kb pdf) February 2007
Sponsor(s): Iowa Department of Transportation