Show simple item record

A method for estimating delta-V distributions from injury outcomes in crashes

dc.contributor.authorFlannagan, Carol A.C.en_US
dc.date.accessioned2016-04-15T14:59:14Z
dc.date.availableNO_RESTRICTIONen_US
dc.date.available2016-04-15T14:59:14Z
dc.date.issued2013-06
dc.identifier103241en_US
dc.identifier.otherUMTRI-2013-15en_US
dc.identifier.urihttps://hdl.handle.net/2027.42/117575
dc.description.abstractTwo key national crash data samples from the National Automotive Sampling System (NASS) program are the General Estimates System (GES) and the Crashworthiness Data System (CDS). The former is a larger (50,000 crashes per year) sample of police-reported crashes and contains only information coded from the police report. The latter is a smaller (~3,000 crashes per year) sample of light-vehicle towaway crashes that are investigated by trained accident investigators. One key advantage of CDS is that it includes estimated crash severity, or delta-V, assigned to each vehicle in a crash. Delta-V is the best single predictor of injury outcome and thus is a key variable in prediction models. However, because it requires special data collection, it is absent from police-report-based datasets. In contrast to CDS, GES is a much larger sample and includes more than just light-vehicle crashes. Thus, it would be ideal to have delta-V available for GES crashes to improve models of injury outcome from those data. Some attempts to do this for individual crashes were unsuccessful (e.g., Farmer, 2003). However, in some analyses, especially statistical simulations, it is sufficient to have a distribution of delta-V for a given crash mode rather than tying a specific delta-V to a specific crash-involved vehicle. This report presents a method of estimating a distribution of crash severity using only police-reported crash data. The approach uses an injury risk curve developed from CDS and a parametric distributional assumption for the delta-V distribution. While the distribution can take any parametric form, I use the lognormal in this report. The method uses maximum likelihood to fit parameters to the delta-V distribution based on the observed injury distribution using the police-reported KABCO scale. That is, individuals in crashes fall into one of five injury categories. Each pair of lognormal parameters produces a distribution of injury when multiplied by the injury risk curve. Thus, the parameters that produce the multinomial injury distribution that best fits the observed distribution are chosen for the estimated delta-V distribution. The report includes results from a simulation study as well as a fit to CDS data with known delta-V distribution.en_US
dc.description.sponsorshipNational Highway Traffic Safety Administrationen_US
dc.format.extent16en_US
dc.languageEnglishen_US
dc.publisherUniversity of Michigan, Ann Arbor, Transportation Research Instituteen_US
dc.subject.otherStatistical analysisen_US
dc.subject.otherInjury severityen_US
dc.subject.otherVelocity measurementen_US
dc.subject.otherTraffic crashesen_US
dc.titleA method for estimating delta-V distributions from injury outcomes in crashesen_US
dc.typeTechnical Reporten_US
dc.subject.hlbsecondlevelTransportation
dc.subject.hlbtoplevelEngineering
dc.description.bitstreamurlhttp://deepblue.lib.umich.edu/bitstream/2027.42/117575/1/103241.pdf
dc.description.filedescriptionDescription of 103241.pdf : Final report
dc.owningcollnameTransportation Research Institute (UMTRI)


Files in this item

Show simple item record

Remediation of Harmful Language

The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.

Accessibility

If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.