01123113

Component

Police reported crash data are the basis for traffic safety research studies which have developed numerous countermeasures to treat deficiencies in highway design features, surface conditions, traffic control devices, and driving behavior. However, a significant number of crashes go unreported. Using underreported data tends to produce a biased picture of traffic safety and thus results in ineffective treatments. This paper presents a new approach to modeling underreported traffic crash data with a modified latent Poisson regression model. With recent advancements in crash modeling and Bayesian statistics, the parameter estimation is implemented within the Bayesian paradigm, using two Gibbs Samplers and a Metropolis-Hastings (M-H) algorithm. The methodology is empirically applied to investigate traffic crashes which occurred on Washington two-lane highways. The underreporting Poisson regression model results differ substantially from standard Poisson models that do not consider underreporting.

01120148

09-3192

English

15p

2009

Transportation Research Board 88th Annual Meeting

DVD

References (50) ; Tables (5)

Crash data; Poisson distributions; Safety; Traffic control devices; Traffic crashes; Traffic data; Traffic safety; Two lane highways

Underreporting (Traffic accident injuries)

Highways; Safety and Human Factors; I80: Accident Studies

Transportation Research Board Annual Meeting 2009 Paper #09-3192

TRIS, TRB

Jan 30 2009 7:36PM