Keywords and Phrases
Autocorrelation; Autoregressive; Crash Frequency Model; Generalized Estimation Equation; Longitudinal Analysis; Temporal Correlation
"This study comprises mainly of three papers. First, a systematic evaluation of the effects of Missouri's Strategic Highway Safety Plan between 2004 and 2007 is presented. Negative binomial regression models were developed for the before-through-change conditions for the various collision types and crash severities. The models were used to predict the expected number of crashes assuming with and without the implementation of MSHSP. This procedure estimated significant reductions of 10% in crashes frequency and a 30% reduction for fatal crashes. Reductions in the number of different collision types were estimated to be 18-37%. The results suggest that the MSHSP was successful in decreasing fatalities.
Second, ten years (2002 - 2011) of Missouri Interstate highway crash data was utilized to develop a longitudinal negative binomial model using generalized estimating equation (GEE) procedure. This model incorporated the temporal correlations in crash frequency data was compared to the more traditional NB model and was found to be superior. The GEE model does not underestimate the variance in the coefficient estimates, and provides more accurate and less biased estimates. Furthermore, the autoregressive correlation structure used for the temporal correlation of the data was found to be an appropriate structure for longitudinal type of data used in this study.
Third, this study developed another longitudinal negative binomial model that takes into account the seasonal effects of crash causality factors using Missouri crash data. A GEE with autoregressive correlation structure was used again for model estimation. The results improves the understanding of seasonality and whether the magnitude and/or type of various effects are different according to climatic changes. It was found that the traffic volume has a higher effect in increasing the crash occurrence in spring and lower effect in winter, compared to fall season. The crash reducing effect of better pavements was found to be highest in spring season followed by summer and winter, compared to the fall season. The results suggest that winter season has the highest effect in increasing crash occurrences followed by summer and spring"--Abstract, page v.
Samaranayake, V. A.
Civil, Architectural and Environmental Engineering
Ph. D. in Civil Engineering
Missouri University of Science and Technology
Journal article titles appearing in thesis/dissertation
- Safety effect of Missouri's strategic highway safety plan - Missouri's blueprint for safer roadways
- Crash frequency modeling using negative binomial models: An application of generalized estimating equation to longitudinal data
- Seasonal effects of crash contributing factors on highway safety
x, 143 pages
© 2014 Mojtaba Ale Mohammadi, All rights reserved.
Dissertation - Open Access
Traffic safety -- Research -- Missouri
Traffic safety -- Missouri -- Statistical methods
Traffic accidents -- Missouri -- Statistical methods
Electronic OCLC #
Ale Mohammadi, Mojtaba, "Longitudinal analysis of crash frequency data" (2014). Doctoral Dissertations. 2498.