Masters Theses


"This research concentrated on two major aspects of highway safety analysis, hotspot identification and identification of crash contributing factors on highways. The objective of the first part of this thesis is to evaluate two crash frequency, three crash severity and two composite hotspot ranking methods. Based on the limitations of the existing hotspot identification methods (HSID), crash factor method (CFM), a crash severity based HSID method is proposed. A composite rank method (F 1) based on principal component analysis is also proposed. These methods are evaluated using three tests, the site consistency test, the method consistency test, and the total rank difference test. The evaluation is carried out in detail for more than top 1-10% hotspots. From the results of evaluation, the empirical Bayes, the crash factor, and Fl methods performed better than other methods. These methods are, therefore, recommended for HSID.

The analysis of highway crash data for a county as a sample to identify common crash contributing factors for several counties categorized by taking effects of spatial autocorrelation (SA) into account is the focus of the second part of the thesis. SA was determined using Moran's I and Getis-Ord Gi* statistic. The crash contributing factors are identified using the multinomial logistic regression (MLR) technique. Three factors were found to influence the identification of common crash contributing factors: i) GetisOrd Gi* statistics, ii) percentage of data represented by the selected county, and iii) the crash frequency for different levels of crash severity for the selected county. The MLR results indicated the following factors have positive association with crash severity: rural highways, roads on grades and curves, driving under the influence of alcohol, head-on and sideswipe same direction collisions"--Abstract, page iv.


Bham, Ghulam

Committee Member(s)

Samaranayake, V. A.
Lord, Dominique


Civil, Architectural and Environmental Engineering

Degree Name

M.S. in Civil Engineering


Mack-Blackwell National Rural Transportation Study Center (U.S.)
National Science Foundation (U.S.)


Missouri University of Science and Technology

Publication Date


Journal article titles appearing in thesis/dissertation

  • An evaluation of crash frequency, crash severity, and composite rank methods for hotspot identification
  • Identification of crash contributing factors: effect of spatial autocorrelation and sample data


ix, 75 pages

Note about bibliography

Includes bibliographical references.

Geographic Coverage



© 2011 Uday Raghavender Rao Manepalli, All rights reserved.

Document Type

Thesis - Open Access

File Type




Subject Headings

Traffic safety -- Statistical methods
Traffic safety -- Research -- Statistical methods
Traffic safety -- Arkansas
Traffic accidents -- Arkansas

Thesis Number

T 10277

Print OCLC #


Electronic OCLC #



Thesis Location