Geography Jobs First. In Canada

Job Seeker

My GeographyJobs Login


Featured Article

Introduction to Highway Safety Analysis – Part 3

By Matt Hiland
August 31, 2009

This article is the third in a series that provides a high-level look at the typical data and processes for highway safety analysis.

Because of their inherent differences in conditions, analyses of intersections and non-intersection sections of roadways are handled separately. Also, before performing analyses of the crashes that occurred on roadway sections or intersections, the sections and intersections must be classified into groups of similar locations. For example, it would be invalid to compare the number of crashes on a two-lane rural state highway with the number of crashes on a major urban expressway because they are very different in terms of traffic volume and road configuration.

Classification of intersections and sections into comparable groups can vary widely according to the available data and the professional opinions of the investigators. Some of the attributes typically used for classifying roadway sections include but are not limited to:

  • Annual Average Daily Traffic (AADT)
  • Functional Classification
  • Number of Lanes
  • Median Type and Width
  • Shoulder Type and Width

Some of the attributes typically used for classifying intersections include but are not limited to:

  • Annual Average Daily Traffic (AADT)
  • Functional Classification
  • Number of Through Lanes
  • Presence and Type of Traffic Control Devices
  • Presence of Dedicated Turning Lanes

After crashes have been accurately located on roadway sections or at intersections and the roadway sections and intersections have been classified into comparable groups, the analyses mostly fall into the following categories. Michael Pawlovich of Iowa DOT provides an excellent and more technical summary of analysis methods at

  • Crash Frequency/Density – Over a given time period, how many crashes occurred at a location or on a section of roadway? This method is one of the simpler analysis methods and is often used to generate an initial list of locations to be evaluated using other methods. Because it typically does not incorporate traffic volume into the calculation, this method can provide misleading results if not interpreted carefully.
  • Crash Rate – Over a given time period, how many crashes occurred per vehicle mile travelled? This method incorporates traffic volume and provides a more directly comparable value for similar locations. Variations of this method look at similar locations across a state and identify those that vary significantly from the expected rate.
  • Crash Severity – These methods calculate crash frequencies or rates, but weigh crashes according to their severity. This helps identify locations that may have fewer but more serious crashes, the mitigation of which would make greater overall safety improvements. However, if there are few total crashes, a single fatal crash could skew the results over other locations with many non-injury crashes.
  • Over-Representation of Crash Types – This method looks for deviations from the normal distribution of crash types (rear-end, side-swipe, etc.) occurring at comparable locations. This method can often give direct clues to potential engineering corrections to improve safety. For example, an over-representation of side impact crashes at an intersection with stop signs on the side roads and no traffic control on the main road may indicate that visibility is limited from the side roads or that speed reduction is needed on the main road.

These analyses are only the starting point of the highway safety improvement process. After identifying the locations that have the most, most frequent, or most severe crashes, highway safety engineers will review the results and the common attributes of the crashes and locations to identify possible ways to improve their safety. These projects become part of the state’s Highway Safety Improvement Plan. Some systems are starting to provide assistance to engineers by suggesting countermeasures for particular types of crash conditions and by providing tools to estimate cost-benefit ratios for selected projects.

Law enforcement agencies can also use these analyses to help assign officers and plan targeted enforcement activities. Emergency response agencies can use these data to help reduce response times to crash-prone locations. These data can also be used to help plan and execute public information and education campaigns to draw attention to specific behaviors that impact highway safety. Other ongoing research is looking into ways to incorporate additional data such as locations of specific types of businesses, demographics, etc. Given the ongoing interest and availability of funding, highway safety analysis and transportation planning for safety improvement seem to be good focus areas for researchers and job seekers and should remain so for some time.

Additional Information