No abstract

In Part C, the HSM presents a predictive crash analysis method that is based on a series of statistical models called "safety performance functions" (SPFs). SPFs are regression models developed from observed crash data that predict the average crash frequency of a roadway on the basis of a set of geometric and traffic conditions. Annual average daily traffic (AADT), segment length, or both serve as explanatory variables, where applicable. The predictive power of SPFs can provide extremely useful information for departments of transportation (DOTs) and other agencies for purposes of funding, project planning, and other applications. However, the HSM SPFs were developed with data from select states across the United States; thus these predictions are not always accurate for all jurisdictions. With the calibration of the HSM models, the SPFs can be adjusted to better predict safety performance for these specific areas.The Ohio DOT desired to incorporate the HSM procedures into its project development process, beginning initially with its Highway Safety Improvement Program and design exception processes. To facilitate these actions, it was necessary to calibrate the provided SPFs for all roadway types currently covered in the HSM for the state of Ohio or develop jurisdiction-specific SPFs. Because developing unique, agency-specific SPFs is a labor-intensive and more difficult task, the Ohio DOT elected first to develop calibration factors and test the fit of the adjusted SPFs to assess whether the calibration factors were sufficiently reliable to be used for predictive analyses. If not, Ohio-specific SPFs could be developed for select roadway facility types that were identified for improvement. Calibration EffortsWhile more states and jurisdictions are beginning to look into developing agency-specific information, whether for calibration factors or full SPFs, there are not many studies that consider evaluating calibration factors before developing SPFs. Shortly after the release of the HSM, Xie et al. calibrated the HSM models for the state of Oregon with the method outlined in the HSM (1). The calibration calculations resulted in factors consistently below the expected value of 1.0, an outcome that was likely attributed to Oregon's property damage only crash self-reporting procedures and its higher property damage only reporting thresholds compared with the states that were used in developing the HSM crash proportions (1).Comparing the fit of calibrated and uncalibrated models is a common approach across multiple agencies around the United States. Srinivasan et al. compared the mean error of the precalibration results with the postcalibration results for the Florida DOT (2). The study also created agency-specific SPFs and compared them with the calibrated HSM SPFs. The findings indicated that the calibrated The AASHTO Highway Safety Manual (HSM) provides methods for departments of transportation (DOTs) and other agencies to incorporate high-quality quantitative safety analyses into project development and decision m...

Ohio faces the challenge, as do many other states, of how to utilize Highway Safety Improvement Program (HSIP) funding to improve safety on its low-volume roadways while meeting the data-driven safety funding requirements of the Fixing America’s Surface Transportation (FAST) Act. Low-volume roads present unique challenges because data is rarely available and other factors, such as roadway ownership, affect the implementation of safety countermeasures on this system. Beginning in 2014, the Ohio Department of Transportation (ODOT) created a township safety signage grant program designed to address the issues with utilizing HSIP funding on low-volume roadways. The grant program’s goal was to drive down the number of fatalities, serious injuries, and overall crashes occurring on Ohio’s low-volume roads. ODOT took into consideration overriding issues regarding low-volume roads in how it structured the grant program. ODOT also utilized the direction from its Strategic Highway Safety Plan in choosing a safety countermeasure which met the needs of its roadway departure and intersection crash trends. The program has been actively engaged in by Ohio townships and now has a large enough amount of post-safety countermeasure installation data available to quantify its initial success. This paper presents the successful results by highlighting the human capital and comprehensive societal benefit/cost analyses for the first 24 townships with 12 months of post-grant completion crash data.