Developing a Procedure for Estimating the Required Sample Size for Safety Performance Function Calibration Factors

The predictive procedures included in the Highway Safety Manual (HSM) are known as safety performance functions (SPFs) and are based on observed crashes for a select number of locations in the United States. For a jurisdiction to be able to fully benefit from applying the SPFs to their region, it is necessary to calibrate these models for local conditions, being a state or one of its regions. Currently, existing manuals recommend a "one-size-fits" all sample size for calibration procedures that require crash data from 30 to 50 locations with a minimum number of approximately 100 crashes per year. This recommended sample size is not fully supported by documented research and several agencies that have initiated SPF calibration efforts indicate that this sample size is impossible to obtain due to infrequent crash conditions associated with certain facility types. The objectives of the proposed research are to (1) review and document issues with the existing calibrating method in the HSM, (2) develop a procedure for the calibration of crash count and severity-level models developed using a discrete choice model, (3) identify factors that influence the calibration procedures and the selection of the sample size, and (4) determine how frequently an agency should update their calibration factors. The study objectives will be accomplished using simulated and observed data. The anticipated guidelines will include a discussion on (1) the minimum sample size, (2) the characteristics of the data used for the calibration procedure, (3) potential issues with the calibration procedure, and (4) when or how often models need to be re-calibrated.


  • English


  • Status: Completed
  • Contract Numbers:



  • Sponsor Organizations:

    Department of Transportation

    Office of the Assistant Secretary for Research and Technology
    1200 New Jersey Avenue, SE
    Washington, DC  United States  20590
  • Performing Organizations:

    Texas Transportation Institute

    Texas A&M University System
    3135 TAMU
    College Station, TX  United States  77843-3135
  • Principal Investigators:

    Geedipally, Srinivas

    Lord, Dominique

  • Start Date: 20141001
  • Expected Completion Date: 20151218
  • Actual Completion Date: 20160218
  • Source Data: RiP Project 37101

Subject/Index Terms

Filing Info

  • Accession Number: 01536419
  • Record Type: Research project
  • Source Agency: Center for Advancing Transportation Leadership and Safety (ATLAS Center)
  • Contract Numbers: DTRT13-G-UTC54, 602761-00014
  • Files: UTC, RiP
  • Created Date: Aug 31 2014 1:00AM