Local Calibration/EB Method

Safety performance functions, development (calibration, crash modification factors [Part C], crash proportions), IHSDM and other software

Moderator: khardy

Local Calibration/EB Method

Postby jmarandino » Tue Apr 26, 2011 9:53 am

Am I correct to assume the Empiral Bayes Method can not be applied without having a local calibration factor?
Posts: 3
Joined: Fri Apr 15, 2011 9:45 am
Location: Vineland, NJ

Re: Local Calibration/EB Method

Postby Mohamad Banihashemi » Tue Apr 26, 2011 12:11 pm

In general, the predicted number of crashes is valid only if you have calibrated the prediction models for your area. However, if you have not done the calibration, the relative comparison of two or more alternatives to your project still could be valid to certain extend. I say certain extend because in a single project you may use different models (because of having different segments and intersection types) and the calibration factors for the same area but for different models might be very different from each other, and even on different sides of 1.00 (greater or smaller). So, if there is only one model (one type of highway segment or just one type of intersection) the relative comparison is pretty valid but if there are several models used in the project, even the relative validity is under question. I believe this is true for both with and without-EB evaluations.
Mohamad Banihashemi
Posts: 17
Joined: Thu Dec 02, 2010 11:49 am

Re: Local Calibration/EB Method

Postby lubhead » Wed Jul 13, 2011 11:50 am

I have a similarly related question. If a jurisdiction specific calibration factor is calculated it is unclear in Appendix A of Part C how/when that calibration factor is to be applied when using the EB method to predict future crashes. My assumption would be that it should be included in equation A-15 somewhere. How was this addressed in the IHSDM?

Posts: 3
Joined: Wed Feb 09, 2011 1:40 pm

Re: Local Calibration/EB Method

Postby mdimaiuta » Thu Jul 21, 2011 2:37 pm


HSM Figure C-2 is helpful in seeing where in the process the calibration factor is applied, in relation to the EB method. Steps 9-11 (Select and apply SPF; Apply CMFs; and Apply a calibration factor) are conducted before "Apply site-specific EB method (if applicable)" in Step 13 or "Apply project-level EB method (if applicable)" in Step 15. The description of Step 13 on page C-11 includes the instruction to, "Apply the site-specific EB method to a future time period, if appropriate."

In the IHSDM Crash Prediction Module, when the user selects an Empirical Bayes analysis, he/she must identify the "Highway with Crash History" (i.e., the "past time period" highway in eqn. A-15). The highway that the user is evaluating is the "future time period" highway in eqn. A-15.

Posts: 29
Joined: Tue Nov 30, 2010 1:50 pm

Return to Part C - Predictive Methods

Who is online

Users browsing this forum: No registered users and 1 guest