Page 1 of 1

Possible Part B Ch 4 Error - Crash Types Exceeding Threshold

PostPosted: Tue Aug 05, 2014 1:03 pm
by benpdx

I've been using the Excess Proportion of Specific Crash Types metric for a network screening project. I believe that there is a typo on page 4-54 in Equation 4-20 where the sample variance is calculated for the subcategories:

Equation 4-20: Sample Variance
Screenshot_080514_115554_AM.jpg (221.75 KiB) Viewed 4240 times

Based on the discussion surrounding this section, the formula seems like it should be "for N observed,i >= 2" instead of "for N observed,i (total) >=2"

I believe the sample variance should only include locations where there were two or more of the target crash type. As written, it includes all locations where there are two or more of any crash type.

I appreciate any clarification on this point.

Thank you,