Probability of Crash Type Exceeding Threshold

Network screening methods and data preparation, diagnostic and selection of countermeasures, economic and prioritization methods, HSIP and HSM, SafetyAnalyst and other software

Moderator: khardy

Probability of Crash Type Exceeding Threshold

Postby mdgiroux » Tue May 06, 2014 7:02 am

I did the sample variance calculation that is shown at the bottom of page 4-54 of the HSM. I obtained the same 0.037 answer. When I changed the observed number of angle crashes for TWSC #2 from 21 to 11, I then obtained a negative answer, -0.000006.

Is it normal that Equation (4-20) can produce negative numbers? I thought that the variance could not be negative.
Posts: 1
Joined: Mon May 05, 2014 9:32 am

Re: Probability of Crash Type Exceeding Threshold

Postby benpdx » Tue Aug 05, 2014 1:17 pm


I also had a similar problem, where many of my variances calculated to be small, negative numbers. My impression is that variance should not be negative. I would appreciate any guidance on the matter, as it was very frustrating.

I eventually "solved" it by using including all sites in the variance calculation, and not excluding the locations where fewer than two target crashes were observed. However, this does not feel right.

Posts: 7
Joined: Fri Aug 01, 2014 5:51 pm

Return to Part B - Roadway Safety Management Process

Who is online

Users browsing this forum: No registered users and 1 guest