Page 1 of 1

NHTSA database includes reported negative crush?

Posted: Thu Jul 23, 2009 1:36 pm
by MSI
Q: In looking at the NHTSA crash test database site, I've noted several instances where one of the crush measurements was a negative value of significant magnitude. Can anyone explain why they report the value in this manner. All crashes are 30 mph or 35 mph rigid barrier impacts and in reviewing photographs, the crush profiles seem consistent across the front of the vehicles. All tests were done at different facilities, hence I'm thinking there is a reason for reporting it this way.

A: The question demonstrates some database reporting errors. When you compare the reported values with the values in the full NHTSA report, you will see the error in the reporting process/transcription to the database.
The question emphasizes the need for a careful review of the full report(s) from the NHTSA site (You should use more than 1 test/data point!) NHTSA has a link to the full report for most of the tests. You'll need the actual test number. In the NHTSA reporting there is also listed overall length pre- and post- impact which should provide a clue that the data may be in error.
It is convenient to use 3rd party software vendors for the NHTSA data but that doesn't excuse your duty to take the time to review the actual NHTSA reports and do the calculations independently to verify the veracity of the underlying data and calculation procedures.
There has been waffling on the techniques used to estimate the effect of `air gap' and it boils down to "WAGs gone wild". The whole `air gap' hocus pocus demonstrates the need for looking at more that one test/data point.
Greg Anderson of Scalia Engineering has done an outstanding job of preparing the Sisters/Clone database which can assist you in finding additional data points.
McHenry Software with our medit32/msmac produce includes the NHTSA crash test database, the Sister/Clone Database and direct links to the individual NHTSA crash tests
Please also see:
medit32 databases
NHTSA crash test database site

Re: NHTSA database includes reported negative crush?

Posted: Thu Jul 23, 2009 1:56 pm
by brian
From another news group there was posted the following:
The NHTSA database has experienced a "glitch" where it has replaced the DPD4 (Damage Profile Distance 4) measurement with the value of -581 in a large number (all???) of tests. Two examples of this are tests 3438 and 6357, where the previous values of 299 and 278 have both been replaced by -581.
Yet another reason to be sure to check, and double check, your crash data test points with the full NHTSA report.