The Kansas City Star analyzed about 1.9 million computer records of the National Highway Traffic Safety Administration, the federal agency responsible for motorists’ safety.
Information about traffic deaths came from the national Fatality Analysis Reporting System, which counts deaths that occur within 30 days of accidents. FARS contains data gathered from law enforcement reports and other sources.
The newspaper examined fatal head-on crashes from 2001 through 2006 in which airbags did not deploy.
In an effort to be as conservative as possible, the newspaper’s analysis focused on head-on crashes into the front ends of other vehicles and objects such as trees and embankments.
The Star did not include victims identified in the database as having been killed in crashes to the left or right fenders of their vehicles.
Nor did it include victims who were ejected or died when their vehicles rolled over, caught fire or were submerged in water after crashes. And the newspaper excluded victims in vehicle models where airbags were not standard. To verify that on specific models, The Star checked Ward’s Automotive Yearbook and Automotive News.
These steps eliminated at least 3,000 fatalities.
That left at least 1,400 victims that The Star used in its analysis. If the newspaper includes front-end crashes into the side or rear of other vehicles, the number climbs to at least 1,900.
The newspaper consulted with automotive safety researchers, statisticians and other experts — including those who have worked for NHTSA — in formulating its methodology for the analysis. All found it acceptable.
A NHTSA spokesman said that the newspaper could not use its database to reach conclusions about airbag deployments because some of the information about crash speeds it receives from law enforcement agencies is insufficient or inaccurate.
The Insurance Institute for Highway Safety, which is funded by the insurance industry, also uses the data for research. “FARS data is very reliable,” said spokesman Russ Rader. But Rader later noted that the institute does not use the database for airbag studies because it sometimes contains errors.
However, Institute President Adrian Lund said the database is the best available and getting better because of improvements in quality control. Lund said he was not surprised by the newspaper’s findings, because the institute’s limited studies on other databases also had found instances where people died in frontal crashes where airbags didn’t deploy.
“It is something that would have to be studied further to be sure of what’s going on,” he said.
In fatalities, law enforcement takes extra care in producing accurate accident reports because of the heightened potential for future litigation, said Dennis Hallion, chairman of the National Troopers Coalition.
Regarding nondeployments in fatal wrecks, “I’d say the (FARS) numbers are very accurate on that: Did the bag deploy or did it not? From a law enforcement standpoint, that’s an easy determination. … It’s one of the first things you look at.”
Rather than disregard The Star’s findings, Hallion added, federal regulators “need to … validate their own numbers. It’s their database.”
Although NHTSA said it does not do a scientific random analysis of its database, it does take some steps to check the data and is confident of their accuracy.
“Quality control is a vital system feature. One important part of the quality control program is a series of consistency checks, which ensure that no inconsistent data are entered,” NHTSA says on its Web site. “Statistical control charts are also employed to monitor the coding of key data elements.”
Join the discussion
Share your observations and experiences about news. Lively, open debate is the goal, but please refrain from personal attacks or comments that are racist, vulgar or otherwise inappropriate. If you see an inappropriate comment, please click the "Report as violation" link to notify a KansasCity.com editor. Thanks for your feedback.