WHITE PLAINS, N.Y. -- Consultants hired by Consumer Reports to investigate how it botched a story about infant car seats concluded Tuesday that a major misunderstanding between the magazine and the lab that conducted the test resulted in the error.
The findings of the test -- that most seats "failed disastrously" -- were withdrawn two weeks after their Jan. 4 publication when the magazine learned its side-impact tests had simulated speeds twice as fast as it reported. The error prompted criticism from the manufacturers involved and confusion among readers, especially parents of young children.
The results of the magazine's investigation were released Tuesday and will be published in the May issue, among reports on Teflon and lawn-care products. The issue also includes a pledge from Jim Guest, president of parent company Consumers Union, that in the future he will have to sign off on "any report that calls a product Not Acceptable or raises questions about an entire group of products."
According to the report, Consumer Reports wanted to test the effect of a 38 mph side impact on children in car seats. The magazine chose that speed because that is how cars themselves are tested by federal regulators.
When such a crash occurs, much of the momentum of the striking car is absorbed by the struck car and the struck car moves away at about half the impact speed. But the lab tested the car seats as if they moved off at 38 mph, which would have been the result of a much more violent crash, the probe concluded.
Consumer Reports then presented the findings as the results of a 38 mph impact and said only two of 12 seats tested were worth buying.
The experts -- Kennerly H. Digges, former director of vehicle safety research at the National Highway Traffic Safety Administration, and Brian O'Neill, former president of the Insurance Institute for Highway Safety -- said the misunderstanding persisted throughout the testing because the test was devised with little input from outside consultants.
Consumer Reports said that reflected "the organization's long-standing policy of limiting contact with government and industry to avoid compromising the independence of its judgment."
In an interview Tuesday, O'Neill said he and Digges recommended that Yonkers-based Consumers Union "be much more open when it is devising new tests."
"During the test development process, CU needs to make sure that it's got consultations with the relevant experts outside the organization because almost by definition their staff cannot know everything because they test so many different products," O'Neill said.
CU's staff "assumed these tests that they were trying to run were simple and straightforward when in reality they were very complicated," he said.
In his message to readers, Guest said the magazine would indeed consult more with outside experts when devising tests; would check results when using outside labs and better inform readers when an outside lab is used; and would double-check any findings that seem to go against real-world experience.
The magazine took full responsibility for the error, not blaming or even naming the lab involved, Buffalo-based Calspan. Calspan spokeswoman Lissa Carroll refused to comment.
Guest said in an interview that Consumer Reports, which refuses advertising and does not permit use of its reviews in others' advertising, would not lose any of its independence as a result of the error and its fallout.
"We can consult at the front end when we're developing our protocol but we don't have to share what our final protocols are and we don't have to share our results ahead of time," he said. "We'll continue to be vigilant about that."
O'Neill said he found no evidence that cast doubts on Consumer Reports' tests of any other products.
He said the magazine is retesting car seats and will do a new report, but that report will describe front-end impacts only because proper side-impact testing is still being developed.
"In the meantime, the key advice is that any car seat is better than no seat," the magazine said.
On the Net: http://www.consumerreports.org