Search This Website
 Understanding Testing, Standards Development

Behind the Consumers Union Infant Car Seats Fiasco

CU first published and then—two weeks later—retracted its test results of 12 rear-facing child restraints that had been widely publicized on January 5 and printed in the February 2007 edition of Consumer Reports. The report drew considerable media attention and led to panic on the part of many parents, who snatched up the only two that were recommended or, by some reports, tried to return other models to retailers.

Several expert advisors to Safe Ride News suspected from the very beginning that the side-impact testing was not performed in the usual manner due to the unexpectedly disastrous test results. CU refused several requests for basic information about their side-impact test procedures.

On January 18th, CU withdrew its article, stating that they had learned two days earlier about the way in which NHTSA conducts its side-impact tests.

On January 19th, NHTSA put on its web site the video of its own tests of 11 of the 12 seats in the CU test program. The NHTSA tests properly simulated the NCAP side-impact tests and apparently were used to convince CU its test procedures were wrong. (See “Explanation of the CU Side-impact Test Error,” page 4.) NHTSA also posted a video of what would happen in a crash to a baby held in an adult’s arms.

Injury problem in the real world

Even if it had run the side-impact tests correctly, CU made no effort to put the higher speed tests in perspective. In reality, current and correctly used RF CRs are estimated to be 96% effective at reducing fatalities compared to non-restraint. The speed required under the current federal standard, FMVSS 213, is 30 mph in a frontal test. The resulting change in velocity is more severe than 97.6% of real world frontal crashes. A portion of the remaining crashes is completely unsurvivable.

Some child fatalities occur in side-impacts. Protection from this type of crash is largely related to the vehicle structure, as most fatalities are related to massive intrusion. However, tests for child restraints have been under development internationally for some time, and NHTSA has been collecting supporting data through an Advanced Notice of Proposed Rulemaking. This is an area being worked on in a scientific manner so appropriate interventions can be determined and regulated.

Other points in the article itself

The CU frontal impact crashes probably were done correctly, since a frontal crash test is easy to simulate on a sled. However, although it lists various types of failures, the article provides very little hard information about how the seats fared in this test mode.

Other than the major testing error, the main flaw in the article was the lack of specific data that would allow readers to adequately assess CU’s conclusions. Consumer’s Union does not submit its research for peer review, as is required for scientific papers. In addition, and far worse, it does not publish the actual detailed test results. Therefore, it is extremely difficult to understand the findings in order to use, support, or take issue with the conclusions.

CU made some good points in the article but was by no means the first to raise them. The comparison test of the British CR showed how well rigid ISOFIX lower attachments performed, yet the rigid ISOFIX was not mentioned in the article as one reason for its good performance. The improved performance, particularly in side-impacts, of CRs with rigid attachments is known. The Britax USA version, BabySafe, and the Baby Trend LATCH-Lok were not mentioned. (The issue of flexible vs. rigid lower LATCH attachments on U.S. models has a long and complicated history.) Other valid points include: the lack of lower LATCH anchors for use in the center rear of most vehicles; the difficulty of installing the old version of the base of the Dorel/Eddie Bauer Comfort (since improved); and the fact that current compliance tests do not allow the use of additional installation features such as a tether for a rear-facing CR or a stabilization foot.

Regarding the Evenflo Discovery test failure in the CU 30-mph test, Evenflo and NHTSA both state unequivocally that it passed actual compliance tests. CU had found a number of complaints to NHTSA and Evenflo regarding separation from the base, which apparently included six fatalities, but NHTSA’s Office of Defects Investigation had investigated and had been unable to find a defect or non-compliance, so it had closed the investigation.
Scant information was provided in the article about how the LATCH systems on the various CRs failed. SRN was told verbally that there were all sorts of failures, from base-shell separations to strap failure or shell damage at the strap path. The details are essential to understanding whether the LATCH system failures were significant and how it might be improved.


The forthcoming revised article on infant restraints will be of considerable interest. In addition, CU says it will test convertibles and boosters in the same manner in the near future, so CPSTs and advocates should learn all they can now about side-impact testing to be prepared.


Sherwood, CP, Ferguson, SA, Crandall, JR, Factors leading to crash fatalities to children in child restraints, AAAM Proceedings, 2003.

© Safe Ride News, January/February 2007


Safe Ride News Publications, P.O. Box 38, Edmonds, WA 98020-0038
Phone: 425-640-5710 / 800-403-1424 • Fax: 425-640-5417
All rights reserved, by copyright E-mail: Webmaster