Bicycle Helmet Safety Institute

Consumer-funded, volunteer staff

Helmets Children Promotions Statistics Search

Report on 1997 Consumer Reports Article

Summary: Consumer Reports tested bicycle helmets in 1997 for the first time. See this page for more recent helmet articles they have published.

Consumer Reports Publishes
Helmet Article

The June, 1997 issue of Consumer Reports had an article on bicycle helmets.

Consumers Union (CU) tested 24 helmets this time by increasingly higher drop heights until the best continued to protect while the others failed. This is in contrast to their 1994 article, where they chose one drop height and rated helmets by the softest landing. We liked that approach in 1994 because we suspect that the foam in today's helmets may be too stiff, and the softest landing would help to avoid concussions and other less-than-fatal brain injuries. Those apparently occur more often than previously realized. In 1997 Bell began using ASTM rather than Snell certification, so an important question was whether or not their helmets were still capable of performing at higher drop heights than ASTM's. That question was answered when Bell helmets placed second and third in the CU ratings. The Giro Helios (a very expensive helmet) took first place, with the Bell Evo2 Pro (a helmet we don't like because it lacks an outer shell) and a Bell Psycho Pro. The Pro-Action Illusion (which we think is a revision of the Troxel Diva, equipped with pony tail port) was fourth and rated as the only Best Buy.

CU did not attempt to test the dozens of brands on the market, settling for only the biggest sellers. They tested 10 adult helmets, of which seven were Bell or Giro models, one Pro Action (made by Troxel), one Specialized and one Rollerblade. There was somewhat more variety in the youth models, but Bell and Giro accounted for five of the ten. The top youth model was the Specialized Air Wave Mega, followed by the Bell Warped, Bell Maniac Pro and Trek Navigator.

The attention-getting headline in the article concerned buckles. CU broke some, and urges consumers to avoid them. They showed up on 1997 Bell, Giro and Pro-Action models. Only some of the tested buckles on those models broke, and CU did not consider the problem acute enough to warrant downgrading the helmets, which is always a significant distinction in their ratings. If you have a 1997 Bell, Giro, Pro-Action or Rollerblade helmet and it has a buckle labeled "ITW Nexus TSK63" they recommend you replace it. Some production batches of the same helmets had different buckles in 1996, and those buckles did not fail. CU also recommended replacing helmets with "Pinchguard" buckles, found on Bell and BSI toddler helmets.

Bell is currently trying to determine how the tests could have turned up defective buckles, since they say that their own thorough quality control testing, and that of third party labs they use for confirmation, did not. Meantime, there are no reports of the buckles breaking in the field, and it may be possible that the CU testing was somehow more severe than the ASTM and Snell standards. Then again, it could be a case of some defective buckles in certain production lots. CPSC is investigating, and hopefully will provide the definitive conclusion. (Footnote: they did, sort of. See below.)

The Snell Foundation issued a press release saying their testing shows no buckle problems with the Giro models they have certified. Harborview Injury Prevention Center's Dr. Frederick Rivara issued a press release saying the article may damage public confidence in helmets. The Protective Headgear Manufacturers Association wrote a letter to CU protesting the article's emphasis on the buckle failures and questioning the validity of the test results. CU responded in the form of an open letter.

The article could in fact put a damper on some decisions to buy a helmet, particularly for those who would buy in discount stores where there is no sales help knowledgeable enough to know one buckle from another. Stay tuned. In any event, the buckles are not breaking easily, and your chief concern with any plastic buckle should be that there are no pieces broken off, and that the prongs are indeed snapping tightly into place every time you use it. We think that is why CU did not consider the problem serious enough to require downgrading the helmets. But if the buckle problem was not worth downgrading the helmets, it probably was also not worth as much emphasis as CU put on it, given the likelihood of spooking consumers.

We wish CU had tested more helmets, and particularly more brands. There are lots of cheap helmets out there that meet the stringent Snell B-95 standard, and not many people who want to shell out for the Giro Helios or one of Bell's more expensive models. We also wish they had included the "softest landing" information, as they did in 1994.

In addition to impact protection, the article rates helmets for ventilation, rolloff resistance (secure fit), ease of adjustment and strap creep (loosening of straps by slippage over time). Of the tested helmets the best vented one was the Giro Helios, and the second-ranked ones were all Giro and Bell models. The Pro-Action Illusion was rated a best buy for its combination of impact performance, stability and price, not to mention the extra space for your pony tail.

As always, we are delighted to have CU publish another helmet article. Whatever your views on their conclusions, they are the only institution in the U.S. testing helmets thoroughly and publishing actual crash lab test results. We hope the buckle question can be settled quickly. Somebody is going to have a lot of explaining to do!

Added on September 1, 1997

The Consumer Product Safety Commission has just released to us, following a Freedom of Information Act request, their results from buckle testing they did in June. They were reacting to the Consumer Reports article.

CPSC tested two helmets from Troxel, ten from Giro, four from Rollerblade and 12 from Bell Sports, for a total sample of 28. Most of them had the ITW Nexus TSK63 buckles that Consumer Reports broke, but it is not clear from CPSC's lab data if all 26 buckles were actually identical. Two had Bell's unique "Pinchguard" buckle for toddlers. In short: The two Giros broke when CPSC tested them to the ASTM standard, but neither helmet was certified to the ASTM standard to begin with. Both were certified to Snell and ANSI. All Giro samples passed the Snell and ANSI tests. So from Giro's point of view, the buckles did everything Giro ever said they would do. CPSC seems to have agreed, since they have dropped the matter rather than seek any action from the manufacturers. We have a page with Giro's comments and CPSC's letters to the manufacturers, obtained by the Freedom of Information Act request.

We have no reports that buckles that pass Snell and ANSI break in actual field crashes, so even if they can't pass ASTM, (supposed to be roughly equivalent to Snell, but apparently more difficult to pass--see our Standards Comparison for details) we conclude that the buckles that broke in CU and CPSC labs were probably adequate in strength. We would not have any real reservation about riding with them ourselves, and do not consider them cause for consumer alarm. We are, however, still disturbed by the earlier reports from Consumer Reports that their lab broke buckles on helmets certified to the ASTM standard with the ASTM test. Nobody has found anything wrong with their lab or testing, even though some dispute the results.

The Consumer Reports article concluded that their results did not even warrant downrating the helmets, so the consumer is left to conclude that the issue is probably not very important, and to wonder why Consumer Reports made such a point of it in their article, leaving readers with the impression that there was a real problem. Or to wonder if there is more that is not being made public about buckles or certain batches of buckles or a thin margin of strength in buckles or whatever. We do know that there are normally small variances between two labs testing the same product, although they are normally less than five per cent in impact testing.

Our best judgment at this point is that there is a problem, but it is a small problem, and can probably be cured by adding tiny amounts of plastic to the buckle designs to strengthen them and provide a more generous allowance for lab testing variance, plastic variance and molding problems. We are probably never going to have more definitive public data on the subject.

Added on November 17, 1997

Consumer reports has tested some more helmets for their December, 1997 issue. They selected a Louis Garneau model, the Globe, as a Best Buy. They also suggested that consumers avoid ITW Nexus TSK63 buckles with date codes of 8/96, 10/96, or 11/96, or ITW Pinchguard buckles with a date code of 9/96 and said "Due to the inconsistent nature of the failures, we can't be sure that the problem has been solved." Their bottom line: "If you already have a helmet with one of these buckles, keep wearing it. Any helmet is safer than no helmet."

Our own advice, worth exactly what you are paying for it: if you have a TSK63 buckle, check the date codes. If you find one of the ones identified above, you can decide whether to replace it or not. Consumer Reports did the testing, and they stopped short of recommending replacement, although they did recommend avoiding those buckles if you are buying a new helmet. The buckles all passed the Snell tests, indicating that they are not going to break in actual use, and failed only on the more severe ASTM test. But if you do have one of those date codes and the idea of it bothers you every time you use the helmet, it will probably be worth it to you to replace it even if you have to pay for the new buckle. We would, just because the lab tests indicate some marginal buckles in those batches, and some individual buckle in those batches may be even worse than the ones CU broke. Unfortunately, everyone in the industry wants to avoid admitting that a buckle improvement may be needed, so the replacement might be exactly the same as your old buckle, or could even be weaker. Until we have more publicly-available test data than we are getting at present you won't have any better basis for making your decision than the small samples reported on in this case.

Perhaps more important in terms of actual lives at stake is the information that a plastic buckle must be replaced if any part of it breaks off. Some riders break off one prong of the male section of their buckle and continue to use it, with the second prong alone holding a tenuous grip. That is a genuinely unsafe buckle. It is entirely inadequate for nearly any type of crash. It is much weaker than anything Consumer Reports tested and broke. Any rider who is using a buckle like that should contact their bike shop or the helmet manufacturer for an immediate replacement!

Added on November 21, 2006

In November of 2006 the ASTM F8.53 headgear subcommittee discovered that there were subtle buckle test equipment differences between labs that could affect results. As of the date below, the subcommittee is still working on a clarification of its F 1446 Test Methods to eliminate the differences. We suspect this was the root of the problem.

You can access complete Consumer Reports helmet articles by subscribing to their website.