Forthcoming news related to this webpage

will be posted at VETA Homepage
.... new pages also at VETA web-guide (NEW in Dec. 1996)

Under Check in and save bookmark! Construction
If access to Sweden is slow, try the last mirroring copy at GeoCities:
Please, be patient if some links have expired. VETA is moving.
Under Construction

VETA is a Swedish acronym for Science and Experience in Traffic safety Work
In the VETA research society we support drivers
by providing knowledge rather than blame

Accidents are unintentional like diseases
Prevent them accordingly!

VETA logo
To VETA Homepage

Antilock (ABS) brakes
Automobile industry contribution to safety
devaluated by inexperienced safety researchers?

New evidence of bias
ABS-risk graph

Access no.

4 Nov, 1995

Dec.27, 1996


Links section have moved from here

... though GeoCities still host some stuff on ...
E-mail directories and other Internet Utilities

To top of page

Antilock (ABS) Brakes

Safety Effects Underestimated?

ABS = Anti Blockier System (German abbreviation)

In spite of disagreements from other researchers, reports from (the industry independent) HLDI and NHTSA have resulted in a widespread and common belief ('authorized' even by Time) that accidents are not reduced with antilock brakes. A German University study 1983 on Munich taxi drivers is still referred to in this context (see link to excerpt from Target Risk). I (Lennart Strandberg) commented on it in a paper to the 13th ESV Conference, see page 819 in the proceedings from U.S. Department of Transportation (No.DOT HS 807991, paper S7-O-08).

Objections from people in the automobile industry have been neglected, though the conclusions are biased.

Insurance study overlooks mileage?

HLDI does not take mileage into account when
claim frequencies are compared among the 1992 year models with standard ABS and the 1991 models without ABS. Though annual mileage tends to decrease with vehicle age, the HLDI unit of exposure (insured vehicle years) makes no difference between the younger ABS and older non-ABS cars. In the first HLDI report, twelve groups are compared (on an annual basis only), and in the Summary it is stated: "No differences in claim frequencies or average loss payment amounts, under either collision coverage or property damage liability, were found for cars with standard antilock brakes compared with the same models without antilocks. In almost every instance, the difference in the frequency of claims was not statistically significant." ...
... "the comparisons presented in this report provide no evidence that the introduction of antilock brakes as standard equipment reduces collision or property damage liability losses resulting from real-world crashes."

Quoted from Insurance Special Report A-41 (January 1994) with last quote almost identical in second Report A-47 (January 1995), Highway Loss Data Institute, 1005 N. Glebe Road, Suite 800, Arlington, VA 22201, USA. See also the overview at AutoSite.
Unfortunately, I have not heard of any attempt to clarify the influence from US differences in mileage. Therefore, I made a rough estimate with HLDI's own data from Report A-41 and Swedish mileage distribution between different year models.
Then I found that the ABS-cars exhibited less damage per mileage unit in all 12groups distinguished by HLDI.
The ABS-favourable difference was statistically significant in 11 out of 12comparisons.
You may find one of them in the diagram below.

Clicking on the image will enlarge and display it separately for downloading.

Evidence of mileage bias in data on the Web

Without any reservation or considerations of the mileage bias, HLDI and IIHS (Insurance Institute for Highway Safety) continue their devaluation of ABS with statements like "Antilock Brakes Don't Reduce Fatal Crashes".
Like the past HLDI studies, this one may confirm the advantages of ABS, too. It is remarkable that IIHS does not present results adjusted for that the newer (ABS) cars are likely to be driven longer distances per year with less mileage on urban (low-speed) roads.
Such explanations may invalidate Risk Homeostasis speculations like "Drivers might feel overconfident and drive faster or take more risks" (quoted from the body of IIHS news).

In this study, IIHS/HLDI's own data provide strong evidence of a mileage bias:
In 20 of 24 rows of their web page table, the right-hand column exhibits greater values than the second column from the right. This is consistent with a greater mileage bias for the right-most column, where the average age difference between ABS and non-ABS cars is likely to be more pronounced.
The substantial column differences for pedestrians and bicyclists on the last rows are as expected, if the older cars are used more in urban areas.

The ABS cars seem to have a greater fatality rate in Single-Vehicle compared to Multiple-Vehicle crashes. This may also be an effect of confounding variables, such as urban mileage ratio.
However, steerability may have been put forward of stability in the first generations of ABS, resulting in more violent yaw motions and side impacts than if front wheels lock up (occupants are more susceptible to injuries in side- than in frontal impacts).
Such problems are taken into account in the ongoing development of GMA (Gier-Moment-Abwäschung) and modern ABS software.

Note: Since this was written without access to the full IIHS report, you are kindly asked to consider my comments with scepticism.

NHTSA report on percentages vs. absolute numbers

The NHTSA report confuses me by putting forward some subgroup percentages that are unfavourable to ABS, though the total absolute number of crash cars is 1000 less with ABS (20000 cars) than expected from a random outcome (21000 cars). This is significant, indeed (Chi2-probability 3E-29, if my calculation is correct).

In spite of these data, clearly favourable to ABS, you may read on pagev in the NHTSA report:

"The increase in run-off-road crashes approximately offsets the reduction of vehicle-to-vehicle collisions. Thus, NHTSA estimates that there has been little or no net accident reduction with ABS, to date. NHTSA's finding is consistent with the accident analysis published by the Insurance Institute for Highway Safety in January 1994."
Quoted from NHTSA Technical Report DOT HS 808 279 (May 1995). Document available from National Technical Information Service, Springfield, Virginia 22161, USA.
(The referred Insurance Institute is part of the same organisation as is Highway Loss Data Institute, HLDI, mentioned above.)

To top of page .

Misleading anti-looks at Antilocks

Free Speach 4 U 2 on Internet On November 15, I finished two VETA papers in English on these observations: In the VETA society we are not yet perfectly prepared for international distribution. However, we will try to send these reports with an invoice for postage and package (no more than 100 SEK), if you write down your name and address in an Email to You will receive these and all other VETA reports issued 1994, 1995, 1996 and 1997 at no extra costs, if you become a member of the VETA society. The annual fee is 300 SEK (about 50 US dollars) for people outside Sweden. The following question is discussed in both papers listed above.

Is ABS used for exercise by inexperienced safety researchers?

Perhaps the speculations on risk compensating behaviour are the main reason why people question the safety effect of ABS. At least, I have found certain investigators very eager to use ABS as an object for research exercise. It is quite easy to get data for statistical testing of hypotheses in this context, since ABS is so common without being excluding.
When the outcome of a statistical test does not show any significant differences, medical researchers suggest that the results are negative. However, it has not of course been proved that ABS brakes or other measures are without any effect, only that nobody has succeeded in proving any effect. This is a vital scientific principle which should apply in the area of traffic safety as it does within epidemiology.
It may be reasonable to assume that there is no real effect, if many independent studies show negative results. In the ABS issue, though, literature surveys and references often can be tracked down to one single study: the Munich taxi experiment in the early 1980s. In an assessment of the driving behaviour (claimed to be of the double-blind type), the investigators found that the "drivers of cabs with ABS made sharper turns in curves, were less accurate in their lane-holding behaviour, proceeded at a shorter forward sight distance, made more poorly adjusted merging manoeuvres and created more traffic conflicts."
(quoted from link to Excerpts from Target Risk).

This driver behaviour and attitude towards ABS may be representative for the early 1980īs. Unfortunately, another part of the Munich study is also referred to in many contexts as proof that ABS brakes do not improve traffic safety. In retrospective accident analyses the results were negative: no significant differences were found between ABS and non-ABS taxi cabs. Therefore, it was concluded that ABS had no effect.
A closer inspection, however, reveals that there are many confounders in the first analysis (1981-1983) based on 21 cars equipped with ABS brakes. With so small numbers it is remarkable that these negative results are considered a proof of ABS inefficiency more than a decade later.
In the second Munich evaluation period (1985/86) there were even fewer cars: 10 cars without and 10 cars with ABS brakes. Based on these circumstances it seems fairly natural that the risk indexes for all accident types did not differ with any degree of statistical significance.

If I am not mistaken above, ABS has been devaluated through overinterpretations of immature research. However, since the HLDI and NHTSA studies have been reported in detail, it is possible for the scientific community to check the methods. If we don't, nobody should blame the misunderstandings on NHTSA or on HLDI with their well-defined evaluation of car crashworthiness.
Without such clear definitions, conclusions could not be questioned so easily. It is also thanks to the NHTSA open-mindedness and to the NHTSA Library, that it is possible to discover and constructively criticise research, as in all scientific work. I can recommend a visit to the NHTSA report download service.

To top of page.

Antilock (ABS) Brakes: How to use them?

Irrespective of accident statistics, I support the warnings (on inadequate driver reactions) given by HLDI associates and by NHTSA, also explaining ABS-technology in the downlodable PDF-file dated 03/21/95 (92K).

I strongly recommend ABS-car drivers to study the advice from one of the auto-manufacturers. And don't miss Autonet's comprehensive guide with technical details and pictures of ABS and its components or their ABS-questions and answers. An excellent overview of ABS-related Tech & Trends from a driver's perspective was given by Bill Visnic on page 101 in the December -95 issue of WARD'S Auto World.

Whether ABS or not, brakes must be properly maintained to avoid failures and spin-outs from rear wheel lock-up , which is likely to occur if you try to stop the car by braking the rear wheels only (for instance with the parking brake or by downshifting in a rear wheel driven car). DANGER!

If you excuse my SwEnglish, I would also like to contribute a translation from a recent Swedish press-release on the VETA study of braking capability on slippery surfaces in experiments with 66 ordinary drivers.

New website and 2-language Homepage
In English VETA På svenska
Tvåspråkig hemsida och ny domän

In addition GeoCities are hosting webpages ...

To be conscious that you are ignorant is a great step to knowledge
(Disraeli quoted by Bartlett)

To top of page

Who is the guy writing by I, just nearby?
Brief curriculum vitae for Lennart Strandberg,
founding chairman of VETA society and its temporary webmaster.

Swedish flag Born in the Dalecarlia province of Sweden. Living outside Linköping 200kilometres south of Stockholm (capital city of Sweden) together with a wonderful family.
My son Victor (born 1983) has opened a site by himself on StarWars and games like Need for Speed, Descent, Slipstream 5000, Doom, Dark Forces etc.

Practical experience as a test and racing driver (more the European Rally or BTCC style than the American Indy Car or NASCAR style, though).
My driver's licence covers all road vehicles (A BE CE DE in Sweden).

Occupation: Traffic safety research consultant.
Charter chairman of the VETA research society.
Professor in accident research at NIWL (presently off-duty).
Until December 1994: driving and traffic safety research at the Swedish Road and Transport Research Institute, VTI

Member of Linköping-Filbyter Rotary Club
(district: 2410, classification: Driving safety research).

To top of page

Best Wishes from Lennart Strandberg
Fax: +46 (0)13 52992
GSM cellular phone: +46 (0)70 54321 00
Mailing address: VETA, Box 1, S-59054 Sturefors, Sweden.

Webspace kindly provided by the American company GeoCities
Return to GeoCities Athens