Home - About Us - Riding Safer - Contact Us - Blog - Disclaimer - Links - Sitemap
Risk Hierarchy: Information - Rider Ed - Driver Ed - Conspicuity - Bike Defect - Ultra-Defensive Riding - Crash Avoidance - Injury Mitigation - Crash Scene

The 'Naturalistic' MSF/Virginia Tech Bike Study.

In the light of MSF's March 31st, 2010, announcement of a Naturalistic Motorcycle Crash Study with Virginia Tech, we need to pay attention to this new alternative to the troubled OSU Crash Causation Study.

This study technique, pioneered at Virginia Tech, has previously been developed in truck studies and with cars in the 2005 100-Car Study. It hasn't been done in motorcycles before, and adapting the machinery to fit on a bike poses several challenges. But monitoring vehicles in this way gives very accurate information on the events leading up to crashes, but also all sorts of near-crashes and incidents, and provides opportunities to examine successful evasions, which, by definition, are not seen by conventional, Hurt-style studies.

We're going to examine the 100-car study in greater detail. This page is a work in progress. We read the 100-car study report, but we haven't got a good handle on its cost, or how much MSF is planning to invest in this new motorcycle study.

The 100-Car Study.

Here's an overview of the Virginia Tech Study, here's VTTI's page on the study, and here's a .pdf of the full report.

Crashes, Near Crashes and Incidents

The 100 cars in the study were followed for about a year and recorded a little over two million miles. They logged 69 crashes, 761 near-crashes, and 8,295 'crash-relevant incidents'. Only 14 of the crashes were police-reported, the majority of crashes were less severe. In a standard, Hurt-style study, where only about a third of police-reported crashes are responded to by the research team, this would have resulted in maybe 5 study incidents, assuming the crashes were within the geographical area of the study. Some of the 'crashes' were as minor as a tire strike. All of the crashes and incidents involved an event that could have required evasive maneuvers, so this is a huge body of relevant data.

The relatively high average mileage and number of accidents were because the researchers tried to recruit higher-mileage and younger drivers to increase the probability of crashes. This might also have skewed the results.

Study Technology.

The cars were equipped with:

As this was early 2000s technology, we'd expect that the equipment is capable of being miniaturized,and factors such as improved cellular bandwidth, newer video compression technologies like mpeg-4 and larger capacity disk drives will make fitting the equipment to a bike somewhat easier, but still a challenge.

The 100-car study was limited in range because they had to send a chase car to download the data on the test car's hard drive from time to time. This would be a problem for a bike study as bikers tend to take long trips from time to time. But the 100-car study yielded some slick techniques for finding incidents of interest among the millions of miles of incident-free driving, and this could probably be placed in the bike's computer to make the chase chore unnecessary. The study yielded a lot of information and techniques which will be useful in adapting the technology to bike use.

One wonders how this equipment will fare in the fairly common bike drop events that every biker has experienced. Stuff mounted on the handlebars might be vulnerable to the casual drop incident. The challenge of mounting fragile equipment and keeping the weight down will be huge, we think. It's the standard motorcycle engineering problem, which will probably be solved the standard motorcycle way, by better engineering and more maintenance. MSF's involvement and their ties to the motorcycle manufacturers should be a great help with this process.

One of the factors in the 100-car study was the need to analyze the electrical and computer systems in each model of car, which restricted the vehicle selection to six models in specific ranges of model years. Because of the greater specialization in the motorcycle breeds, such as cruiser, dual sport, sports touring, crotch rockets, commuter bikes etc., it will be harder to have such a small list of eligible bike models, especially as they'll want to have ABS and non-ABS versions of similar bikes. They'll also probably be very interested in measuring traction issues, and possibly bike-specific issues of tire pressure and the level of pre-ride checks the rider does. We think assessing the rider's skill levels and previous training might be important, and, as well as ABS, we hope they'll be paying more attention to weather and animal issues, to correct oversights in Hurt and other studies.

We also think that they'll naturally collect information on both helmet and non-helmet wearers, but the lesser importance of actual crashes and the relative lack of follow-up on injuries that they can do may downplay the role of helmet and protection issues in the study. It would require the researchers to follow up on injury crashes. Our take on this issue is that the information on helmet effectiveness is there for anyone who wants to see it, and this study getting into the trench warfare of the helmet debate might detract from its effectiveness.

Study results

The study had a number of objectives, including assessing the naturalistic technology itself, and developing the techniques. The study was effective in these areas.

The study did produce a large body of information on the crashes and incidents, and developed some major themes, especially the effect of distractions. This includes cellular devices, but also talking, eating, fatigue, grooming and paying attention to another driving task. The study found driver distraction to be a factor in 78 percent of crashes, more than twice the number they expected. In retrospect, this makes sense, as many drivers might downplay these factors in the police reports of crashes, whereas the video cameras show all.

Driver skill levels were also important. All these factors are likely to be issues in a bike study, although there are probably fewer opportunities for distraction on a motorcycle. In all, the final report has almost 500 pages with detailed breakdowns of the crashes and incidents, with analysis of pre-crash maneuvers, contributory factors like environmental problems, evasions, post impact maneuvers etc. There's a lot of information, and it is not directly relevant to motorcycles, so we refer readers to the project summary and main report for further details. It is probable that this study and the truck studies done by VTTI have played a large role in the recent attention that USDOT have been paying to distracted driving, which is especially important to motorcyclists.

We'd love to see this level of information on motorcycle crashes, and especially the effectiveness of the various evasive and predictive techniques in avoiding them.


It's early days yet. This type of study has not been done in motorcycles, and there are only a few previous studies on cars and large trucks. We think that this is a viable and creditable effort to improve our knowledge of bike crashes and their causes.

We would hope the VTTI will be making some effort to encode their data in a format compatible with the OECD guidelines. This might make the study database capable of being merged with the OSU study data, which could strengthen both sets of data. We hope the fact that both studies are competing to some extent won't eliminate the possibility of co-operation in this regard, or at least the door might be left open to someone performing this combined analysis of both studies sometime in the future.

Paper trail

This study was paid for under NHTSA contract DTNH22-00-C-07007 Task Order 23. We haven't yet found out what the project cost was, we will get this information soon. We assume, from the fact that MSF chose to support this project instead of the very expensive OSU study, that the cost was attractive, and we'll submit another FOIA request for this information on the 100-car study if we have to. As the current study is being privately sponsored by MSF, we might never learn the true cost. We might not know if the funding is adequate until we see the final report. For now, we have to hope that MSF is doing the right thing here.

Tom Dingus, the current director of VTTI, was an author of the 100-car report, so we are sure that VTTI is fully invested in the motorcycle study. We are not aware of any current motorcycle expertise at VTTI, and we hope that their researchers are busily getting up to date on the motorcycle study literature and learning to ride. Any riders who would be interested in earning a post-graduate qualification in this area should contact VTTI.