By on March 4, 2010

And the data beat goes on. I asked Edmunds if they had updated model information to filter out the spike of UA reports to NHTSA after the 9/29/09 Toyota mat recall in order to improve my attempt at coming up with a model-specific UA rate. Not only did they oblige, but they already did all the work! A big hat tip to Edmunds, who has taken a lead in the quest to make sense of the data as well as the whole UA fiasco.

A note of explanation from Edmunds:

Attached are the UA complaints against sales by Model for MY2005-2010 and complaints received through 9/30/2009.  A couple things to note about our “complaints per 100k sold” measure: if the model sales was less than 30k, then we excluded it from the list and if the model sales was between 30k and 100k, then we extrapolated the complaints to 100k.

Furthermore, I removed a few cars that had very low UA reports, generally less than four. But I left other in with low numbers because they were essentially “twins” of other models (Fusion/Milan), in order to test how reliable and consistent they are between the. I’m happy to see that generally that is the case: (Grand Cherokee: 7.57; Commander: 7.30); (Fusion 2.91; Milan 3.18) (Vibe 2.85; Matrix 2.75).

Obviously, the same question as to the Panther triplets comes up: they’re all high, but by varying degrees. And of course the biggest on is the discrepancy between the ES 350 (32.03) and the very similar Camry (6.52).

One more minor note: the Lexus LS 430 has an old-time bottom-hinged accelerator pedal, so mat entrapment is not an issue with it. And it hasn’t been implicated with a sticky pedal either. So in its case it must either be human error or…

Finally, the next step would of course to cross tabulate both sides of this chart to specific model years, since awe saw in our previous post how much variation there is from year to year. And then…

Get the latest TTAC e-Newsletter!

53 Comments on “NHTSA Data Dive 3: 117 Models Ranked By Rate Of UA Incidents...”

  • avatar
    Telegraph Road

    Paul, Edmunds has indeed done some excellent analysis on this issue. What the B&B stat-heads would like to see now is how this ranking compares to the average age of buyers by model line.

    • 0 avatar

      That, and also cross-referenced by annual average milage per model (data has to exist, but probably only in the OEM’s databases based on mileage recorded in service, warranty calls, and at lease expiration.) I expect that it is the different types of customers buying the variants of the vehicles (Toyota v. Lexus), and the related usage (company car or private car) that is accounting for some (but not all) of the statistical variation between similar models.

  • avatar
    fred schumacher

    Re: “So in its case it must either be human error or…”

    I think the statistics show pretty clearly that human error is not the primary variable. If it were, we would see a high degree of uniformity among vehicles; instead, there is a tremendous variation. So much for the “little old lady who doesn’t know how to drive” argument.

    • 0 avatar

      Yes and no.

      Let’s assume something similar to the Audi 5000 scenario: that there’s something wrong with the physical layout or behaviour of the pedal in Ford or Toyota vehicles that is primarily affecting older drivers.

      For some reason, that same behaviour does not exist in GM, most Chrysler, Nissan, Honda and suchlike. So, the question becomes: “What is it about Ford and Toyota pedal characteristics that cause drivers to misapply the pedal”. A laundry list might include:
      * Placement relative to the brake
      * Offset versus the brake
      * Tendency to physical obstruction (mats)
      * Cruise behaviour
      * Pedal angle versus seat (which could explain the Tacoma’s ranking, as well as the Venza, Matrix and Sienna; the Taco sees the driver’s legs go forward; the various Sports Buses see them drop down)

      This makes some sense. Physical pedal issues are more likely to a) not show in the EDR and b) not trip a sanity-check.

  • avatar

    From everything I have read, the ES350 has the Denso pedal, and the Avalon has the CTS pedal, so the data says to me that CTS, (an American company) has been thrown under the bus by Toyota. It is amazing how data, when sorted properly, gives a much more accurate view of the situation. I’m starting to think that the problem is partially being caused by the age of the drivers involved. I’m glad that I treat my recalls like I treat my computer operating system upgrades, let others be the guinea pigs to see what problems develop. Seeing how the major news organizations are now reporting that the “fix” may not have solved the problem, or created new problems, it appears I have made the right decision. I hope Toyota gets control of this situation soon, as I would hate to see American workers suffer needlessly. As others have mentioned, I think Toyota needs to install a throttle cut off upon brake application to all models involved, no matter what the cost, or the downside. Their reputation literally hangs in the balance.

    • 0 avatar
      crash sled

      Good point. The Tacoma at #4 on this list also is fitted with the Denso pedal, so that shouldn’t be at issue (although it does have the floor mat issue).

      I have to assume Tacoma drivers would be a “young” demographic, unless you’re implying that I’m “old”, in which case I’ll have to slap the taste outchyermouth. ;-)

      Many of those vehicles would have a manual, as mine, and that may be a factor, in some strange way. The clutch is a failsafe you would think, but in practice and under stress, maybe some confusion there? The brake and accel pedals are very close together, per many Tacoma owners, so this could also be an adder in some cases.

  • avatar

    I had to Google what a Pontiac torrent was and the first thing that came up was “Pontiac torrent problems”.

  • avatar
    Greg Locock

    Yup, you need another column. Average age of driver at time of incident.

    Then compare that with the average age of the owners for that model.

    Then you’d really be able to figure out whether it is little old ladies panicing or whatever.

    Good stuff.

  • avatar

    @ fred shumacher

    I disagree. Since the dats is not sorted by driver age, it is hard to make that conclusion, however, if you look at the type of cars at the top of the list, assumptions can be made, although they MIGHT be wrong. ES350: usually older driver. Lincoln town car: same. Mercury Grand Marquis: same. Toyota Tacoma: probably younger, therefore a wrench in the works. Could possibly be pedal spacing or other indeterminate cause. I think the data needs further refinement.

    • 0 avatar
      fred schumacher

      If older drivers are the problem, then GM should head the list, since by a lifetime of customary buying, GM should have the greatest number of old drivers.

      There are more than two orders of magnitude difference between the Lexus at the top of the list and the Chevy pickup at the bottom. What is your independent variable? I don’t think it’s the driver, not with that kind of spread.

  • avatar
    Gardiner Westbound

    BMW, Audi, Mercedes-Benz and Volkswagen models all have brake override software. Significantly, none are on the Edmunds report.

    Is it really that expensive or difficult for Toyota to flash the systems with revised code?

    • 0 avatar

      In the House Hearings, the U.S.Toyota CEO stated only about 71% of their current product could be flashed. The remainder would have to be replaced.

      Extend this back to 2002 for all models and it’s likely a larger percentage would have to be replaced.

    • 0 avatar

      @CarP: I don’t recall him saying anything about replacement. Could it be that we drew two different assumptions about the non-flashable vehicles, you that they (might) be replaced, or that if something had to be done, they would have to be replaced (based on what you wrote, I’m not exactly sure how you interpreted what was left unsaid, and don’t want to be unfair and put words in your mouth), me that no action was contemplated for these vehicles?

    • 0 avatar

      @ Robert.Walter

      I was addressing Is it really that expensive or difficult for Toyota to flash the systems with revised code? A Representative asked the same question in the House Hearings.

      The response did not address what they were going to do but the capability to do such a flash. I was very surprised to learn not every ECM is flashable.

      The few cents they saved added up to a nice sum over a large production run but on a completely new system? In my view, proper risk management would not have allowed that on an unseasoned ECM.

    • 0 avatar

      Toyota must fix this with a rewritten ECU. Whether they flash it or replace every unit, they simply must if they want to continue selling cars into the United States market.

      It’s clear there are too many problems in too many places. For customer relation reasons, you can’t blame the drivers, but in any case the data screams it’s impossible the majority of cases are down to elderly, or clueless soccer mom types, or cell-phone-distracted drivers. Where’s Buick? Cadlliac? BMW 5 and 7, Mercedes E and S class? Kia and Hyundai?

      Audi 5000 recall actions:

      1) Issue recall relating to floormat-pedal entrapment.
      2) Issue recall installing shim to better differentiate brake and accelerator pedals
      3) shift interlock
      4) decade of meagre sales

      Bertel, would you happen to have an idea of how costly the shift-interlock retrofit was for Audi?

    • 0 avatar
      Mark MacInnis

      Ironic, that bit, eh? Audi supposedly had the problem in the 1980’s, worked on it, developed an apparently reliable solution….which other German companies adopted. American and Japanese cars….not so much. Wonder what the per-unit cost savings of not having the brake override software is? We bitch and moan about the nanny state taking over our cars, but when something like this happens….we WANT them on that wall.

      Full disclosure: Audi A6 owner

    • 0 avatar
      crash sled

      “…the data screams it’s impossible the majority of cases are down to elderly, or clueless soccer mom types, or cell-phone-distracted drivers.”


      I wouldn’t be too quick to dismiss any of that, CE, especially the elderly, a demographic that is increasing as the boomers age, and will increasingly impact these types of analyses.

      And the cell phone distracted are an issue, and we all see this as we drive. I read a survey the other day that 2/3’s of drivers aged 19-29 have texted while they drive. Think about that. They’re keying-in data, not just talking into a phone. I’d like to see that survey data from other countries, just as a comparison.

  • avatar

    The Toyota Camry data is so sharp-edged it is dificult to believe the demographic, age, mental capacity, cell phone use, shoes, or anything else with a gradual trend line could be the culprit. Direct current (DC) circuits have a built-in natural immunity to EMI/RFI, so that doesn’t seem too promising.

    That only Toyota has experienced the spike and only after a new product launch also tends to exonerate outside sources.

    By sharp-edged, I mean you can almost determine the day the MY2002 Toyota Camrys started rolling off the dealer lots by the UA complaints to the NHTSA: They took off like a shuttle launch.

    People continue to die. Eight years is enough. It’s time to drop the secrecy and allow outside experts to review the firmware code.

    • 0 avatar

      The last paragraph of my post is a little more strident than what I indended. Pressing the “Enter” key seems to do that.

      But hey, if Toyota isn’t able to get a handle on this, why not bring in outside help with some knowledge of the topic and see if they can find something.

    • 0 avatar

      “Direct current (DC) circuits have a built-in natural immunity to EMI/RFI”

      That’s not true. DC circuits are susceptible to all sorts of noise. Take a look at the shielding and those big cylindrical inductors on your laptops power cord and USB cables.

    • 0 avatar
      Rod Panhard

      Really? Is that why we have “noise suppressing spark plug wires” so we can’t hear DC interference in our car audio systems?

      Carperson, methinks you try to hard.

    • 0 avatar
      crash sled

      Yeah, and the 6 wire DC bundle to/from the Toyota pedal sensors is shielded, so I assume they think it’s susceptible to noise.

    • 0 avatar

      DC circuits are more resistant to EFI/RFI than AC circuits. Steady-state, decrete voltages give you that advantage in a dirty environment. Part of it is that they are much less likely to have to deal with any sinusoidal or coupling issues.

      If you are interested, you can trigger off either the leading or trailing edge of the DC voltage change, or both edges.

      However, try to split 0.1Vdc in a vehicle like Toyota is apparently trying to do, and you will likely find yourself in bed with Mr. Floating Ground. DC LOVES it’s Earth and is most intolerant of an interloper.

      My guess is you could drive the car into a microwave and it’d burn the paint off the hood before it disrupted the DC signals, hence the comment this line of inquiry (EFI/RFI) would not appear to be too promising.

    • 0 avatar

      I have read that the Toyota throttle system calibrates at start-up as well as periodically. Let’s presume that is true.

      How does it calibrate when I have the throttle half-way pressed on startup? Is that the new idle value?

      Of the accidents I’m familiar with, all appear to be within the first five minutes after the car was started. Is this at or near a re-calibration point?

      I’m getting the feeing if I’m in a Toyota, I’d want to kick it into Neutral, shut it down, re-start it, and put it back in Drive about every 30 minutes to re-boot the ECM.

  • avatar

    @ crash sled,

    I would never presume to call anyone old. Sounds like you are about the same age as me, and I prefer the term “seasoned” LOL Be well!

  • avatar


    Can you post the tables in excel format? Thanks!

  • avatar

    Hi Ed, may I request the addition of additonal columns? Or maybe you can suggest it to your contacts at Edmunds.

    With an “x” in the corresponding boxes for 2007 mat-entrapment recall, 2009 mat-entrapment recall and 2009 shim recall. If anybody knew which of the 29% could not be reflashed and which vehicles accounted for the reported deaths and fatalities, this would have the potential to be very telling (I don’t think the details on the 29% were released by TMC, but the fatalities and injuries should be in FARS and in the recall notices ((but probably not updated or broken out)).)

    I think this would be enlightening.

  • avatar

    Wow, that’s interesting.

    Since the xB has only 4 complaints, it ranks 62nd. If you add my experience and PN’s (which were not reported to NHTSA), it jumps to #36. My curiosity is piqued. How many unreported incidents are out there?

    • 0 avatar

      The current thinking is that drivers rarely report the problem. A few will talk to the dealer and far fewer will contact the automaker or NHTSA.

      Dealers may discus this with the automaker on an informal basis but NHTSA has stated they all but never get an individual report of any kind from a dealer.

      The current media frenzy and the NHTSA publicly soliciting comments is helping to address how many more are out there unreported.

      Thinking the Japan market might be a mine of information, Mr. Toyoda was asked what data they had. His response was far less than what was available in the U.S. In Japan, people are much more reluctant to talk about it or file a claim.

  • avatar
    John Horner

    It is interesting to see how different the reported rates are between the Mercury Mountaineer and the Ford Explorer. Aren’t those two vehicles nearly identical twins?

    But, driver age can’t be the whole story. How would you explain the Grand Marquis being near the top of this list and the Buick Lucern near the bottom if driver age is the primary factor?

    There seems to be a great deal of noise obscuring whatever signal might be buried in this data.

    • 0 avatar
      Rod Panhard

      I’m with you on that John Horner. A lot of noise.

      Give me 8 million units of anything. I can find 50 incidences of unintended acceleration in whatever you’ve got … coffee cups, gumballs, bookends, door stops, anvils, flat screen TVs…you name it.

      Follow the money on this one. It’s curious that reports of UA are non-existent in the rest of the world. But here in the US?

    • 0 avatar
      Christy Garwood

      @ Rod, are there mechanisms in other countries for people to report any safety issues? The USA’s NHTSA ODI can be accessed by anyone around the world. Where can we go to see the same in Japan, Germany, France, UK, Brazil, China, etc.?

  • avatar

    Looking at the top 20, it’s clear that Toyota and Ford dominate, with a Volvo, a Pontiac and a couple of Jeeps thrown in. GM cars are more likely found in the lower end of UA ranks. In a previous post, I suggested with significant doubt that, “GM owners more than any others carry on the historic American garage-mechanic cultural value where car owners actually know how cars work, take an interest in the mechanicals and have a higher awareness of what to do and not do when operating a vehicle.” Paul N. agreed with me that this was, “A romantic notion, but not likely.” Can we be so sure? Or are GM cars better designed and built on this particular issue. It’s one or the other as the slicing and dicing of data, keeps reaffirming the rankings.

  • avatar

    In the Bush/Gore presidential race recounts in Florida, history was given a permanent icon of a be-speckled man staring at a card looking for a hanging chad. Who can ever forget that image?

    That image was prominent in my mind from the moment that I saw the above chart and while reading all of the above posts.

  • avatar

    I worked, a little, on GM’s electronic throttle development, back int he 90s.

    There are several safety assumptions built into the design.

    There are 3 variable resistors (rheostats) in each pedal. One operates in reverse of the other two. That is, as the pedal is depresse, the resistance decreases in 2, and increases in the other one. (or vice versa, it’s been a few years). The point is that the computer has to get consistent readings in both directions. Therefore a simple short or disconnection will tell the computer that there is a problem, and can’t be construed as a wide open, or closed, throttle.

    The brake override was also incorporated.

    And, GM’s pedal design standards mandated that feet in work boos cannot depress the gas pedal more than a little bit even by carelessly pressing the brake.

    I suspect that this is why GM cars are at the bottom of the list when it comes to UA complaints.

    The highest GM cars on the list are cars not designed by GM, and, presumably, not held o GM’s own design standards.


    • 0 avatar

      thanks for the info bob …

      i mused in one post that Sorels more likely contributed than flip-flops, and indicated in other posts over the past month or so that it would be interesting to see a) accl gap to the tunnel, b) the pedal-pedal spacing and c) pedal-pedal offset dimensions …

      aside from bad design of the pedal internals, i pedal-pedal packaging and pedal-vehicle packaging had a major contributory effect to creating the problem (as well as a lack of override having a major contributory effect to not stopping it)…

      cross-referencing these pedal-pedal and pedal-vehicle dimensions with reports and the 2 mat recalls and the 1 shim recall would be very telling and go a long way toward helping to clear up this mystery (even if ruling out things rather than ruling them in.)

  • avatar
    Rod Panhard

    I wonder how many Tacoma instances of Unintended Acceleration at the Wheel are tied to a workers comp claim? Pick-up trucks are used in the construction trades.

  • avatar

    I quit flying years ago, primarily because I was not a very good pilot, and it was too easy to get behind the curve.Is it possible that this is just some kind of mass distraction leading to panic?

  • avatar

    Based on the rankings of the top 20, I find it interesting that Toyota is being thrown under the bus (their own fault to be in this position), and Ford is skating along under the radar.

    Toyota = 6 of top 10 and 9 of top 20, Ford = 4 of Top 10 (inc. Volvo which the S40 was under their watch during the development and qc process, or should have been) and 7 of the top 20.

    Either way, Toyota is going to be forever changed from the events of the last 30 days, and probably a whole lot less profitable, based on having to throw cash on the hood or extend the warranty to bring customers back.

    FWIW, RE: distracted driving, yesterday on the highway I saw a mom, PUMPING BREAST MILK, while driving. If you know what this process entails, this would make TEXTING while driving look relatively safe…

    • 0 avatar

      Wow, talk about a most inappropriate place to get “a tit in the wringer”…

      Regarding the Toyota-Ford comparos, it would be necessary to add data for accidents, injuries and fatalities (from various sources including FARS, and several big insurance companies) into the comparison … as Bertel previously pointed-out, the NHTSA d.b. covers reports, but not all of them are created-equal.

  • avatar

    For anyone who’s still interested in diving into the data, here is a file that lets you chart the number of “vehicle speed” incidents and injuries by year for ANY model that you want.

    To add a model to the chart, click on the “Model” menu (gray box on the right of the chart). Then select the checkbox for any model you want to include.

    Unfortunately, I don’t have sales data by model and year, so unlike the great table from Edmunds, you can’t account for the relative number of cars that were sold. Also, the NHTSA data do not include the driver’s age.

  • avatar

    So can it be assumed that 100% of the UA vehicles in that list are equipped with automagic transmissions? I find it very hard to make my 5 speed Jetta move when it’s in neutral or when it’s in gear and the clutch is pushed down fully. It’s also very interesting when I attempt to press the go and stop pedals at the same time since the ECU basically says “what the fsck are you doing?” and lurches to a stop.

    I think Barry should ban automatic transmissions in all new vehicles. It’ll weed out the people that can’t handle the complex skill of driving a 3 pedaled car and UA incidents will mysteriously disappear. It’s fun to make believe…

    • 0 avatar

      Interestingly, no: manual transmissions don’t seem to offer much protection against UA. Among all UA incidents, the breakdown is:
      Automatic: 41%
      Manual: 4%
      Unknown: 55%

      Let’s leave aside the unknowns. That means that there were 10x as many incidents in automatic transmission cars as manual transmissions. That’s roughly the same as the ratio of automatic vs. manual cars in the general fleet (I think???), so we can say that UA probably occurs at the same rate in both types of cars.

  • avatar

    Chevy and Howie Long should break out the ads: Chevy Impala and Chevy Silverado: THE SAFEST CAR AND TRUCK IN AMERICA*.

    *Safest claim based on Edmunds compilation of UA complaints against sales by Model for MY2005-2010.

  • avatar

    Am I missing something? VW/Audi were listed as highest make in UA incidents for ’08 and ’09 here:
    But there is not one VW/Audi/German car on this list.

    • 0 avatar
      crash sled

      friz, the bloggers on this site have a suspicious and peculiar predisposition, re German vehicles.

      There’s only one thing for us to do about it:


    • 0 avatar


      VW is worst offender in 2008 and 2009. Why is it off the chart?

      VW sells about 200k+ cars per year. That should be a statistically significant number.

  • avatar

    “Furthermore, I removed a few cars that had very low UA reports, generally less than four. But I left other in with low numbers because they were essentially “twins” of other models (Fusion/Milan), in order to test how reliable and consistent they are between the. I’m happy to see that generally that is the case: (Grand Cherokee: 7.57; Commander: 7.30); (Fusion 2.91; Milan 3.18) (Vibe 2.85; Matrix 2.75).”

    Then explain how the 300 is on the list, but not the Magnum, Challenger, or Charger.

    • 0 avatar

      You’re making me wonder again about driver demographics. Are 300s being sold to Lee Iaccoca’s age group or to Snoop Dog’s? There is an elderly coupe that goes to our local Catholic church and shows up in a DUB edition 300. Seeing them pull up always makes me chuckle.

  • avatar

    Interesting stat about AT/MT cars having the same UA rate. Do MT drivers find that the clutch is also disabled in such occurrences?

Read all comments

Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

  • Corey Lewis: From what I see on other examples, that nice wheel was installed by an owner after the fact. 82:...
  • bullnuke: ToolGuy – figure 2/3 of a bale of hay per day @ $7.50/bale (price depends upon where you live)....
  • Jeff S: Most of us have more than 1 vehicle especially if you are married and have children. Owning any motorized...
  • Jeff S: @ToolGuy–I have noticed fewer gas stations in rural areas as well but I don’t know if that is...
  • RedRocket: Thanks for the laugh. Voice recognition is the most worthless thing ever introduced to the modern software...

New Car Research

Get a Free Dealer Quote

Who We Are

  • Adam Tonge
  • Bozi Tatarevic
  • Corey Lewis
  • Jo Borras
  • Mark Baruth
  • Ronnie Schreiber