mkaresh

Updated results soon for both the 2007 and 2008.

Additional participants remain needed, for all model years.

Car reliability research

mkaresh

We have updated results for the GL-Class to include owner experiences through March 2009. In terms of successful repair trips per 100 cars per year:

2008: 166, worse than average, decent sample size

2007: 110, worse than average, small sample size

We don't have enough participants to report separately by powertrain. If we did, it is likely that the diesel would have a higher repair rate, thanks to an intercooler hose that tends to detach. This is one reason for the higher reported repair frequency for the 2008.

Another common problem, for all GLs: the exterior door handles tend to break.

A big thanks to everyone who has been helping. We'll have further updates in August and November. With more participants, we could provide more precise information and cover all model years.

Mercedes-Benz GL-Class reliability comparisons

mkaresh

91 GL-Class owners now signed up to participate--a good start, but more remain needed to properly cover all model years and to report results separately for the diesel (if this proves necessary). Especially need more 2009s.

Car reliability research

mkaresh

Updated results in a couple of weeks. Additional participants would be very helpful, for all years.

Franman

dude - give it a rest. JD Powers has already thought of this idea and has a much better web site and results too.

mkaresh

Only if you don't look all that closely.

J.D. Power doesn't provide any numerical stats--only dots. Good for people who prefer idiot lights to instruments, I'll grant.

J.D. Power only covers the first 90 days and the third year. Some people care about what happens in between those points, and after the third year.

J.D. Power includes complaints along with things that can actually be repaired, with no way to separate out the latter. This supposedly led to poor MB results from brake dust.

Fourdiesel

JD Power also tends to rate models highly if the manufacturers support and advertise the JD Power name - Is there possible bias there?

StevethePilot

Nnnnnnnnnnno! Never!

But even as a subscriber to a magazine that reports to consumers (cough, cough), I'm participating in the survey when and where possible, because I'd like to see as much future data available as can be gathered.

mkaresh

Maybe I'm naive, but I don't personally believe this. Manufacturers pay for the right to use an award in advertising. But that's a minor income source. The real money comes from selling detailed results and consulting services--where J.D. Power teaches the company how to improve its scores.

Because J.D. Power makes its millions selling detailed results to the OEMs, they cannot provide detailed information to the general public. Essentially, you've got to choose whether your primary clients are manufacturers or consumers. Like CR, we've chosen to focus on consumers.

One thing I forgot to mention earlier: speed. Because of how our research process is designed, our results are, on average, nearly a year ahead of CR's. We're about to release results based on owner experiences through the end of June 2009. CR's results are still based on a survey conducted in April 2008. As you might expect, neither J.D. Power nor CR makes it easy to find out how old their data are. When looking at their results for a car, you're often looking at how it was doing a year ago, when it was a year younger and with 12,000+ fewer miles.

Franman

so what type of sample size are you using to conduct your analysis? I hope you have more thorough data sources than message boards, but based on the fact you're flooding this one across all model types, it makes you seem quite desperate for data. I could be wrong (wouldn't be the first time)

mkaresh

For some models we have over 100 participants, but the minimum for a full result is just 25.

Message boards are the source for about 8 percent of our traffic. But since there are cases where we're just below the minimum, even another participant or two can be very helpful. And with sample sizes larger is always better.

I don't think you'll find a single successful survey that does nothing to attract participants. People don't actively look for surveys to participate in. As for the number of message boards: if you want to provide information on all cars, and each board covers only one brand or even only one model...it's not very complicated logic.

Fourdiesel

You may be naive - I may be overly cynical - but I have learned a long time ago to distrust and heavily discount the accuracy of so called 'ratings' where ANY group with a vested interest; ie a manufacturer, gets to use the ratings provider's name in his advertisements for a fee! Built in conflict of interest!
Consumer Reports fought many legal battles over this 35 years ago and still, correctly in my mind, won't allow their name to be used this way. As the saying goes - "Even the appearance of impropriety is to be avoided".
In my mind, allowing use of their name this way puts JD Powers "in bed" with the highest bidder for their name.

mkaresh

CR isn't as touchy as they used to be on this issue. During this year's press days at NAIAS two manufacturers cited CR in their presentations to the media. The CR people in the audience did not object. That said, I highly doubt they paid CR for the rights to cite their results.

At the same time, I doubt J.D. Power's results are influenced by who pays for what.

Fourdiesel

As far as I can remember, CR has not objected to accurate quotes of their ratings in news articles. But I cannot recall a single instance where Honda, for instance, bragged about their CR rating in a paid advertisement - even though CR is clearly 'in love' with the Honda product.

mkaresh

They actively seek citations in the media. That's what the highly paid PR agency is for.

The two manufacturers that used their ratings in prepared presentations to media were Toyota and either Hyundai or Subaru (memory is a touch cloudy). I don't see anything wrong with this except that it is against CR's rules.

StevethePilot

I can see "actively seeking citations" as against CR's rules, but I don't know how a manufacturer could be violating a CR "rule" by citing CR's own published results in their prepared presentations? Without having to qualify (unless I misunderstand their entire system) to have a product tested, I don't know what rules a company could've broken by quoting a published test result?

Not that I doubt you, I'm just confused by the whole thing. (I am, by the way, both a supporter of you and your method, which I like a lot, and CR and their method, which I take with a couple of grains of salt.)

STP

Fourdiesel

As I understand it, CR owns the copyright rights to any new material they publish. It has long been held in court that a news article or report that quotes other copyrighted material [source given] is 'fair use' and no copyright infringement exists. However, any advertisement for the benefit of say, a car company, that quotes either CR or JD Powers without their consent (ie the payment of a fee) DOES violate the copyright owned by the publisher.

JD Powers clearly charges substantial sums for the use of their 'results'. That is how they exist. They are also very close mouthed about the details of precisely how they arrived at the conclusions they publish (or allow to be published) thus it is virtually impossible for anyone to reproduce their results with any degree of certainty. For the same reasons, peer review is not possible. Thus the credibility of their conclusions is weakened. For example: JD Powers rates car X model Y as "Best in initial quality". Just what does THAT mean? Initial quality?! Fewest defects from the factory or as 'detailed' by the dealer? Does it matter if the transmission shifts or the tires fail when you are a block from the dealer? Apparently not since that is not initial quality!! The basis for their conclusion is never revealed because, if it were, the weaknesses and limitations would be revealed and the paying customer (the manufacturer of Car X) wouldn't like it.

If CR's rule is that they DO NOT sell their results for the benefit of the makers of the products they test then they improve their credibility. However, this only works to the extent they enforce their copyright ownership rights.

StevethePilot

Wow. Thank you. A clear and conscise explanation. Thanks!

mkaresh

It might well be true that CR cannot legally prevent manufacturers from citing CR's results in their presentations to media. But it's a fine line between using CR's results in advertising and encouraging media to cite these results when writing up a product. PR and advertising aren't the same thing, but they are both marketing activities.

On peer review, etc.--no car reliability survey releases the data this would require. CR is even more secretive than J.D. Power about their data collection and analysis methods. J.D. Power, after all, must answer to market research pros at their clients, and these pros want the details. CR, in contrast, refuses to divulge more than the bare minimum of information as a matter of "policy," and they get away with this because the average car buyer doesn't demand more than black and red dots.

TrueDelta already publicly releases far more information about its methods than either J.D. Power or CR does, and I've been thinking about making the dataset open source. Haven't had the time to give this sufficient thought, though. I'm certainly open to suggestions.

Fourdiesel

True Delta does an OUTSTANDING job in being open about the basis for the conclusions reached. I would not recommend making the database 'open source' because the efforts to accumulate this data are so great. Too much work goes into doing what you do to just 'put it out there' for others to use.
That said; the supporting data, analysis, and algorithms should certainly be available to 'members', contributors, etc if they have questions or concerns.
The work is yours, the data is ours (individually), but the collective impact is the real worth.
Peer review is a VITAL tool to establish credibility. Both CR and JD Powers are much less credible because of this factor. Most people feel that they cannot completely trust data if they are not confident in the integrity of the sources.

mkaresh

Unfortunately, in most cases people's confidence comes from NOT looking into the details.

We have updated results for the GL-Class to include owner experiences through June 2009. In terms of successful repair trips per 100 cars per year:

2008: 145, worse than average, decent sample size

2007: 111, worse than average, small sample size

We don't have enough participants to report separately by powertrain. If we did, it is likely that the diesel would have a higher repair rate, thanks to an intercooler hose that tends to detach. This is one reason for the higher reported repair frequency for the 2008.

Another common problem, for all GLs: the exterior door handles tend to break, and the air suspension also appears to be a trouble spot.

A big thanks to everyone who has been helping. We'll have further updates in November and February. With more participants, we could provide more precise information and cover all model years.

Mercedes-Benz GL-Class reliability comparisons

mkaresh

Updated results again next month. It's looking like results should be reported separately for the gas and diesel. Need more participants to make this possible.

Car reliability research

mkaresh

Updated results later this month.

We're going to start providing new stats for the odds of requiring no repairs and the odds of getting a lemon. These require larger sample sizes, so additional participants would be helpful.

mkaresh

We have updated results for the GL-Class to include owner experiences through September 30, 2009. In terms of repair trips per 100 cars per year:

2008: 164, worse than average, decent sample size

2007: 81, has improved to about average, small sample size

We still don't have enough participants to report separately by powertrain. The diesel intercooler hose, exterior door handles, and air suspension have been common problem areas.

A big thanks to everyone who has been helping. We'll have further updates in February and May. With more participants, we could provide more precise information and cover all model years.

Mercedes-Benz GL-Class reliability comparisons

StevethePilot

You don't mean the diesel intercooler hose, the diesel door handles and the diesel air suspension, do you?