Forums | MacLife

Ah, typical consumer reports.  They base their articles and ratings on people's opinions of how reliable the car is.  Not on dollars spent on repairs, not on number of issues, not on any hard data.  Its all opinion.  Opinions don't mean crap to me when I am trying to buy something.  I want the hard data and nothing else.  Gave up on Consumer Reports over ten years ago, and its painfully obvious that they haven't improved in that time.

Bren wrote:

and the best thing they can say about other VW models is that their reliability has impoved to merely average.

Volkswagen and Audi are also staging a nice recovery. The Volkswagen Rabbit (Golf) and the new CC get top scores. The VW Jetta's recommendation now extends to the diesel version, making it the only diesel we currently recommend.

and... Consumer reports' "method" of testing is ridiculous.

Our data are based on an annual survey of subscribers to Consumer Reports and ConsumerReports.org and are not derived from our vehicle testing. A model needs at least 100 responses for us to score it.

From the survey, we create a reliability history for each model over the course of 10 years, 2000 to 2009. We use the data, in part, to help readers determine which used cars to choose. We also use it to forecast how 2010 models will hold up. That predicted-reliability score is based on a model's overall reliability for the latest three model years, provided that the 2010 version hasn't changed significantly. If a model was new or redesigned in that period, we might use one or two years of data. We will make a prediction for a newly redesigned model only if previous versions had outstanding reliability.

Only 100 responses ... come on. Thats all they need.

And they only poll there own customers. Which is largely bias otherwise they wouldn't be subscribing to the magazine. They should send out surveys to everyone and let them all give input.

next thing ... they haven't even driving the cars. I sure as hell wouldn't take advice from someone who never used the product ... and with that they should have to publish what cars the employees actually drive.

Last edited by mrreet2001 (2009-11-03 12:29 pm)

yes I have and it's easy to lie with any.

Consumer reports asking it's own customers is bias. Would you consider a poll by Glenn Beck magazine asking there customers "how people feel about the president" to be a true sample of the American public?

Last edited by mrreet2001 (2009-11-03 9:41 am)

radarman wrote:

mrreet2001 wrote:

and... Consumer reports' "method" of testing is ridiculous.

Our data are based on an annual survey of subscribers to Consumer Reports and ConsumerReports.org and are not derived from our vehicle testing. A model needs at least 100 responses for us to score it.

From the survey, we create a reliability history for each model over the course of 10 years, 2000 to 2009. We use the data, in part, to help readers determine which used cars to choose. We also use it to forecast how 2010 models will hold up. That predicted-reliability score is based on a model's overall reliability for the latest three model years, provided that the 2010 version hasn't changed significantly. If a model was new or redesigned in that period, we might use one or two years of data. We will make a prediction for a newly redesigned model only if previous versions had outstanding reliability.

Only 100 responses ... come one. Thats all they need.

And they only poll there own customers. Which is largely bias otherwise they wouldn't be subscribing to the magazine. They should send out surveys to everyone and let them all give input.

next thing ... they haven't even driving the cars. I sure as hell wouldn't take advice from someone who never used the product ... and with that they should have to publish what cars the employees actually drive.

You've never taken a statistics class, have you.

He probably has, I have, and I'm not sure if you have because you should know that CR's sampling method is flawed.
Additionally, their scoring is vague and subjective, and often is contradictory and shows a bias towards Japanese and against American and European vehicles.

CR is great for appliances and suchlike where they take the product and beat the smurf out of it in a lab but anything subjective and they're crap. This is the same magazine that recommended against buying Macintosh computers for the entire 1990s because "they don't have software for Macs", the same magazine that tested consumer photo editing software, included Photoshop in the ratings, and then downrated Photoshop for being "too hard to use".

The same magazine that included "poor fuel economy" as a criticism of the last-generation Regal - the EPA ratings give it the same fuel economy numbers as the Toyota Avalon, which didn't get that criticism.

radarman wrote:

It is flawed, but mrreet was pointing out the sample size - which isn't necessarily an indication of a bad statistical process.

That was just one of the things I was pointing out.

First of all, CR does, in fact, road test the cars they write about. It's impossible to find even one issue that doesn't contain extensive auto road tests.

Secondly, how does polling their readers result in biased responses? Is it your assertion that CR readers as a group would have an axe to grind against particular car manufacturers? If so, why? The Glenn Beck analogy doesn't work, because Beck's viewers can be expected to hold specific, psychotic, hateful beliefs, because he is a psychopathic hatemonger, and that's his demographic. The same can not be said about CR. CR readers are just people who want to spend their money wisely and not get ripped off.

Thirdly, it's not that CR is biased against American auto manufacturers; it's that anybody who's ever owned or operated an American car is biased against American auto manufacturers! It's wonderful that Ford's made such strides in reliability and quality of workmanship, but even they have nearly a century of unreliable products to live down. Ditto for Chrysler and double ditto for GM.

Bren wrote:

First of all, CR does, in fact, road test the cars they write about. It's impossible to find even one issue that doesn't contain extensive auto road tests.

Secondly, how does polling their readers result in biased responses? Is it your assertion that CR readers as a group would have an axe to grind against particular car manufacturers? If so, why? The Glenn Beck analogy doesn't work, because Beck's viewers can be expected to hold specific, psychotic, hateful beliefs, because he is a psychopathic hatemonger, and that's his demographic. The same can not be said about CR. CR readers are just people who want to spend their money wisely and not get ripped off.

Thirdly, it's not that CR is biased against American auto manufacturers; it's that anybody who's ever owned or operated an American car is biased against American auto manufacturers! It's wonderful that Ford's made such strides in reliability and quality of workmanship, but even they have nearly a century of unreliable products to live down. Ditto for Chrysler and double ditto for GM.

So that justifies flat out bias like reporting a GM vehicle to have "poor gas mileage" and not a Toyota that has the exact same numbers? Or saying that a Volvo has "poor leg room" when the measured leg room is better than the Acura that gets praised, and in the same review the Volvo is "sluggish" and the Acura "peppy" even though the objective acceleration numbers were the same? How about the validity of a polling system that gives results like one Oldsmobile having "Excellent" manufacturing quality and a different model that rolled out of the same plant as "Poor" manufacturing quality? That means that the sample size is not large enough. Or a Chrysler having a "harsh" engine, but not mentioning that about a Mitsubishi with the exact same engine?
How about their blanket policy of praising Toyoda drivetrains even with the notorious Honda transmission and Toyota sludging of the 90s?
On the flip side of wildly varying ratings for what are practically the same car, there is lumping different cars together like rating my Sebring the same as a Sebring sedan, even though the only thing they have in common is the name. One is the same as a Stratus, the other is similar to an Eclipse or Galant.

No, CR has smurfy statistical methods and clear bias. Read JD Power, read Edmunds, read Car & Driver, read Motor Trend, but don't bother reading CR for reliable car information.

Bren wrote:

Where can I find JD Power and Edmunds data on any given model's predicted reliability?

You can look at their past reliability and guess just as well as CR does. Maybe even better if you cover up the make and model so you aren't biased.

Bren wrote:

Well that's what I mean. Where can I find this? Do they have fun 'n' easy-to-read charts, like CR does? Got some URLs for me? I'd Google it, but I think Google's closed right now, 'cause it's after 5:00PM.

JD Power has two studies each year - Initial Quality, which is driving impressions and problems after the first 90 days and the Dependability survey which is based on the first three years.

2009 Dependability Survey, which ranks 2006 vehicles.

Bren wrote:

So...

Just by brand, as opposed to individual model ratings? And just for new cars?

Geez, I gotta do everything for you?
Dependability ratings by category.

It erases your hard drive.