Edmunds dealer partner, Bayway Leasing, is now offering transparent lease deals via these forums. Click here to see May lease deals!
Options
Popular New Cars
Popular Used Sedans
Popular Used SUVs
Popular Used Pickup Trucks
Popular Used Hatchbacks
Popular Used Minivans
Popular Used Coupes
Popular Used Wagons
Comments
Actually it's not. They don't use equal samples for each vehicle so they can't all have the same sampling error. Here in medical research we have to use equal samples when comparing say, two or more medications in order to get good results. Will it work with unequal samples? Sure, but that will raise some red flags.
USAToday wrote a column about this last year but I can't link you to it anymore because they want me to pay for it now.
Anyway the jist of it was that for 2001 CR received fewer than the required 100 completed surveys to rank the Pontiac Grand Am but yet they received more than enough to rate the MB SLK convertible. Sales for those models were 182,000 and 11,000 respectively.
Does that sound like a good sample to you?
"CR sends out questionnaries to their subcriber base and gets about 675K back. From this they report the number of problems people have with their cars. That's all there is to it. No secret or hidden agendas, no biases."
Two problems with that. One, it's not random because the same people fill out the same survey and mail it back in every single year. Two, CR does not tell us how many surveys were returned for each make and model.
You have to remember that the 675k they receive back is then sorted/divided out for each of several hundred different vehicles. A lot of those are probably thrown out for various reasons too.
"For there to be consistent bias, say against the M6, then you'd have to argue that somehow M6 owners are reporting problems more frequently than other owners."
See, now that's another problem because that CAN happen. It becomes an even bigger problem if the vehicle's survey count is low. I'm quite sure it doesn't happen often, but I'm also quite sure that it does happen.
"And their data seems to back up the anecotal data we get from boards like this"
I always like to use the Escape and Tribute as examples because they were initially rated very poorly by CR despite what owners on various boards like this were saying about their reliability. Now they are recommended even though nothing has changed since they were introduced. It took three years for CR to figure that out. That means for three years CR and board members disagreed.
I understand that many people read and trust CR and that's fine. To reiterate what was said before, it's all we've got. People just shouldn't be surprised if they get a lemon after buying a CR top pick. Their circles are only a guide, and not fact. They prove nothing.
I'm sorry to go on about this but I can't resist a good CR discussion. I'll try to stop now.
There is nothing to indicate people with higher-rated cars or more expensive cars will report more problems. In fact, just the opposite has been hypothesized - that owners of MB, BMW, Volvo, etc. cars will actually be much more picky than the owner of say, a F-150 pickup. In any event, the thing that keeps getting missed is that one should use the CR data to compare across similar cars, not as absolute measures. If one is interested in buying a sports car, then one would look at the relative data for a Corvette vs. a S2000, not the Corvette vs. the Prius.
All the other issues about sampling error are have some validity - it's not a perfect sample, by any stretch. That's not the issue. The issue is whether the sample is statistically significant in predicting the likely reliability of a given car. I think it is, and statsticians who have studied CR's data think it is.
- Mark
When people agree with them they are the "bible", when they don't it's "biased".
Anyway, local dealers here still have quite a few left over 2003 models, mostly 4 cylinder 6i's. But considering the 6 was introduced late into the 03 model year, it's not too bad.
I'll agree with you there. >= 100 surveys for each vehicle is statistically significant. I have no problem with that. But people use their data for comparisons. That's where things start to get iffy.
"The Tribute and Escape are rated higher now because their reliability data has improved. There was no "mistake": The early ones sucked, and the ones they're producing now are better."
So the magic dust is working then?
I'll just agree to disagree.
"The Tribute/Escape are just following the typical trajectory of new models - poor reliability initially improving as the factories learn how to built them better."
Then why do we need CR? We all know you should never buy a first year model because it's bound to have some bugs that the factory has to work out. All CR does for us is represent that information graphically. With little black and red circles.
Even the almighty Camcords follow that rule. You just won't see it represented in CR because they PREDICT the reliability for most first year models based on the last gen model if one exists. And they seem to work the bugs out a little faster than average.
tomcat,
Most of the '03 6i's left around here have an AT too.
Even though I'm committed to buying an '05 or '06 Mustang I still visit the Mazda lot on a Sunday every now and then to check out the latest deals and wheels. I guess you could say my test drive of the Mazda6 left a good impression on me.
Predicting reliablity is not some intractable problem that defies a statistical analysis. You process 675K surveys and you glean some useful information. It's not gospel, and it is one of many things to use in selecting a car, but it is definitely VERY useful.
- Mark
Actually quite useless. But then again you have to get the lemon to understand why, it is useless.
Not according to CR they don't but if you visit their forums you'll find out they actually do have the same little problems all other vehicles do.
You can't base that on what CR says about first year vehicles because they usually don't have enough data to use in their final rating. They will use, and correct me if I'm wrong or if they have changed their methods, a previous rating and note that they didn't have sufficient data. How is that useful? We've seen how ratings can vary from year to year so how can we expect that method to work all the time?
What's the 6 going to look like next time? With all the rust you hear about around here and the addition of a recall or two I'd expect it to rate poorly.
But does it really deserve it? Is a bad paint job on a specific batch of copies grounds for a well known magazine to move it to its "avoid" list? If Mazda6 owners actually subscribe to CR and are honest on their surveys it will.
In April, they publish a comprehensive year-by-year, system-by-system reliability report - this is the one with the columns of red, black, and gray dots.
This report is a straightforward reporting of the survey numbers. If they have a statistically significant number of surveys received for any given year/model, they'll make a report, otherwise, they say "insufficient data". The report simply gives the number of "defects" reported by owners for that particularly make/model/year, expressed as a percentage of the total number of surveys they got (e.g., 3% of 2003 M6 owners reported problems with their A/C system).
They also publish an "average chart" which allows one to see whether the reported percentage is above or below average for that vintage of car for each system (e.g., the average 2002 car had 2% of the owners report problems with their A/C system). They also report an aggregate percentage number for the entire car.
That's it - just a straightforward reporting of how many defects were reported by owners in any given car on any given year in any given system which you can compare to average numbers for cars of that vintage. No interpretation. You can object to the surveys not being representative, but you have to argue that somehow defects are overrepresnented or underrepresented for a given make/model. Frankly, I don't see why M6 owners would be any more likely to report (or not report) problems than Accord or Camry owners.
Then there are their reliability predictions which are a different animal and do have some interpretation involved. Throughout the year, they publish ratings of cars. In it they'll attempt to make a prediction of the reliability of the model they are testing. If it is a new model (or a heavily revised version of an existing model) they'll say "unknown - new model"; if it is a model for which they have data from a previous year and the car hasn't been radically changed, they'll infer the reliability from the years previous.
As you note, this may tend to penalize a new model in the 2nd year of production if it had a bad first year and the 2nd year it has improved. Such is life; a prediction can't be perfect and all you can do is use the data you have available to you.
I think they do make exceptions to their rule of not making a reliability prediction for certain makes that have a history of very reliable first-year models. Every Accord produced since it first came (25 years now) has shown good reliability, so when a new Accord model comes out, often a very evolutionary update of a previous model, they'll go ahead and make a prediction. You probably call foul - I call good judgement.
The final bit of interpretation they do is to never recommend a model for which they do not make a reliability prediction or where the reliability prediction is below average. Their philosophy on this is that in any given car category there are good reliable models, so why would anyone gamble on a car which is showing below average reliability? However, it should be noted that they never penalize an unreliable car in their relative rankings, so one can always ignore their binary "Recommended" list and go with their raw ratings. (I did this when I bought a Mercedes ML - I was willing to look past the spotty reliability of this SUV to get the SUV I otherwise wanted. And FWIIW, while I'm happy with the SUV, but I have had a few problems with it - Okay with me so far, I went in with my eyes wide open.)
Again, this methodology is not perfect, but it is the best we've got and quite a bit better than throwing darts, relying solely on a car's reputation, or relying on anecdotal data we get on forums like this.
- Mark
I've read posts from many users who bought their vehicles based on reports like CR's. Are they planning on buying a better car, like the Mazda6, in their next life?
Unless something is terribly wrong with a particular model, a la the Suzuki Samurai, I don't see why any enthusiast would question their gut feeling and buy what a magazine tells them to buy.
ramped,
This board had very little action as of late. Someone had to keep it open. I guess all of you are just happy as pigs in poop driving your 6's you don't have time to type!
(just kidding of course, couldn't stop myself!)
2. Fun to Drive
Well, maybe. It'll help if Mazda advertises them ("The Mazda6 Family") because very few people will know about them when they come out. I do think that people looking a midsized wagon will become aware of them, because there are so few out there (Passat, Saturn, uh...the Focus is smaller and the Taurus is bigger). As opposed to midsized sedans, where the 6 gets lost in the crowd.
The 6 will never sell as well as the other (you know which ones..) 2 Japanese cars, not because it is not reliable, or gets less mileage, but because it is not as large! Considering the size of the average American nowadays, you need more than zoom zoom-ability to get the public to bring you home. Consider the new Mitsu Galant! What an ugly car (for a brand new redesign)!
For the 6 to be successful, the buyer has to 1) be slimmer (rather than hefty), or 2) enjoy the driving experience (as opposed to own an appliance). Just a couple pennies...
They could call the 6 wagon an "SUV" and people would line up to buy it.
...posted from Houston, Texas. No longer the "fattest city in the US" but still the fattest state!
As for the 6 platform Mazda wanted to sell 70,000-80,000 6's last year but like Anoy said they didn't quite make it. To me Ford and Linclon variants don't count for 6 sales. The 6 is successful where I live so I think its doing ok. Remember what I said 65,000 would be a good number as far as sales go and they did match what I had in mind. Lets see how it sells this year and the next year.
I've seen sales goals anywhere from 60000 to 90000. What's your source?
I really like the silver w/sport package as my favorite combo. Not a fan of the sport grille too much. I like black but the grille doesn't stick out enough in that color for me as much as the silver shows off the front end. Silver looks good with the windows tinted out too. Not a fan of tinted out windows but works well for a Silver 6.
2) Honda Accord - 26,443 (397,750 -0.3%)
3) Ford Taurus - 15,182 (300,496 -9.7%)
4) Chevrolet Impala - 20,467 (267,882 +34.7%)
5) Nissan Altima - 13,795 (201,240 -0.3%)
6) Chevrolet Malibu - 7,972 (122,771 -27.5%)
7) Chrysler Sebring - 8,445 (100,109 -11%)
8) Nissan Maxima - 5,319 (86,758 -11.9%)
9) Hyundai Sonata - 6,053 (82,330)
10) Mazda 6 - 8,778 (66,118)
11) Saturn L-Series - 2,818 (64,957 -20.0%)
12) Mercury Sable - 4,082 (61,342 -38.0%)
- Mark
LOL mazda6s!!!
The Impala and Taurus are doing even better.
take away sales in midwest and remote areas where there are no nissan dealers and then we'll see where its at.
I'm thinking about buying an 04 M6s-AT with the S-plan. I know the $1500 rebate would come off of that price but what about the owner loyalty rebate? Has that expired or is that in play?
Any other advice for/against S-plan purchases?
Thanks in advance!
It's all good - I kinda missed it ya know?
Dinu
The Impala has been climbing up the charts in the past 2 years since so many are being dumped into rental fleets adn admitted by GM.
Here's an article that some might like where it mentions Mazda thinking of placing AWD into the M6.
http://www.detnews.com/2004/autosinsider/0401/25/b01-44608.htm
The Impala has proven to be one of the most reliable American cars. I drive one owned by my office for work fairly often, and it's OK in a utilitarian sense but it's not by any means a car an enthusiast would want to drive.
What is a surprise, despite what has been said here by some for the past year, is that the Mazda6 has done as well as it has, considering how Mazda had driven its buyers away by the mediocre 626 during the 90's.
The 6's sales in December were at a rate that would put it at 100,000 units annually, above Mazda's goal. It's taken a little while, but I think it's doing just fine (sorry Gee; I know this is a great disappointment to you).
I'm thinking about buying an 04 M6s-AT with the S-plan. I know the $1500 rebate would come off of that price but what about the owner loyalty rebate? Has that expired or is that in play?
Any other advice for/against S-plan purchases?
Thanks in advance!