By accessing this website, you acknowledge that Edmunds and its third party business partners may use cookies, pixels, and similar technologies to collect information about you and your interactions with the website as described in our
Privacy Statement, and you agree that your use of the website is subject to our
Visitor Agreement.
Comments
Those data are not fact. I'll keep repeating it until you let it sink in. :P
They are relying on subscriber survey data to make their conclusions. Do they have a standard way of determining what sort of problems constitute an unreliable vehicle? Or even a standard definition of a problem for the surveyed to go by? I didn't think so. You can't group a clogged windshield washer tube with a failed transmission then fill in a little circle and call it a day.
Then we have the issue of sample size. CR does not report the number of returned surveys for each vehicle for some insane reason. Some vehicles, like the Pontiac Grand Am when it's sales were 200k+ 3 or 4 years ago, don't even receive enough returned surveys to be rated. In statistics, the smaller the sample the more chance there is for error. What's the minimum # of surveys returned to be rated now? IIRC it was as little as 100 according to the article my Grand Am example appeared in (USA Today). Their other example was some MB model which I can't remember. I'd really like to know if other poorly rated vehicles have such low survey returns too. Maybe one of us would have to change our minds! :surprise:
Given the above statements, don't you think it's a little off-base to call CR's data fact when they don't even have the readership to make fair comparisons? Think about that Grand Am for a minute now. Say CR did receive 100 surveys for it and could rate it. Don't you think it's possible that 100 people could have signed up for a CR subscription because they hated their Grand Ams and wanted to be "better informed" for their next purchase? What then happens to that rating when 50 of them report 2 or more problems? What about the other 200k+ Grand Am owners? Did they all have 2+ problems too? Does CR know that answer? No they don't, but GM does.
Once more, if you want the facts about how many problems a particular model has had then ask the manufacturer. They have the EXACT numbers whereas CR offers us an educated guess. The latter is not fact just as often as it is. In order for a statistical result to become fact it has to be proven over and over again using the same methodology. CR's methodology varies from year to year and therefore can never, ever be factual.
I asked you for evidence, but all you did was boast about being an uninformed buyer.
How does my not reading CR make me an uninformed buyer? Do you mean to tell me it's the only source of information out there? Boy, where have I been all these years? :surprise: Here I've been buying vehicles because I like them, and driving around with a big smile on my face when I could have bought what a guy who tests toasters told me to buy. Thank you for introducing me to this whole new world! I'll see you at the cliff with the rest of the lemmings some day. :sick:
I'm boasting? I'm very sorry that you're so jealous of someone who can make a decision on his own but that's the way the world turns. It's also funny how you can say that about me when you continue to tell all of us how smart and informed CR readers, like yourself, are. We're talking about buying cars here not which asteroid defense system is going to save the most lives in the event of an asteroid shower.
I forget now. Am I the pot or the kettle?
Sorry, but going out of your way to avoid factual information isn't a tribute to your methodology.
Oh, I don't avoid factual information. I just haven't seen any yet. Have you?
That may be true regardless if one buys a Chevy or Honda. It would have to be some kind of incredible coincidence that average Chevy owners treat their cars statistically worse than average Honda owners. There may be some significant differences between say owners of $40K+ and those of $15K+, but within the same segment - come on! However, if this was really the case and Honda people actually were taking better care of their vehicles than Chevy people, it would be an argument AGAINST owning Chevy, rather than for it. If I were to buy a new product, I associate myself with careless people who don't change oil in their vehicle (not a big deal if I were to own that car for next 10 years, but a real potential monetary loss if I were to trade it after 3-5 years), OR why should I buy a used car from those kind of people?
2018 430i Gran Coupe
I used to theorize that many people buy a Japanese car just thinking that it's automatically more reliable than a domestic, but somehow get it in their head that that means that they don't have to take care of it. And hence the problems when it does break down.
But if that were the case, and masses of people were neglecting their Hondas and Toyotas, you'd think that in the long run that would bring their reliability ratings WAY down. But it doesn't.
2018 430i Gran Coupe
Instead, you raised a series of rhetorical questions without answering them. A question is not the same as a factual response.
Do they have a standard way of determining what sort of problems constitute an unreliable vehicle?
Is it standardized? Yes, all respondents are asked the same questions, so the method is standardized.
Your definition of "standardization" is different -- you want to exclude problems that you've determined aren't important to you. (I presume that you must believe that the domestics would require a lower bar to hurdle in order to get a better ranking, an ironic position to hold.) That would be vastly more biased than it would be to simply ask people to indicate the problems they have had in each category.
Some vehicles, like the Pontiac Grand Am when it's sales were 200k+ 3 or 4 years ago, don't even receive enough returned surveys to be rated.
If you've read anything in this forum, then you should know by now that most of these cars are bought by fleets, not retail consumers. If the survey size for a given model isn't adequate, then there is no reason to include it.
In statistics, the smaller the sample the more chance there is for error.
And the converse of that is that a survey with a large enough sample size should produce data that is reliable within an acceptable standard error. A sample size of 100 should be good enough to generate a usuable response.
Lets see someone buys a Japanese car doesn't take care of it, it breaks down and its a sign of how reliable the car is. Now when someone buys a GM or Ford and doesn't take care of it and it breaks down its a sign that its a POS. Hows that again?
Here is my theory, perception creates it own reality. Person 'A' believes that Toyota makes bulletproof cars and buys one. Person 'A' in two years has taken their Toyota back for "issues" three times yet goes around telling people how great and reliable their car is. Now Person 'A' knows person 'B' who bought a GM that has only one minor issue in two years. At that point person 'A' tells person 'B' "oh you should have gotten my car it is so much more reliable.
2011 Hyundai Sonata, 2014 BMW 428i convertible, 2015 Honda CTX700D
90,000 miles was when the tranny started to go. And it needed exhaust work as well. And some other issue or two that on an older car would have been either minor or for a part that hadn't even been invented yet!
The Stanza got replaced by a 1999 Altima, which I thought was an ugly, grubby looking little car. Uncomfortable, kinda cheap on the inside, just mediocre all over. But my stepdad just loved the thing. I remember when I bought my '00 Intrepid, my stepdad was looking under the hood of it and saying "Nissan does this better, Nissan does that better, this isn't too bad but Nissan still laid their out better, etc). Well, one thing Nissan DIDN'T do better was the transmission in that car, because it failed at 35,000 miles!
And c'mon, I drive an INTREPID! Isn't it almost mandatory that the trannies fail in them? :P
Now, to be fair, Nissan replaced that tranny under warranty, with no fuss. And now the car has about 200,000 miles on it, and has been pretty reliable since then, so ultimately it did get to prove itself in the long run.
But still, Consumer Reports, anal maintenance, and all the research in the world won't always protect you from being stranded on the side of the road.
It's been repeated often but a few just don't get it. Keep trying though, it's been done in several different forums.
>they don't even have the readership to make fair comparisons?
We don't know about who the subscribers are let alone how few of them respond and what their individual purpose in responding is.
How many here subscribe to CR? I don't. I check it at the library occasionally. I subscribed one year and they didn't send me a questionnaire. I guess they saw I subscribed to Automotive News and wouldn't fit their demonographics and purpose.
Let's do a CU type poll. Those who subscribe to CR voluntarily make it known here. Return my questionnaire here. Does socal subscribe?
2014 Malibu 2LT, 2015 Cruze 2LT,
According to this link, CR surveyed five million subscribers. That's a sizable pool by any measure.
I will admit I subscribe to the online portion of CU. I don't get the magazine, but I have access online. I also subscribe to Fortune, Money, Motor Trend, Road & Track, Car & Driver, Popular Science, Trailer Boats Mag, and the WSJ,plus the local paper, no shortage of reading material around here.
My wife has had 4 company cars since 2001, we have not filled out a JD Powers or CU survey on any of them that I can recall.
Since I'm a CU subscriber I guess I throw the demographics off because my wife and I have owned 10 domestics, 1 Asian, and 1 German, vehicle in the past 19 (well we've only been married 10, so I'm counting my cars back to when I was 16). That does not count my wifes 4 company cars since we didn't buy them.
With all of these cars, CU's info has been consistant with the results I experienced with those vehicles.
I just don't see how CU is any more biased than JD. Does JD even drive any of the cars they report on?
Let's face it -- the CR foes don't really have any evidence to back them up. At best, it's based upon speculation (generally from people who don't even read it), and the few rebuttals that they do offer are poor. Think about it:
-CR received more than one million responses to its surveys last year, yet the opponents would have us believe that the pool is not large enough.
-CR sends surveys to five million households, yet the sample is allegedly inappropriate (as if you could ever get five million American families to agree completely on much of anything.)
-The survey results are generally consistent with other sources, but that consistency is ignored.
-CR allegedly pans domestic cars for the sake of it, even though it is on record as having recommended some of them.
It comes down to this -- CR says stuff that some people don't like, and those people choose to attack the messenger, irrespective of the quality of the data. They reject a survey with one million responses if it contradicts their own personal experience, opinion or wishful thinking about a given product.
Again, I would like to see some evidence that there is a bias in the survey results that renders it invalid or inaccurate. Not a single shred of evidence has yet to be provided, the only response is for these same people to just restate their opinions or to pose rhetorical questions as if those substitute for evidence. I'd welcome a well-documented answer based upon solid reasoning and data, but I don't expect that we'll be finding one here any time soon.
And I am telling you that CR and the like do not provide us with facts. They all have their own flaws and differences but the one thing they all have in common is that they are not stating facts.
Did you happen to catch the news a year or two ago about the Hummer H2 and it's poor JD Power ratings? In summary, people listed the poor gas mileage as a problem and put the H2 in the ratings basement. I'd be willing to bet that other things like rattles and squeaks are reported as problems on these surveys too.
Is it standardized? Yes, all respondents are asked the same questions, so the method is standardized.
No. You just told me that the form is standardized not the method. They are two completely different animals.
If you've read anything in this forum, then you should know by now that most of these cars are bought by fleets, not retail consumers. If the survey size for a given model isn't adequate, then there is no reason to include it.
So out of 200,000+ Grand Ams sold fewer than 100 of them were delivered to non-fleet customers? What about the Mercedes then? Were they all sold to fleets too?
A sample size of 100 should be good enough to generate a usuable response.
Not when it is being compared to another vehicle that may have had a sample size of 10,000 or more it's not.
I counted the number of models available right now (here on Edmunds) and the total is roughly 336. All things being equal, each model from each manufacturer should have received roughly 3000 surveys from subscribers. Yet some vehicles don't even get the required 100. Where did the rest go then? A well-documented answer based upon solid reasoning and data would be welcome here too.
FWIW 3000 is a pretty good sample size when done right.
That is not a fact, that is a repetition of your opinion. Rather than restate your thesis, prove it -- repeating it just an effort to evade providing the evidence that you are obliged to give.
You just told me that the form is standardized not the method.
The method is also standardized. CR uses its mailing list to send out surveys to five million households, and gets back about one million responses. Compare that to typical opinion polls by firms such as Gallup, that can draw reasonable conclusions based upon surveys of but a few hundred people.
So out of 200,000+ Grand Ams sold fewer than 100 of them were delivered to non-fleet customers? What about the Mercedes then?
If you are going to discuss statistics, you really ought to do the math:
1,000,000 respondents
divided by
Approx. 1500 models (250 nameplates X 6 years of data)
equals
An average of approx. 667 responses per nameplate.
Obviously, that's a rough approximation, and there will be variation based in part of the popularity of certain models of cars, i.e. there should be more respondents from Honda Accord owners than there would be from owners of 3-series BMW's. However, in both cases, there should be enough data from each to render the data to be accurate.
So fine, let's look at your Grand Am. According to Fleet-Central, here are the retail sales figures for the first half of MY 2005:
Car / Retail sales / Percentage of sales to fleets
Grand Am / 9,657 / 71.5%
Accord / 160,522 / 2.1%
Mercedes C-class / 26,381 / 8.6%
Obviously, Grand Am retail sales were pretty low, about 6% of the level of Accord's retail sales, and even about one-third of a Mercedes C-class. So it's quite understandable if the Grand Am didn't generate enough responses to merit reporting it.
"52% of regular Fox News viewers consider themselves conservative — 16 percentage points higher than the population as a whole — according to a survey conducted by the Pew Research Center" -USA Today
Consumer Reports is simply the best source for honest, reliable, accurate information regarding vehicle dependability.
Any source which doesn't randomly buy their own cars, is unacceptable. Car manufacturer's most definitely "quality control" their "test" cars more so than their regular average run of the mill vehicles that the average consumer would buy.
I believe Honda, Toyota, and the other reliable Japanese auto manufacturers have a high level of quality control for all of their vehicles. I believe the domestics have a very low level of quality control. I think this explains in part why CR may appear to be harsher on domestics than other sources.
2014 Malibu 2LT, 2015 Cruze 2LT,
CU says they surveyed 5 million subscribers. They didn't get 5 millions responses. Maybe 1 million on a good day?
Also CU played games with the response numbers when some were supporting it in another discussion. They counted each car reported as a separate response rather than counting the number of people responding to their convenience survey.
2014 Malibu 2LT, 2015 Cruze 2LT,
They have sent me surveys in the past, but I haven't seen one in a while. I know it has been a year or two. I guess it's possible my wife or myself accidentally tossed it.
Or maybe I have to many domestic vehicles. LOL.
I'm glad to see there's still a sense of humor here! You probably rated your domestics too well on the earlier surveys so they took you off their list.
2014 Malibu 2LT, 2015 Cruze 2LT,
According to this link, CR surveyed five million subscribers. That's a sizable pool by any measure.
**********************************************
You just answered your own query. The surveyed 5 mill SUBSCRIBERS. Therefore, their survey is based solely on the fact that you have to subscribe to their magazine to be counted. That is biased. They've skewed their results to only include a certain demographic. I'm really not sure how you could NOT call that biased.
'11 GMC Sierra 1500; '98 Alfa 156 2.0TS; '08 Maser QP; '67 Coronet R/T; '13 Fiat 500c; '20 S90 T6; '22 MB Sprinter 2500 4x4 diesel; '97 Suzuki R Wagon; '96 Opel Astra; '11 Mini Cooper S
2018 430i Gran Coupe
The reason people have a bias against a manufacturer is because they have either been directly burned by them, or know someone personally that has been burned.
If you sell lemons, only an idiot will let you sell them again.
I already told you that the exact numbers are in the hands of the manufacturers and therefore anything else is just an educated guess. What more proof do you want? I've also given you several other reasons why those data are not fact but you keep repeating that I'm just repeating myself. When are you going to prove me wrong I wonder?
The method is also standardized. CR uses its mailing list to send out surveys to five million households, and gets back about one million responses. Compare that to typical opinion polls by firms such as Gallup, that can draw reasonable conclusions based upon surveys of but a few hundred people.
How does any of that prove that the method is standardized? All I know now is that the form and whom they mail it to is standardized. That tells me nothing about how they reach their conclusion, little red and black circles if you will, based on the results of those surveys.
If you are going to discuss statistics, you really ought to do the math:
1,000,000 respondents
divided by
Approx. 1500 models (250 nameplates X 6 years of data)
equals
An average of approx. 667 responses per nameplate.
Since I do not read CR the 6 years back was not accounted for in my EXAMPLE. Yes, you are correcting and critcizing my example.
Your supposedly accurate math above shows us that each model should average 667 responses which would be great and a big step in the way of these data becoming fact. However, that average does not apply to most, if any model and we're stuck with high margins of error on some models.
So fine, let's look at your Grand Am. According to Fleet-Central, here are the retail sales figures for the first half of MY 2005:
As stated previously my Grand Am example was from three or four years ago. The Grand Am is dead now so obviously it doesn't apply anymore. I think the article was for MY02 Grand Ams but it could have been MY03. I guarantee you fleet sales were now 71.5% in either of those years.
Does CR still have some models marked as "Not enough data for review" or something like that? If so, what are they?
It's not a random poll. It's called a convenience poll IIRC. The group is whoever subscribes to CR. Could take a poll of MT subscribers and use their results just as easily. But its meaningfulness is not that of a random poll. If CR called households around the country based on population and collected 800 results on each car model and year, they would be closer to a random sample with a meaningful accuracy.
They don't. And their personal opinions show through. Even when "buying" the cars on the open market, I don't trust them not to be pickier about the prep by the dealer; they are often so critical of little details on the car, i.e., delivered with low air pressure in right front tire--a dealer item, not a manufacturer defect.
2014 Malibu 2LT, 2015 Cruze 2LT,
True, but reflect on manufacturer, as it is an indication of a problem with quality assurance process with their retail network.
2018 430i Gran Coupe
I should also note that CR did rate the Caddy STS slightly above the Acura RL and Lexus GS 300. The STS recieved an average score of 78 in CR while the RL and GS 300 reciveed a score of 76 and 75 respectively. Cr also liked the Ford 500's ride but hated the engine because they claimed it was underpowered. Heck CR even liked the interior of the Ford Fusion and Buick LaCrosse listing it as one of its positives for both cars. CR even rated the Chevy Malibu 4 cyl sedan slightly over the Nissan Altima(4 cyl.) Based on what I said it doesn't like that CR is biased. Lastly CR did rate the Buick LaCrosse reliability wise over the Toyota Avalon. The Pontiac Grand Prix in CR's ratings was as reliable as Honda Accord V6. CR couldn't reccomend the Grand Prix because they called it a mediocre sedan.
I agree as Honda fan the dash theme needs to be watered down. I just don't like the digital guage at all.
You can repeat yourself all day long, but this is a false statement. JD Power is in the business of selling data, in part because automakers do not have complete information.
All I know now is that the form and whom they mail it to is standardized. That tells me nothing about how they reach their conclusion, little red and black circles if you will, based on the results of those surveys.
It tells you nothing because you haven't conducted basic research. The red and black circles are clearly designed to provide a lot of data quickly in a user-friendly manner. You are confusing the presentation, which is simplified for common readers, with the methodology behind it.
I think the article was for MY02 Grand Ams but it could have been MY03. I guarantee you fleet sales were now 71.5% in either of those years.
Rather than "guarantee" something without data, I'd prefer the data. The Grand Am (and now G6) have both comprised a significant component of GM's fleet sales (the G6 is a Top Ten fleet seller), so there's not much reason to believe that it was any different a few years ago than it is now. After all of the posts about fleet sales throughout this forum, this dominant trend should be fairly obvious.
This is a no-brainer: the survey asks for responses from owners of late model vehicles.
-Not everyone surveyed has a late-model vehicle that would qualify them to respond to the survey
-Not everyone qualified to respond to the survey bothers to respond.
Unless you are silly enough to think that all five million households have at least one vehicle new enough to merit a response and that every single person answers the survey, this should be pretty obvious. Having designed a few surveys myself, I can tell you that there isn't a survey on the planet that includes a 100% response rate.
For your point to have any merit, you would have to demonstrate a couple outlandish points:
(a) The ownership experience of CR readers would have to be much different than those of the general population, i.e. CR subscribers who buy domestics end up with a disproportionate number of lemons, while CR subscribers with imports have significantly better experiences than do the general population
(b) CR subscribers are significantly more dishonest than are the rest of the population
It's pretty obvious -- sending the survey to subscribers is the cheapest way to reach a large number of people. I suppose that they could spend $30 million per year sending surveys to 100 million US households, in the hope that they reach those who own late-models AND that they respond to the survey. This would obviously be costly and would yield a lower response rate, particularly because the contact information would likely be less accurate. (Since CR has good mailing addresses for its subscribers and a relationship with the recipient, the response rate should be much higher.)
But there's no reason to believe that the results would be much different from that than they would be by polling the mailing list that they already have. The survey isn't so much about the owner's subjective opinions as it is about obtaining a list of repairs and defects. Unless the owners of one class of car are significantly more dishonest, stupid, or deceptive than the general population, there is no reason why you shouldn't take the data. Counting problems per number of vehicles is a standard survey method, there is absolutely nothing wrong with the methodology in that respect.
But there's no reason to believe that the results would be much different from that than they would be by polling the mailing list that they already have. The survey isn't so much about the owner's subjective opinions as it is about obtaining a list of repairs and defects. Unless the owners of one class of car are significantly more dishonest, stupid, or deceptive than the general population, there is no reason why you shouldn't take the data. Counting problems per number of vehicles is a standard survey method, there is absolutely nothing wrong with the methodology in that respect.
I agree with you Socal. Sure it would be nice to send a survey to every car owning person in the US, but I think my $20 subscription would have to be increased about 100x to meet the added costs;)
I don't think this debate is going anywhere. I've yet to see any concrete evidence that CUs results are less reliable than JD Powers and those against CU are not going to change their mind either.
They are in denial! They are not using their brains.
Demographics do play a role, I know very few people under 50 or so that have a subscription to CU. Interestingly enough, that ties in with the mean age for Toyota buyers.
"For your point to have any merit, you would have to demonstrate a couple outlandish points"
It's called Recall Error. Its the same thing we run into when I was in graduate school doing accident surveys. People can't remember exactly what went wrong and when, and they don't double check.
In addition, part of the problem is the tendency to confuse CR's car reviews with the reliability surveys.
The reliability surveys are what they are, they are based upon owner's lists of problems and defects. Pretty good data, and a useful starting point to avoid some obvious dogs with fleas. I have yet to see either here or anywhere else illustrate how this is flawed.
CR reviews may be a different matter, because they are based upon a blend of the objective (reliability data, crash test data) and subjective (their judgment of the driving experience, packaging, etc.) They also tend to regard cars in a somewhat appliance-like fashion, so it is possible that their tastes don't match with yours.
And if you don't care so much about reliability or crash test results, you may find your favorite car is not on their recommended list. (For example, CR removed the Focus from its recommendation list because of crash-test data, even though it was their favored choice in their subjective testing.) If you find that you differ with CR's reviewer as to what constitutes a good driving experience, then do the easy thing -- don't take their subjective opinions into account. But that's no reason to ignore the reliability survey.
So you are asserting that Toyota and Honda owners have terrible recall that benefits them, and Dodge, GM, Ford, Chevy owners have terrible recall; that harms them?
It's funny that some people can make seemingly logical arguments, but still forget the fact that their own argument favors believing CR completely.
There is an element of self-selection in any survey. The fact that many of the cars surveyed exhibit average results without extreme outcomes in the data would indicate that your supposition is invalid.
Demographics do play a role, I know very few people under 50 or so that have a subscription to CU.
That should not impact a reliability survey. If the car has a failure or requires some sort of repair, except if that repair was self-inflicted (i.e. running the car into a wall), the repair need is based upon the vehicle, irrespective of the user. I don't see any data that indicate that the reliability of an Accord is considerably different for a 35-year old owner than it would be for a 55-year old owner.
It's called Recall Error.
So you're claiming that recall error rate of Honda Accord owners is dramatically different than that of Cobalt owners, while Focus owners come out somewhere in between?
Over a broad pool of respondents, the recall error rate should be the same for one pool that it would be for another. I doubt that the Honda owners are more forgetful or dishonest than are the Chevy owners. That's a pretty weak theory to hang your hat on.
I am one who doesn't trust CR further than I can throw a bull elephant.
More than once I have bought crap that CR highly recommended that simply just was junk.
So tell me why should I change my mind about CR?
More than once I have seen CR do comparisons giving brand 'A' as the best of product 'X' yet fail to even mention in their reports other brands which everyone else rates better than brand 'A'.
So tell me why should I change my mind about CR?
More than once I have seen CR give company A's product high marks and give company B's product poor marks while all the while both companies product are the exact same thing coming off the exact same assembly line made by the exact same producer, the only difference is the name.
So tell me why should I change my mind about CR?
There are other things that make me distrust CR but I will leave it at that.
They are in denial! They are not using their brains.
Nope no denial here, I used my brain saw the evidence and made a much informed opinion.
2011 Hyundai Sonata, 2014 BMW 428i convertible, 2015 Honda CTX700D
I don't think what you are saying is quite correct.
- I'm certain that there are clear demographic trends among CR subscribers.
- People (i.e. the "average" consumer) are notoriously bad at objective opinions. Personal expectation plays a HUGE role in such evaluations.
Obviously, for various reasons, CR can't eliminate all bias from its results, but you do have to take them with a grain of salt.
From my personal experience, I find that I tend to disagree with many of CR's recommendations/statements, etc. I believe this is a direct result of the two issues mentioned above (plus my own biases).
When you say product 'x' and brand 'a', are we talking toasters and electronics, such as receivers, speakers, TV's, mircrowaves?
I agree, in terms of electronics products, CR needs to test a whole lot more brands and models up and down the lines coming from manufacturers. I found their year end book to lack way too many popular brands and models. Therefore, what they are recommending as the best, may only be the best out of 25% of what is "out there."
But.... When it comes to cars, I think they covered a vast majority of what is available "out there." I don't think they've missed many models.
Two things come into play. First is perception. If you pound into someone that one car is very reliable and another is very bad then their response will be somewhat biased that way. Since the public has been bombarded with how great Japanese cars are and how bad American cars are people will expect that and report that. In other words people who own Japanese cars would be more likely to view a couple of issues as minor since they believe their cars are better. People who have GM or Ford may view the same thing as more of an issue because they perception they have of them.
The second thing is in surveys people have a tendency to be biased in the direction they think the survey taker wants them to be.
2011 Hyundai Sonata, 2014 BMW 428i convertible, 2015 Honda CTX700D
But.... When it comes to cars, I think they covered a vast majority of what is available "out there." I don't think they've missed many models.
One of my examples was about cars.
2011 Hyundai Sonata, 2014 BMW 428i convertible, 2015 Honda CTX700D
That still doesn't show how they bias the survey towards Japanese car makers. As far as I know, CR sends out the exact same survey with the exact same questions to everyone, and every body. If they had questions like "So how many breakdowns did your domestic car have?" and excluded the same question regarding import cars... that might show bias... show me!
I have a Honda Accord currently, I expected bullet proof reliability, as you state in your case. I disagree with all of your conclusions 100%, however. My expected bullet proof reliability made me much harder and harsher on Honda when I experienced a few minor, and one major problem in my first 56K miles. I had to step back and put things in perspective..... My Honda has had HIGH average reliability, good, but not excellent overall. For a Honda, this is poor, but compared to all vehicles out there, its still on the high end of average. For a time, I felt like Honda gave me a lemon, but then I remembered my old 95 Dodge car. Incomparable. As annoying as it was taking my 03 Accord to the dealer many times for warranty work, I took my 95 Dodge many more times. As annoying as rattles, and minor cosmetic defects are.... I never had to tow my Honda, and was never stranded. 95 Dodge had to be towed 4 times, and I do mean had to.... no choice.
Another point people try to make is reliability depends on maintenance, or lack thereof, and how easy you drive a vehicle (or how hard). I don't think that carries much weight, because.... I babied that Dodge and got oil changes every 3K miles.
My friend raped (yes, raped is a good word to describe it)his 94/95 Geo Prizm (Toyota Corolla with grill and emblem removed) for over 100K miles, with untimely maintenance to say the least. He floored it everywhere he went. He raced it, he tortured it, he even crashed it eventually. His repair costs for 100K miles (no including body work) = $0.
You couldn't make that car break down if you tried. And you couldn't make that 95 Neon run dependably no matter how much money you put into it.
The survey is not rooted in subjective questions, i.e. "Have you had a lot of problems with your car?". The respondents are asked to provide a list of their repairs and defects, and whether the service done to perform the work solved the problem. The resulting percentage of defects are put into a defects per hundred calculation, which gets translated into a colored circle
That's about as objective as it's going to get. If the air conditioner on my car failed, I would indicate that the A/C stopped working. That counts as a defect, regardless of the car that I own. That response should not vary based my age, gender, or income, just so long as warm air blew out of my dash when cold air was required, and I remembered to tell the surveyer about it.
And I would expect the memories and honesty of various pools of owners to be comparable. One purpose of getting a sample is to reduce the noise produced by a few outliers who have amnesia or who prove to be habitual liars, with most respondents in the pool providing reasonably useful information.
I disagree. I've owned one asian made vehicle. An 01 Nissan Pathfinder LE. Yes it was rated well by JD POwers, CU, Edmunds and many magazines etc, but I didn't buy because of those reports, I simply bought it because I liked the way it drove compared to a Trailbazer, Explorer, Tahoe etc.
I did happen to have two problems with it. A faulty 02 sensor @20k miles(under warranty) and a faulty intake valve timing sensor @ 53k miles(or something like that, not under warranty) I was not happy at all and I did report those problems when I was sent both JD and CU surveys. And quite frankly I was more disappointed that it happened since at that time, Nissan had a solid reputation. Granted I still liked the truck, and those problems only caused a check engine light w/o any performance issues, so it rightfully should be considered minor and cost me little.
That's a little different than the brand new '98 Ford SVT Contour I bought died on me at 4000 miles (a simple short to the fuel pump) out in the middle of no where and I had to wait 2 hours for a tow truck. Ironically, that car gave me all sorts of problems, but I still enjoyed the heck out of it. If anything I tend to be more critical if my expectations are higher and I'm let down.
Perhaps it would help if you provided examples of what questions are asked on the CR survey. Many posters here seem to have the wrong idea about what kind of questions are asked, and how they are posed.
TIA.
Then of course there is the resale value. You are simply throwing money out the window by buying domestic.