Supercharger

J.D. Power Initial Quality Study Awards


Premium Luxury: Lexus LS 430

Mid-Luxury: Lexus GS 300/430

Entry Luxury: Lexus ES 300

Compact: Toyota Prius

Entry Midsize: Chevrolet Malibu

Premium Midsize: Mercury Sable

Full Size: Ford Crown Victoria

Premium Sports: Mercedes SLK

Sporty: Mazda Miata


Report: Fall 2003 issue J.D. Power Car Guide

lurcher

So, still no challengers for the Miata after all these years?

How long before the RX-8 can make an appearance - cars have to be out for 3 years or so before JDP list them right?

mikeb

wow
lexus is whooping some ***

Squidward

companies like jd powers has no credibility.. they come up with these lists as if it was the be-all end-all rating.. whatever.. i wouldn't be surprised if auto manufacturers "court" them constantly (if not down-right pay them under the table).

of course there are legal terms they always inject into their ratings, such as "best of it's class" however their definition of class changes from car to car...

Elara

They also just measure how many complaints new owners had- and that means EVERYTHING- from the poor gas mileage on the hummer (DUH) and the lack of cupholders on the minis. So it's not a good test of anything.

FamilyGuy

For instance, if I recall correctly, JD Powers ranks reliability by initial build quality - i.e. complaints the owners have in the first few weeks.

Even the worst car companies are pretty good about catching the really obvious problems before letting someone drive the car off of a lot. The real test of a car's reliability is the number of problems it has over the first four or five years - and you have to look elsewhere for statistics on that.

Quick_lude

JD Power also has long term dependability tests/list. I believe Lexus is on the top followed by Acura, Toyota, Honda and Nissan not necessairly in that order.

FamilyGuy

I sit corrected.


[EDIT] Still, I'm surprised the Chevy Malibu is on the list. The 2003 model was a middle of the pack performer. The 2004 redesign has gotten decent reviews so far, but since it is brand new there are no reliability or crash test numbers yet.

zoom44

what's with the "9" best? they thought "oh everybody does 10 best so lets be different!"?

rloewy

Actually, JD Power have a lot of credibility if you understand the statistics behind the numbers and know how to interpret them.

The study specifically tells you that this is initial quality survey (based on problems within 90 days of purchase) and based on a scientific survey all they do is tabulate the information and present it. It is not the final answer to any question you might want to know about car quality - it is just a simple very specific piece of information that is clearly explained.

The problem with JDPower is that often people that have no clue of basic math try to interpret the information and blame the messanger.

I would suggest reading some of John Allen Paulos books as a good introduction to math applications to real life that does not involve actually understanding the math behind the numbers.

Ron.

blizz81

Math or no math, if anyone wants to believe that the Malibu, Sable, and (well, perhaps more fathomable on this last one) Crown Vic are at the top of their domains, then go ahead.

Don't count me in though.

rloewy

OK. No Prob. I assume that next time you need some help fighting illness you plan on treating it with chicken soup, because any FDA approved drugs are tested (and approved) using exactly the same statistical methods.

When JDPower started (was it the 70's or the early 80's) and claimed that Japanese cars had much better reliability than American counterparts the same kind remarks were made. In the long run - if you understand the math - the numbers do not lie. (If you read the numbers after they have been interpreted by "marketing" or do not understand what the information really is and how it was collected - it's a different thing).

Ron.

mikeb

I dont understand your analogy
but I have always heard good things about JD

rloewy

The analogy is as follows:

You either belive that the research was done scientifically and thus there is a high confidence that it is accurate or you do not. You can not choose to belive the part about the Lexus and not the part about the Chevy.

Since he decided not to belive it despite the scientific research ("math") - he should not belive the FDA when it tells him that certain drugs are safe and will help him - because drug approval is done using exactly the same kind of statistical methods - only with higher confidence level (read - money).

(Drugs are generally approved by sampling against a placebo and applying probability methods to determine if your assumptions are correct)

The only "drugs" that do not require FDA approval are the "granny" kind - read "chicken soup".

Hope this clarifies the analogy.

Ron.

U. N. O.

thats that count include all cars sold regardless? i mean rental car companies buy tons of cars.. that might trhou off the numbers..

mikeb

thanks ron
I get it now

blizz81

Geez...my issues are not with the math involved (I have my own separate issues to mathematics/mathematicians that regard the study as All Holiness that lie more on the theoretical + philosophical side) - just with how the information is thrown to the public in such a way that it's easy for people to take it to be something it is not. I do not follow the JD & Power executives, lackeys, etc, around that are involved in the process of creating a "Initial Quality Study", so I cannot vouch for their entire method down to a T, nor do I closely study their methods as they release them to the public...but to me, surveying a sample group of car owners and asking them what problems they've had in the first 90 days of their ownership (which includes "I don't like the cupholders even though I knew that going into my decision to buy" and "I don't like the gas mileage of this Hummer H2") of a car doesn't really advance any kind of noteworthy conclusion to me. Maybe I'm way off on their methods of collecting information...

Something that would be better in my eyes, but not conclusive, would be collecting information from dealerships on warranty repair/replacement work. There's room for misguidance/error/etc here too, but not as much.

rloewy

This is a much more reasonable answer than the "math or not" response give earlier.

I still disagree with you on the following subjects:

1. How it is "presented" to the public - the information is out there - you just need to read the detail - they made a "ease of use" feature by displaying the best in each category they sampled - but other than that - if Joe Public does not take the time to read the detailed information - he has only himself to blame for misunderstanding.

2. While your complaint about the fact that the public at large considers problems YOU do not consider as problems is valid - it is however, the only way to really perform such a sample. Some people on this very board were upset that the car lost a whopping 11hp from what was on the spec sheet - yet there are a lot of others that could not care one bit. By taking a large sample you basically "average" these differences and get a very accurate average. (Assuing a large enough sample the central limit theorem, IIRC, actually ensures that the population mean will be the same as the sample mean).

3. The idea of surveying the dealerships is interesting for a repair survey but it does not provide you with true "quality" picture because of the following:

3.1 - The dealership has only information about repair - it does not include information about general design issues, stuff that people find problematic but not enough to bother with wasting the time to have it repaired.
3.2 - The dealerships are not the only source of repairs - some people work on their own cars, other take it to independent mechanics.
3.3 - This is the biggie - the dealerships are not an unbiased sample - they have an interest in the result because they also sell and service these cars and make money from them - so they will want to display it in favorable light. The car owners (at large) on the other hand - do not have this liability - thus scientifically an owner's survey is more reasonable.

Finally - please notice that you are confusing the JDPowre initial quality survey (first 90 days) and reliability survey (3 years). The results of these are different and when you use the information in any decision you plan to take - you should be careful to understand what it is and what it means for you.

Ron.

blizz81

I suppose I can work up a few counter-punches (of sorts)...

1) I kind of see it as, Joe Public will misunderstand. I know that's Joe Public's fault, regardless of JD Power & A's intentions, and that there's not much that can be done about that, but many people will attempt to spread information perhaps falsely based on results that they see...and my intentions in the original remark were, in a way, what I could do in my power to convey the message: Take this and any information presented by anyone else with this as the source with a grain of salt and not an indication of "this car is the best in every way period" (because that is basically the way some people would respond). Especially when people explicitely defend the (near-)infalliability of the mathematics involved - this just fuels the fire even more for Joe Public to stand by his/her possibly-not-so-well-founded views.

This was why I said "math or not, .... (which implies "with the mathematics out of the picture, done correctly or not, I don't care)" I don't think anyone here was directly challenging the scientific process nor the mathematics involved, moreso what the "simple very specific piece of information that is clearly explained" really should mean to someone, or maybe at the most, an outside jest on the social factors involved in consumer surveys.

2) I could make up some far-reaching retorts here but as I said I'm not really challenging the math.

3) I would still wave-of-the-handishly assert that these issues would add up to less muddying of the waters than calling people up and asking them questions over a phone survey/having them fill out a written survey. But that's a wash as far as convincing anyone of either side.

4) I do not understand the point of a 90-day initial quality survey in general, but I definitely don't understand it when they do a 3 year reliability survey. Shouldn't the issues that arise within the 3 year survey include those that arise within the first 90 days? If the people forget, was it really a problem worth any concern? What worthwhile guidance can you possibly get out of a 90-day anything, survey or otherwise, on your buying decision?

rloewy

I have no problem with the idea of taking every piece of information with a grain of salt - but in general the survey provides a good indication of the true picture - with reasonably high confidence levels.

Actually it is not a wash. There is an entire science out there that is based on having a SRS (Statistically Random Sample) and perdicting the mean and standard deviation of the population with very high levels of confidence, assuming that you have a large enough sample (assuming that the distribution is normal (Z) the mean of the sample means distribution will match the true population mean and the variance (standard deviation for our purpose) is much smaller than the true variance - which allows you to predict the real information with very high confidence levels. Any other scheme you come up with will not be a truly random sample and your results will be biased. This might be counter intuitive, but it is the way it works.
[/QUOTE]

The way I look at it - the 90 days survey is geared to serve the manufacturers and not the consumers - it tests market perception, marketing and sales messaging success (or flaw) and simple build quality and defect detection.

The 3 year reliability survey is the one that consumers will, IMHO, gain more important information from - because it provides more information about the long term quality of the design, build quality and components.

As for the 3 year survey containing the 90 days - I am sure it does - but it is a reliability survey, not a quality survey - do not confuse the two. For example - if you have a small defect detection problem in your car that caused the manufacturer to ship a car with a loose roof rack, for example - but other than this problem that was fixed after the car was purchase - the car was exemplary - the car will do very well in the reliability survey but my have more problems with the initial quality survey. This example is actually a specific example that we had with my wife's previous car - a 1997 Subaru Impreza. Other than a whistle issue that annoyed us for several weeks before it was diagnosed as the roof rack issue - the car had no problems.

So - they did have an initial quality issue with it, but no reliability issues.

Hope this makes it easier to understand.

Ron.

mikeb

Ron,
you have out done yourself:D

blizz81

Could the same logic not be used for surveying dealerships for similar information, ie, regardless of possible general survey concerns, the math involved will smoothen everything out to "truth", provided a large enough sample? I should probably just stop now - I feel this is going to circle back to 3.1, 3.2, and 3.3 (where I could just say, 3.a, the survey does not contain all the information about general design issues: stuff the people find problematic but not enough to bother reporting on the survey...3.b, some people work on their own cars, and would not report the quality defect they repaired as a problem on a survey....3.c, people are biased - sometimes they will not report issues publicly (on a survey, or with their friends and colleagues, even) on a vehicle if they have make/country of origin/etc pride and history, etc, etc.)

Then give it to the manufacturers privately for their own use, and don't allow people to misconstrue the information into unfounded conclusions, ie don't run an advertisement for your car displaying a big trophy for 15 of the 30 seconds explaining how your car is ranked #1 by JD Power & Associates, and in -1 point font at the bottom, "In one category in the 90-day initial quality survey".

rloewy

The math will be OK - but it will be based on biased data - so it will be biased by itself. When we sample at random - we ensure that the chance we get a very biased sample group is very small - if we limit ourself to a biased group - we increase likelihood that we get "bad" data. Using the same logic - why stop at the dealership? Go and survey the manufacturer - they have the warranty expense information so they can tell you exactly how reliable the car is, right?

Last thing about the idea of dealership survey that would make this kind of survey hard to perform - how do track specific cars. A specific car can be serviced by multiple dealers during the time you are interested in - so how can you be sure that car A really had 3 problems during this time if your survey talks only to one of the dealerships where it was serviced.

Even if you did not have this problem - assume that you talk to dealership A about cars X and Y - and the dealership tells you that they had to perform twice as many warranty jobs on car X compared to car Y - does it tell you anything? What if they sold 4 times as many car X compared to car Y - in this case despite the fact that they had twice as many jobs on car type X - on average car X had 1/2 the problems of car Y.

In other words - what I am trying to say is that the only reasonable way to get this kind of information is through a survey of the owners - they are not biased, they have all the information about a specific car - and the numbers are not skewed because of sales numbers.

Ron.

Supercharger

J.D. Power Vehicle Dependability Study (3 years) Awards


Premium Luxury: Lexus LS 400

Mid-Luxury: Acura RL

Entry Luxury: Lexus ES 300

Compact: Toyota Corolla

Entry Midsize: Chevrolet Malibu

Premium Midsize: Toyota Avalon

Full Size: Mercury Grand Marquis

Premium Sports: Porsche 911

Sporty: Toyota MR2 Spyder


Report: Fall 2003 issue J.D. Power Car Guide

Squidward

JD Powers is a marketing powerhouse. Even if their numbers were mathematically/statistically accurate, they cater to the marketing practices of automotive companies, who generally quote claims that are very vague or use qualified statements such as "best of it's class". With the right wording, they can pretty much make *any* claim, and JD powers can find the statistical data to back it up, and release a supposedly "unsponsored" best-of list, only to be quoted by the intended claimer.

When it comes to money, there's always ways to manipulate consumers into believers. It's called marketing.