Denon, Yamaha or Marantz Receiver

[Woochifer]

actually I worked in accounting for a decade and understand stats and it is you who is trying to weasal out of the truth.

Just because you can work your way around a ledger sheet does not mean that you know anything about survey research design, validation procedures, or how to report statistics using that methodology. And if I'm weaseling out of the truth, then what is YOUR definition of the truth? That CU is lying? That their statistical procedures are biased? That European cars are more reliable than Americanones just because you say so? Sorry, but molding universal truth out of your own biased version of reality doesn't stand up to any kind of objective scrutiny.

You are banking on the following premise 20 Europpean cars fail versus 18 fails of Eurpean cars...therefore European cars are now worse than American cars. Answer yes oer no to this because that is ALL the stats presented by USA indicate. Improvements began when co-productions started. A co-production is not an American vehicle the same way a non co-produced American car is. I give the American car manufacturers credit for finally realizing the don't how to design nor build a good vehicle and to get help from those who are competant and know how to run a business that puts out quality over profit margins...seemingly they could not do that either judging by Toyota and Honda for that matter.

You're just putting words into my mouth for argument's sake. Just because you have a love for exaggerating, does not mean that others share your obsession. Nowhere did I say that European cars are now WORSE than American cars. Whenever I say that one car is WORSE than another, I go by a lot more criteria than just the reliability record. If you go back to my original point, I was simply pointing out that based on the CU survey, American cars now have a lower defect rate than European cars, where's the untruth in that? And why this sets off such an illogically visceral reaction on your part is anyone's guess. If you want to parse that general point to death and try and find exceptions, asterisks, etc. and try and find some hypothetical rearrangement of the data that better fits your biases, go ahead. But, then it certainly would not have more validity than what CU reported.

Reliability is meaningless woithout validity - you studied that right? There are studies with over 250,000 people in longitudinal nutrition studies over a 30 year period which is now kaka because the tested the wrong thing.

So, tell me the invalid part of the CU data collection procedures, sampling, and/or survey form. Just because you can point out one example of bad research design doesn't mean that all research is therefore invalid.

your trial notion is idiotic. In audio they use a 16 trial test - invented by??? some engineer no doubt. The statistical significance of a trial of 10 would be 9 correct in a DBT. OR 59/100(achieved as 6/10 ten times in a row with one 5). You run your 10 trial test and score 6 your test is chucked out and you can't tell the difference between componant a and b. But with more trials achiviing the the 59/100 is the EXACT same statistically significant outcome at the .05 level. Naturally you don't see this little notion dawn on anyone or the reaosn WHY more trials would be used. More trials = more confidence in the result.

What trial notion did I bring up? The DBT research design that you're bringing up is completely off topic because a consumer survey is not about doing repeated trials of a behavioral input and comparing stimulus variables. If you survey a consumer about their vehicle's repair history 100 times, guess what, they will give you the same response 100 times. Now you're REALLY getting desperate by bringing these kinds of irrelevancies into the discussion.

All I said was at the very beginning the USA today article compared European cars with American cars. Well Jaguar is a Ford - it's owned by an Amrican then it is American...and if it doesn't count as an American Car then how the hell does the Japanese co=produced cars count as Amercian. Oh I see only if it serves the American car manufacture best. If Jag scores terrible then it's European but if the Toyota Matrix(ahem Pontiac Vibe) gets great reviews it's a GM?

First off, the main conclusion of the USA Today article is that this year's batch of cars represents the first time that European nameplates have a higher defect rate than American cars in 24 years of tracking the data. You were trying to berate American cars for their reliability, and this is obviously some valid data to the contrary. Volvo's also now owned by Ford, but the cars are still designed and built in Sweden. Jaguar's owned by Ford, but they are still designed and built in Britain. The one strongsuit of the CU survey is that they have been consistent during that time period in tracking and reporting the data.

Sorry it doesn't take a person with a Stats 101 course to know that without knowing specifics then you have nothing. I don't mean a 50 page report. This is not hard...the next time CR comes out with a full listing of cars I can set up a spread sheet with Co-produced cars and results on one side and non co-produced cars on the other - and separate the American Cars and European Cars.

What specifics are you trying to get at? They're tracking the data the exact same way that they did the past 24 years, which makes these kinds of comparisons relevant. The other tracking trends have been pointing in that direction for years (i.e. VW and Mercedes' reliability declines, BMW's recent problems with the 5 and 7 series, Land Rover's perennial reliability issues, etc.).

Coproduced cars represent a relatively small segment of the overall market, certainly not big enough to drive a 675,000 response survey, and you can check the sales charts from R.H. Polk and other sources if you want to look for that. If you want to go on the basis of where they are produced, then that's another story altogether given that many foreign car companies now make cars here, and a lot of domestic cars are made in Mexico, Canada, and elsewhere. But, the country of origin is something that CU does not track from year-to-year, and the occasions (with the first U.S. made Honda Accords I remember) where they did compare U.S. and Japanese made versions of identical models, they found no significant differences in the reliability.

Generalizing American cars witha far higher reliable statistics base (in terms of confidence level) versus that of European cars is ridiculous. I don't blame CR for the stat because i understand what it is they're REALLY saying in that article that the Big three have come a LONG way to curtail their problems - 108 to 18 is terrific...naturally they're going to say wow look how much we improved compared to the Europeans. But then they had SO MUCH farther to come. What you expected the Europeans to improve to 1?

Considering how far ahead European nameplates were, you would expect that they would be able to maintain at least a slight edge. But, like I said, Mercedes made the mistake of not changing the fundamental way that they design cars. They still do a bumper-to-bumper redesign with every new model cycle, and their desire to maintain profit margins in the face of rising R&D costs led them to economize on the parts used in the vehicles themselves. BMW's recent 5 and 7 series models introduced some new untested electronics, some of which have proven unreliable. Volkswagen let their QC guard down as their sales picked up, and now their reputation's taking a beating, which is too bad because I generally like their cars otherwise and my wife has been saving up for a New Beetle. European car makers at one point had a better than 2-to-1 edge in their defect rate. How they let that huge edge whither away should be the burning question rather than trivialities about how the survey sample was divied up. And as I've pointed out repeatedly, this is not some one-year outlier result (and with survey samples this large, outliers are not going to drive findings), it's part of a general trend that's been going on for the past few years.

For this to be worth thing one to a BUYER who is looking at manufacturers you NEED to know the numbers for GM, Chrysler, Ford, BMW, Mercedes. What are the percentage for each company. Surely if they have the numbers to make a broad and useless generalizaion they have the percentage for each car manufacturer - and then for each car. I mean it would make sense for Ford to know that hey the Focus and or Explorer are bringing the entire company into the mud maybe it's time to dump the sh!t and move on to something else. Or at the very least phone Toyota and ask them to re-design the car for them so wheels don't fall off and engines don't catch fire. This method saves both companies money - build them in the same plant.

No, what matters to the buyer is how reliable the model that they are test driving has been. They don't care that the VW Passat has had an average reliability record for the most part if the VW Jetta that they are interested in buying has had high defect rates in most categories. If CU is reporting the data, then it meets their minimum confidence level, if the data's insufficient, then the buyer's on their own. All that other background about which car maker or model has a higher proportion of the sample, or how the company does in aggregate, makes for good bulletin board fodder and fanboy discussions, but pretty much irrelevant otherwise.

You don't seem to want to look at individual cars - or the co-production issue because I suspect you've read CR like I have and know yourself which ones get recommended and which ones are dung heeps.

Quite the contrary, that overall defect rate is very indicative of a general trend that I've seen among individual European models over the past five or six years. For example, before 1998, Mercedes were mostly above average or much better than average in their reliability records, but around the introduction of the M-class, successive model introductions have shown more problems than before. VW has had a similar slide in the reliability of most of its models as well. BMW was strong for the most part, but CU's now reporting on problems with the 7 series, and I've been reading other stuff about the 5 series as well. Frankly, I'm surprised that it took until this year for the CU findings to confirm what's been showing up on the JD Power rankings and CU's own reliability information on individual models the past few years.

Coproduction's not an issue because examples like the Pontiac Vibe/Toyota Matrix or the Nissan Quest/Mercury Villager or Chevy Prizm/Toyota Corolla make up such a minimal portion of the overall auto sales (at least on the American side of the sales ledger) that they would hardly make a dent in an aggregated summation like the one reported in the USA Today article, and it's not like coproduced cars have grown so much the past five years that they would completely explain the improvements in American car reliability (in fact, both the Villager and Prizm were discontinued for 2004, so the coproduction cross-badging strategy's obviously not a major part of American car makers' plans). If you're talking about shared platforms or drivetrains, then there might be an argument there. However, even in those cases, you're talking about significantly different vehicles with potentially very different reliability records. I mean, the Sterling shared the platform, drivetrain, and a lot of the same body parts as the Acura Legend, yet the reliability records of those two models were about as contrasting as you can get.