From: larry moe 'n curly on 20 Nov 2009 02:39 Derek Gee wrote: > "larry moe 'n curly" <larrymoencurly(a)my-deja.com> wrote in message > news:d1ac9580-a06c-4691-86f9-302ae1f854c9(a)u25g2000prh.googlegroups.com... > > > So why hasn't CR shown any bias when they've tested American and > > Japanese twins, like the Toyota Matrix and Pontiac Vibe or the Toyota > > Corolla and Geo Prizm? > > They have. Look back through the archives at their ratings for the Ford > Probe and Mazda 626. They were built in the same US plant, using largely > the same parts, yet the Mazda was always given the higher reliability > rating. This was one of the first clues that I had that something was > wrong at CR. > > > Please explain. Here are the reliability ratings of the 626 and > > Probe: > > > > http://farm3.static.flickr.com/2580/4114348350_3d1e3833c8_o.jpg > > What do you mean please explain? Can't you read? Look at the '89 > 4-cylinders for instance. The Ford's have extra black dots (or other lower > ratings) the Mazda 626 doesn't. (Cooling, electrical, A/C, clutch, > integrity, paint-trim) Why is the Ford version rated worse? They were made > in the same plant with the same platform parts. Can't you read? For some components, the 626 rated worse than the Probe. > CR did rate the Mazda auto-trans worse, which is also wrong, being the same > unit as the Ford. Why the difference then? I'm going to blame small survey samples, which tend to lead to wider variations in averages -- notice the * in some spots, indicating insufficient data. You're not doing a Robert McNamara on statistics, are you? ;)
From: clare on 20 Nov 2009 13:13 On Thu, 19 Nov 2009 23:28:39 -0800 (PST), "larry moe 'n curly" <larrymoencurly(a)my-deja.com> wrote: > > >C. E. White wrote: >> >> "Andrew Rossmann" <andysnewsreply(a)no_junk.comcast.net> wrote in >> message news:MPG.256cdade1ef3a1b9896cc(a)news.eternal-september.org... >> >> > CR only reports what their annual survey says. If there is any bias, >> > it's in the subscribers. >> >> I don't completely agree with this. The CR survey is very simplistic. >> A lot of the questions depend on the responder making value >> judgements. These judgements are based on the experience / beliefs / >> opinions of the reponders. > >I don't believe subjectivity affects the reliability ratings much >because CR also asks owners if they'd buy the same vehicle again, and >some vehicles where the owners overwhelmingly answer yes are ranked >among the least reliable. IOW the very same people who love their >cars admit that their cars have been troublesome, the Chevy Corvette >being a prime example. You need to remember too, the JD Power report is not so much about reliability as "initial quality " - at least that's the report most often referred to. That means what the car was like when first purchased - has little to do with what you had 6 months later. A new Caddy is a sight to behold, and for the first couple weeks you may be "in heaven" - from the experience of several friends, that turns to a minimum of "purgatory" within 6 months.
From: Derek Gee on 21 Nov 2009 22:18 "dr_jeff" <utz(a)msu.edu> wrote in message news:ptKdnRH_9eYyYZjWnZ2dnUVZ_v5i4p2d(a)giganews.com... > Derek Gee wrote: >> "dr_jeff" <utz(a)msu.edu> wrote in message >> news:JOKdnUJqm4PvM5nWnZ2dnUVZ_t9i4p2d(a)giganews.com... >>> Derek Gee wrote: >>>> "larry moe 'n curly" <larrymoencurly(a)my-deja.com> wrote in message >>>> news:f152deef-37b4-4a7e-85d7-5f85bd895e61(a)m7g2000prd.googlegroups.com... >>>>> Derek Gee wrote: >>>>>> "Andrew Rossmann" <andysnewsreply(a)no_junk.comcast.net> wrote in >>>>>> message >>>>>> news:MPG.256cdade1ef3a1b9896cc(a)news.eternal-september.org... >>>>>> >>>>>>> CR only reports what their annual survey says. If there is any bias, >>>>>>> it's in the subscribers. >>>>>> There's always bias in humans, but a better survey would do a better >>>>>> job in >>>>>> trying to design it out of the polling. The JD Power data seems >>>>>> better >>>>>> quality. >>>>> JD Powers is worse and has shown more bias in favor of luxury cars and >>>>> cars favored by senior citizens, whether or not those vehicles were >>>>> reliable (Lincoln) or not (Jaguar before Ford bought them). >>>> Not so. They have a much more accurate survey questionaire than the CR >>>> one. >>> How is the JD Power survey more accurate? Has there been a study on >>> this? It may look better on paper, but it may not better in real life. >> >> From what I've been able to gather about the IQS and VDS surveys, it's a >> 44 point questionaire with specific questions about stuff like handling, >> braking, seats, audio systems etc. It's better than just asking which >> areas did you have a "problem" with and asking the user to check a single >> box. >> >> Here's some very specific criticism of the CR methodology I recently ran >> across: >> >> http://www.truedelta.com/pieces/shortcomings.php >> >> http://www.truedelta.com/pieces/cr_survey.php >> >> http://www.truedelta.com/pieces/newdots.php >> >> http://www.truedelta.com/pieces/anomalies.php >> >> >> >> Derek >> >> > > The website is a competitor to CR for car data. How reliable do you think > its criticisms are? Valid enough. I took a year of statistics in college. The True Delta guy is explaining how he plans to avoid those issues with his site. From what I've seen of his site (because I checked it out after reading those pages), his sample sizes are still too small for my liking. Derek
From: Derek Gee on 21 Nov 2009 22:25 "larry moe 'n curly" <larrymoencurly(a)my-deja.com> wrote in message news:4b39b572-6084-4c98-a51a-666de330436a(a)z35g2000prh.googlegroups.com... > Derek Gee wrote: > >> "larry moe 'n curly" <larrymoencurly(a)my-deja.com> wrote in message >> news:f152deef-37b4-4a7e-85d7-5f85bd895e61(a)m7g2000prd.googlegroups.com... >> >> There's always bias in humans, but a better survey would do a better job >> in trying to design it out of the polling. The JD Power data seems >> better >> quality. >> >> > JD Powers is worse and has shown more bias in favor of luxury cars and >> > cars favored by senior citizens, whether or not those vehicles were >> > reliable (Lincoln) or not (Jaguar before Ford bought them). >> >> Not so. They have a much more accurate survey questionaire than the CR >> one. > > Show me. I've gotten the CR survey a couple of times, and it > basically asked if I had any problems in the past year in any of the > areas listed in their car reliability survey tables. What does JD > Powers ask? A Mar. 2004 US News magazine article said they lump too > many aspects of the cars together: I wish I could show you. I've only been able to fill out Power surveys on auto options. I haven't been able to get my hands on a IQS or VDS survey. I've had to rely on other descriptions of them. > "One gripe: In the new car quality survey, a car's basic > attributes, gas mileage, and the placement of cup holders, > for example, are lumped in with problems like rattles, buzzes, > and broken equipment." > > And the fact that JD Powers has rated Cadillac and pre-Ford Jaguar > high in reliability proves that something's been seriously wrong with > their surveys. I don't remember pre-Ford Jags ever being rated high in reliability anywhere. Do you remember which year you saw this? >> Also, don't forget that the quality of the dealership can go a long way >> toward how 'reliable' you think your car is. >> >> > Then I'd expect luxury car brands to fare better because of their >> > dealerships, so why have Cadlllac and Rolls-Royce long fared so poorly >> > in reliability ratings? >> >> Where have you seen any published reliability data on RR? They aren't in >> any published survey I've ever seen. > > RR is pretty famous for bad quality, and I even heard an owner > complain about his to my boss. Name a British car that isn't > unreliable. And how do you explain Jaguar's high ranking in the > Powers' surveys? You know Jaguar, the brand so bad that when Ford > took over the company, one of its ads touted a brand-new wiring > harness? Then there's the Powers' quality award given to the horrible > cable TV company here. Really, you need to demonstrate that Powers > has any credibility at all, especially when their sources of revenue > are a mystery or could be from the very industries whose products and > services they rank. > >> > And Toyota's brand with the highest reliability is budget Scion, not >> > luxury Lexus. >> >> Wrong, Scion is WAY below industry average in the 2009 JD Power >> Dependability study! > >> http://www.jdpower.com/corporate/news/releases/pdf/2009043.pdf > > Wrong, Scion is WAY above industry average in the more trustworthy, > non-sellout 2009 Consumer Reports reliability survey: > > http://farm3.static.flickr.com/2557/4118564115_3013ca4784_o.gif > > Why shouldn't Scion score high when they have only three models, two > of them based on the Toyota Yaris, which is rated high in reliability > (Yet Consumer Reports doesn't recommend it because it scores too > poorly in performance). I tend to believe the Power survey due to the better methodology, plus I have two first hand owner reports of failures (disabled) of Scion models. Granted, that's only a personal sample of two, but it fits the pattern of the Power data - Scions suck. My guess is that even though it's based on a Yaris, the customization is causing quality problems. That's a truism with automakers, the more options, the more potential for problems. Derek
From: Derek Gee on 21 Nov 2009 22:33
"larry moe 'n curly" <larrymoencurly(a)my-deja.com> wrote in message news:7d7642f3-3295-4e59-be33-5c3db331b561(a)i12g2000prg.googlegroups.com... > > > Derek Gee wrote: > >> "larry moe 'n curly" <larrymoencurly(a)my-deja.com> wrote in message >> news:d1ac9580-a06c-4691-86f9-302ae1f854c9(a)u25g2000prh.googlegroups.com... >> >> > So why hasn't CR shown any bias when they've tested American and >> > Japanese twins, like the Toyota Matrix and Pontiac Vibe or the Toyota >> > Corolla and Geo Prizm? >> >> They have. Look back through the archives at their ratings for the Ford >> Probe and Mazda 626. They were built in the same US plant, using largely >> the same parts, yet the Mazda was always given the higher reliability >> rating. This was one of the first clues that I had that something was >> wrong at CR. >> >> > Please explain. Here are the reliability ratings of the 626 and >> > Probe: >> > >> > http://farm3.static.flickr.com/2580/4114348350_3d1e3833c8_o.jpg >> >> What do you mean please explain? Can't you read? Look at the '89 >> 4-cylinders for instance. The Ford's have extra black dots (or other >> lower >> ratings) the Mazda 626 doesn't. (Cooling, electrical, A/C, clutch, >> integrity, paint-trim) Why is the Ford version rated worse? They were >> made >> in the same plant with the same platform parts. > > Can't you read? For some components, the 626 rated worse than the > Probe. Only the auto trans on the year I cited. Looking at some of the other years, the pattern I cited generally holds. Only the Mazda auto-trans seems to get picked on. >> CR did rate the Mazda auto-trans worse, which is also wrong, being the >> same >> unit as the Ford. > > Why the difference then? I'm going to blame small survey samples, > which tend to lead to wider variations in averages -- notice the * in > some spots, indicating insufficient data. Given how widely read and widely responded to that CR survey is, I doubt if small survey samples are the problem, but I don't have access to the raw numbers to prove that. Derek |