NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/rec/autos/26338

Text File | 1992-11-10 | 5.8 KB | 124 lines

Newsgroups: rec.autos
Path: sparky!uunet!caen!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!jnielsen
From: [e-mail] (John F Nielsen)
Subject: Re: Survey impartiality (was Re: Profits to Japan?)
Message-ID: <[e-mail]>
Sender: [e-mail]
Nntp-Posting-Host: photon.magnus.acs.ohio-state.edu
Organization: The Ohio State University
References: <[e-mail]> <[e-mail]> <[e-mail]>
Date: Tue, 10 Nov 1992 22:37:26 GMT
Lines: 111

Think of it in the reverse way:

If a car company did in fact make a very unreliable product. Would
it in fact show up in the ratings? If a car had a very reliable
product would that show up in the ratings?  If not why?

>There is probably some value in trends over a period of years, in which
>case it might give you a clue as to what to look for.  However the summaries
>do not give the important dollar cost and/or difficulty in repairing the
>failure.  Like I said, some people get really snookered on a minor tuneup
>that is required every year; while others may may only pay a small amount
>every couple of years for a new exhaust system.
>
>CU really needs to quantify the magnitude of costs for maintenance and
>repair.  But I think they even realize that just data would be very
>dubious when you consider that most people have trouble with their
>simple checkbook.  I prefer to look at long term reports in the variour
>auto magazines for real costs over a time/distance.  Autoweek also supplies
>the "typical" price for certain repair parts (like alternators, windshields,
>etc.) which can be very revealing.

They have a rating for the cost of having the car, which they call the
cost index which  is not entirely dependant on just repairs. Ever
notice that?


>dig around in my files I can find it.  The point is, who is going to go
>to that much trouble to be part of a "volunteer" survey?  If you don't
>get EVERY owner of the particular model to respond then the ones who
>are satisfied with their car, may not want to bother.  If that is the
>case you will only get the frantics (for and against) to spend the time
>and effort.   CU's method IS NOT, I repeat, IS NOT statistically sound,
>nor factually rigorous.

I agree with you on that except, I think the info is still then
valuable, precisely because you get both extremes. A person
who reports poorly on the car will definitely have the numerous
repairs on the top of their mind. The person who had absolutely no
problems and loves the car will remember that too. IF you get
more people who report badly than report favorably then you
have a valuable piece of information.

>It's a cost/benefit trade-off.   Not everyone can afford a Mercedes.
>Nor are there and infinite supply of quality cars.

You say it then gets scewed by the impression people have of american
or foreign cars in general. I venture that this impression is much
weaker than you guess because who is going to buy a brand new car they
do not like???? It's a big investment and the Japanese compete in the
same price range believe it or not. If a person really wanted a
Japanese car they would have got one. I don't understand your
reasoning.

>

>
>|> Why would the Suzuki Samauri, something CU trashed have an excellent
>|> reliability rating.
>
>What you keep doing is using this word "reliability" everytime you refer
>to their repair "history".  Reliability is measured in mean time to failure
>(MTBF) or number of defects over a population (ppm).  Those terms are never
>used in the CU ratings.   They are reporting incidents and making a judgement
>call on them being above or below average (whatever that is).  It cannot
>predict anything about the probability of a failure occurring.  Your
>paradox example above shows just that.
If a car is reliable, there will be less problems to report per car.
Why won't that show up in the ratings?

And, what paradox example?

>Are you suggesting that a person can judge the expected quality of
>a new car based on the repair incidence of past years?   Not even the
>Japanese would agree to that.   The name of the game for the successful
>players is to constantly strive for improvements.  It may very well be
>that the car with the 5-year record for bad exhausts, finally made a
>design/material change that makes their system 2X better than the
>competition.  To assume the future based on the past is pure negative
>CU thinking at its best.

But, then you are changing the variables. What about a person who is
thinking of buying the car with the old exhasut system would
the info be valuable for them? I cannot see why it wouldn't.
If many  people report problems with the exhaust then I think
that would be a very strong indicator for that model and year of car.
Why wouldn't it be?

>I don't. And I haven't been burned.  No, take that back, I once bought a
>piece of crap that I think CU liked -- a used '75 Honda Civic.  I made
>the mistake of assuming that since it was Japanese in couldn't possibly
>be bad.   So much for collective opinion and predicted "reliability."

Every company makes a lemon, what matters is who makes *more* lemons.

>BTW.  Just once, I wish the consumer advocates would do an analysis on
>the realative merits of various consumer magazines :-)  But, that would be
>like asking the fox to watch the chicken coop :-(

Of course, I am just saying the info is valuable, maybe not
statiscally rigorous but still valuable, because the good and bad
will show up.

john



--
-------------------------------------------------------------------
John Nielsen   MAGNUS Consultant            ______   ______   __  __
"I can't compete with you physically,      /\  __ \ /\  ___\ /\ \/\ \
and you're no match for my brains."          \ \ \/\ \\ \___  \\ \ \_\ \