NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/rec/autos/32792

Internet Message Format | 1993-01-27 | 3.7 KB

Path: sparky!uunet!opl.com!hri.com!ukma!gatech!swrinde!zaphod.mps.ohio-state.edu!saimiri.primate.wisc.edu!ames!sun-barr!male.EBay.Sun.COM!news2me.EBay.Sun.COM!seven-up.East.Sun.COM!phaedrus!ebergman
From: [e-mail] (Eric Bergman - SunSoft Human Interface Engineering)
Newsgroups: rec.autos
Subject: Re: yet another consumer reports article (
Message-ID: <[e-mail]>
Date: 26 Jan 93 17:56:06 GMT
References: <[e-mail]>
Reply-To: [e-mail]
Organization: Sun Microsystems Inc. - BDC
Lines: 69
NNTP-Posting-Host: phaedrus.east.sun.com


>But, it is not an issue because it is the same for every make of car.
>That is a constant factor. Perhaps you can say the problem side is
>more heavily weighed than the non-problem side, but it is more heavily
>weighed for everyone.

This is just wrong. It is also not true to say that since you are
"already looking at people with Type A personality" it is ok to
draw conclusions about just those people.

Your comment above about "a constant factor" *assumes* a constant
factor.  In reality, it is EXACTLY that assumption that scientific
use of statistics seeks to avoid.  Instead of your Stat class, I suggest
you take a Research Methods course in a department of Biometrics,
Psychology, or Epidemiology.

I will make up an example that is a bit silly, but it illustrates
the point.  Remember that this is to explain the theory,
so the particulars are not as important as the idea that there
are many different factors operating at the same time.

Let's say there are 3 attitudes of people when it comes to cars:

Car Fans -- tend to like performance cars & high end sedans w/ latest technology
Neutrals -- tend to like practical 4 door sedans, decked-out econoboxes
Car Haters -- tend to like econoboxes & tiny cars, the cheaper the better

Car Fans know lots about cars, love cars & take care of them on schedule.

Neutrals know a little about cars, take care of them usually, but skip
some maintainance.

Car Haters want to go from point A to point B.  They fill up the tank and
bring in the car if something REALLY goes wrong.  They change the oil
when they think of it.

We all know people who fall into these different categories.

What kind of cars do you think these people drive?  Well, it turns out
that the "hidden factor" of their attitude *influences* what kind of
car they buy AND at THE SAME TIME, how they treat it.

Even if (and that is a BIG *if*) in surveys these people respond at
the same rate about their cars, different types of people have different
cars and treat them differently...Ok, big deal -- what does this mean?

It means that if I as a consumer am comparing a Honda Accord to
a Honda Civic (I can't decide whether I need the extra room), the data
may say that the Accord is slightly more reliable, but in fact it easily
could be exactly opposite.

The *hidden* factor of driver type means that the Civics on the road
are getting worse treatment than the Accords.  There are twice as many
Car Haters driving Civics as Accords.  They are *abusing* their cars.
I don't KNOW that, so I just read the results and say that they are about
the same. WRONG.  If I knew about the hidden factor (and its LEVEL) and
controlled for it in my sample, I would determine that Civics are actually
much *more* reliable than Accords.

As I say, this is an example.  The point is, there are numerous hidden
factors that make any conclusions from the Consumer Reports data
scientifically invalid.  That is just a fact -- not even a point
of argument really.  You wouldn't get past the door in an intro
course arguing that it you can.

Check it out in a course on Research Methodology.


Eric