home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!saimiri.primate.wisc.edu!ames!sun-barr!news2me.EBay.Sun.COM!seven-up.East.Sun.COM!phaedrus!ebergman
- From: ebergman@phaedrus.East.Sun.COM (Eric Bergman)
- Newsgroups: rec.autos
- Subject: Re: yet another consumer reports article (
- Date: 25 Jan 1993 23:31:49 GMT
- Organization: Sun Microsystems Inc. - BDC
- Lines: 46
- Distribution: world
- Message-ID: <1k1t96INN176@seven-up.East.Sun.COM>
- References: <1993Jan25.174221.11394@magnus.acs.ohio-state.edu>
- Reply-To: ebergman@phaedrus.East.Sun.COM
- NNTP-Posting-Host: phaedrus.east.sun.com
- Summary: statistics is a science
- Keywords: consumer reports statistics
-
- > IF you have 10,000 people responding about problems with their Ford
- > Ranger and 50,000 about problems with their Ford Taurus, patterns
- > will show up.
-
- Here is a story about why sample bias is important,
- REGARDLESS OF THE SIZE OF THE SAMPLE:
-
- Literary Digest did a telephone survey of Readers in 1936.
- They asked for whom readers were intending to vote, and on the
- basis of the huge response (over 100,000 people responded!),
- they predicted that Roosevelt would lose in a landslide.
- (For you young folks -- Roosevelt won in a landslide).
-
- You are right to say that "patterns will show up", but you
- are wrong to think that we can draw valid conclusions from
- those patterns.
-
- Your email ends with the note "Back off man, I'm a Scientist!",
- which I think is great, but really, you should know that there
- are people who get PhD's studying how to do valid survey research.
- Doing it RIGHT is a science. It is NOT so simple as to say
- "there are patterns". That is no more science than it is to
- for an Astrologer to say that there are patterns in the stars.
-
- > And, both with have groups of people who respond
- > because they love or hate their machine.
-
- True. But you can't know what factors will cause people to
- respond differentially. For example, in the poll I cited above,
- the reason that Roosevelt did so poorly is that Readers of the
- Literary Digest who *had phones* were more likely to be Republicans,
- because phones were still relatively expensive in 1936.
-
- For all we know, readers of Consumer Reports are highly likely to
- have families -- perhaps in the large samples, the people who
- drive the Taurus treat it with greater care than those who drive
- an Eclipse, for example. Who *knows*. The point is, there are
- too many factors that could cause "the patterns" that you speak
- of to be utterly meaningless.
-
- The useful information, I think, is not comparative, but descriptive.
- Consumer Reports describes cars. If they use the same criteria, then
- comparing the qualitative info in the descriptions is useful.
-
-
-
-