plunkett@rlgvax.UUCP (Scott Plunkett) (02/01/84)
Mike Kelly and others are fond of quoting mysterious numbers
confirming that the vast bulk of America stands resolutely behind
certain liberal notions. For instance, Mr Kelly has written that 80% of
America supports the nuclear freeze. Numerous similar statements
have appeared to unimpeachably confirm one liberal position or another.
No doubt some conservatives use the same shabby propaganda tactics.
While there has been legitimate complaint from others about the use
of numbers like "80%" without attribution, I should like to suggest
that most polls, especially the so-called "opinion polls", are by
their very intent, worthless, meaningless, tendentious, and ought
to be ignored by all thinking people everywhere.
I am not swayed by the statistical proofs of their accuracy; that is
irrelevant. The opinion poll (more often than not, I claim) is
inherently flawed in the framing of the question, and further
bias is easily factored in by using clever sampling techniques.
So, given a statement such as "80% of America supports the nuclear
freeze" we must surely wonder, (a) What is the text of the question
the respondents were answering, (b) How were the respondents selected,
(c) When were they asked, (d) Who commissioned the pollster. There
are probably other factors that could introduce bias into the result,
but these would be enough.
...{allegra,seismo}!rlgvax!plunkettmmt@dciem.UUCP (Martin Taylor) (02/08/84)
================I am not swayed by the statistical proofs of their accuracy; that is
irrelevant. The opinion poll (more often than not, I claim) is
inherently flawed in the framing of the question, and further
bias is easily factored in by using clever sampling techniques.
So, given a statement such as "80% of America supports the nuclear
freeze" we must surely wonder, (a) What is the text of the question
the respondents were answering, (b) How were the respondents selected,
(c) When were they asked, (d) Who commissioned the pollster. There
are probably other factors that could introduce bias into the result,
but these would be enough.
...{allegra,seismo}!rlgvax!plunkett
================
The sky is falling!!! I have found something written by the plunkett
person with which I agree!
I'd go a little further even than Plunkett on this one. Opinion polls,
even when accurate, can be used to drive opinion in directions where
factual analysis might not lead.
--
Martin Taylor
{allegra,linus,ihnp4,uw-beaver,floyd,ubc-vision}!utzoo!dciem!mmt