[thechat] polls / more conservative bias (was: Getting closer)

Erik Mattheis gozz at gozz.com
Sat Feb 8 17:31:01 CST 2003


On Saturday, February 8, 2003, at 09:09 AM, Tony Crockford wrote:
> 50:50 split on a poll of 800 people!
> Polls tell you absolutely nothing.

Are you mostly referring to:
<http://news.findlaw.com/ap_stories/i/1107/2-2-2003/
20030202060006_27.html>

Following is the last sentence I wrote, but is the one I'm most proud
of, so I'll move it up front: That the article (note: an AP wire, not a
FindLaw article) directly attributes the reverse in the trend of
support for war to President Bush is a textbook example of Conservative
media bias; anyone willing to challenge me on that? Mr. Kevin Isom -
you have the floor.

Anyway ...

A series of polls of 800 people can tell you a lot about trends. (I
don't claim to to have authoritative knowledge of statistics, so
someone jump in and correct me if I'm wrong here).

Most polls you see in the news have a sample of about 1000 people, a
margin of error of about 4% and a confidence level of about 95%. This
means that the result stated in the poll is accurate +/- 4% (a range of
8%) and 1 out of 20 times, the result would be outside of that 8% range
had you polled a different group of 1000 people.

This is always in the fine print and ( as in
<http://news.findlaw.com/ap_stories/i/1107/2-2-2003/
20030202060006_27.html> the confidence level is usually completely
absent).

So looking at the statement from the linked article, "President Bush
has reversed the slide in public support for a possible war with Iraq,
with two-thirds, 66 percent, in an ABC-Washington Post poll released
Saturday saying they support military action against Iraq."

As it's completely absent, guessing the confidence level at 95%, that
66% really means "If you asked 855 people, 19 out of 20 times, between
62.5 and 69.5% of the people would say they support military action
against Iraq.

The main jist of that part of the article is that the trend against war
has recently reversed. They base that statement on comparisons to the
same poll conducted in prior months (Jan: 47.5 to 54.5%) (Dec: 33.5 to
40.5%). So if the article had said (again assuming a 95% confidence
level) with "There is a 95% chance that the trend against war with Iraq
has reversed ...", it would have been completely accurate.

-----------------------
Erik Mattheis
GoZz Digital
<http://goZz.com/>
Flash and ColdFusion Development
Minneapolis, MN
-----------------------




More information about the thechat mailing list