Newshound Editorial:
Some of the commenters on this site called my attention to the following issues around the Newsweek poll that showed Bush leading Kerry 54 to 43 percent, giving Bush a 12 point lead.
Out of a sample size of 1,188, a total of 578 were military. This number is 48.6% of the sample, which is much higher than the military in the voting public. Also, the non-white sample was 144 our of 1,188. This number of 12% is much lower than the actual non-white voting population. Therefore, both the military and the non-white samples favor the republicans. This is true because we had at least three weeks of non-stop coverage of the Swift Boat ads, and the military tends to favor republicans as a group anyway, even without the extra incentive. The non-white population favors democrats by large margins, and they were not adequately represented in the poll. See NEWSWEEK POLL: REPUBLICAN CONVENTION 2004 for more detail.
Comment: When I had statistics in college, I learned that you can influence survey results with deliberate selection of the demographics and by using a small sample size. The results can be manipulated to reflect what you want them to reflect, depending on who and how many you ask, as well as how you word the question. Polls can also be influenced by the halo effect. If a candidate is perceived as a winner, people without strong convictions, who can go one way or the other, will tend to vote for the guy they think will win. Thus, the importance of being ahead in the polls, legitimate or not.
I have believed for a long time that most of the polls we hear about have no relationship to reality. It really helps a winner if the actual results match the pre-election polling numbers. An interesting election change since 2000 is the disappearance of exit polls. Exit polls in Florida varied from the actual results, and that added to the confusion when the networks called the election for Gore early based on exit polls. Most of them have gone out of business. In some places, activist groups have replaced them, but that's the exception and not the rule.
http://www.newshounds.us/2004/09/04/politics_and_polls.php#moreThere are demographics linked on this page at
http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=104&STORY=/www/story/09-04-2004/0002244238&EDATE= . Here are some of them:
374 Republicans (plus or minus 6)
303 Democrats (plus or minus 6)
300 Independents (plus or minus 6)
476 Men (plus or minus 5)
532 Women (plus or minus 5)
112 18-29 (plus or minus 10)
349 30-49 (plus or minus 6)
516 50+ (plus or minus 5)
283 Southern White (plus or minus 7)
559 Non-Southern White (plus or minus 5)
144 Non-White (plus or minus 8)
417 Military households (plus or minus 6)
578 Non-military households (plus or minus 5)
310 Republican states (plus or minus 6)
409 Swing states (plus or minus 6)
289 Democratic states (plus or minus 7)
541 Bush/Cheney supporters (plus or minus 5)
403 Kerry/Edward supporters (plus or minus 6)
Republicans are over-represented because: Southern white are over-represented, white are over-represented, men are over-represented, 50+ are over-represented, and military is over-represented. All groups that will favor Bush.
Why would you deliberately choose 1/3 of your poll (283/<283+559+144> = 28%) from states that have 20% of the EV's (<8 + 15 + 9 + 6 + 9 + 27 + 34>/538 = 20%)? Because you want to "tweak" the results. But not TOO much. It's got to look believable, and the timing has to be right (like right after a convention.)
We are being manipulated. The question is, what can we do about it?