Daily Kos: Newsweek Polls: A Tale of Two Countries

[Pirate Smile]

Newsweek Polls: A Tale of Two Countries

by DHinMI
Sat Jul 12, 2008 at 01:50:33 PM PDT

In one country, which Newsweek polled in June, the electorate is split R26/D38/I36. In the other country, which Newsweek polled a few weeks later, that country's electorate is split R32/D32/I36.

In one country, polled in June, 84% of the electorate is white. In the other country, polled a few weeks later, 88% of the electorate is white.

In one country, polled in June, 19% of the registered voters are under 40, 39% are between 40 and 59, and 42% are 60 or older. In the other country, polled a few weeks later, only 17% of the registered voters are under 40, 38% are between 40 and 59, and 45% are 60 or older.

In the younger, slightly more racially diverse and much more Democratic country, Barack Obama was supposedly leading by 15 points. In the country that is somewhat older, is somewhat whiter, and is evenly split between Republicans and Democrats, Barack Obama is supposedly leading McCain by only 3 points.

According to Newsweek, comparing these results shows "what a difference a few weeks can make." A more valid conclusion is that it shows what a difference polling two very different countries can make.

Why does this silliness happen? DemFromCT gave us the answer a few weeks ago.

http://www.dailykos.com/storyonly/2008/7/12/124829/538/624/550584

[depakid]

1. Bottom line: polls in America are designed to influence public opinion

NOT to reflect it.

That's the general rule and nowhere does it apply more clearly than with cheap media polls.