Saturday, January 07, 2006

Polls, Lies, and Statistics

AP/Ipsos released a poll which at face value tells us that the GOP is a dead duck.

Fortunately, the AnkleBiting Pundits dismembers the poll rather thoroughly:

Well, as we've said before you (and the GOP leaders) shouldn't use these polls as any kind of indicator as to what could happen in 2006, which is certainly how the MSM will spin it.But here's what you won't see in the news stories. - the demography of the poll respondents

First, only 81% of respondents were even eligible to vote, and there's no indication of how many of them actually went to the polls in 2004..

1. Party Leanings - The poll is slanted 52-40% towards Democrats, even though the voters in the 2004 election were split evenly at 37% between Republicans and Democrats.

2. Religion - Next, a whopping 19% of respondents had "no" religion, while in 2004 only 10% of voters had "no" religion, and they voted overwhelmingly for Kerry (+36%).

3. Age of Respondents In this poll 31% of the respondents were between 18-34, even though the 18-29 year olds (a slightly smaller demo) only made up 17% of the electorate in the 2004 election. I think it's pretty safe to say that by including 30-34 year olds that number would have come close to the IPSOS sample.

4. Income Level of Respondents - This one is amazing. In this poll 15% of respondents made under $15,000 per year. In 2004, only 8% of voters were in this income bracket, and voted 63-36% for Kerry.

5. Marital Status - In this poll, only 56% of respondents are married. In 2004, 63% of voters were married, and voted 57-42% for Bush.

6. Geography - In this poll, only 17% of respondents were from "rural" areas. In 2004, 25% of voters were from rural areas, and voted 57-42% for Bush.

7. Race - In this poll, there were 71% white respondents and 12% Hispanic respondents. In 2004, 77% of voters were white, and only 8% Hispanic. Bush won the white vote 58-41% and Kerry the Hispanic vote 53-44%.

We're not saying the GOP is going to have an easy time in 2006, but when you read slanted polls like this you really wonder if the purpose of conducting it was to get good results, or push an agenda. Given the history of the AP/Ipsos poll, my bet is on the latter.

No comments: