I counseled that it was a wee bit early for that, but I had sensed that there was something off with the poll, but couldn't quite figure out what it was.
Curt at Flopping Aces found it so I didn't have to look very far.
The poll oversampled Democrats. By a big margin.
One would hope and expect that a poll that claims to be a nationally representative sample would take equivalent numbers of Democrats and Republicans with a mix of Independents. Not the case here - nor in several other polls touted by the media outlets lately.
Curt went through the metrics in the poll to find the key details left out of the poll stories - and one that completely undermines the results:
Project #81-5139-72In other words, the published and touted poll results can be nearly completely explained by the oversampling of Democrats and not because of any trend that disfavors Republicans. So, if Democrats want to keep looking at the polls and thinking that they're in a great position to pick up seats in November, go right ahead. You'll be disappointed yet again because you didn't bother to read more than just the headlines.
Interview dates: January 3-5, 2006
Interviews 1,001 adults, 856 registered voters Margin of error: +3.1 for all adults, +3.4 for registered voters
THE ASSOCIATED PRESS POLL
CONDUCTED BY IPSOS PUBLIC AFFAIRS
NOTE: all results shown are percentages unless otherwise labeled.
REGISTERED VOTERS/ PARTY AFFILIATION
1. Are you currently registered to vote at this address, or not?
ALL
ADULTS
Yes………………………………………..81
No ………………………………………..18
Refused/not sure ……………………..1
REGISTERED VOTERS
Strongly Republican …………………..13
Moderately Republican ………………..27
Definitely Independent/neither……….8
Moderately Democrat…………………..32
Strongly Democrat …………………….20
Refused/not sure……………………….-
Total Republican ………………………..40
Total Democrat ………………………….52
UPDATE:
Newsbusters caught this one too.
No comments:
Post a Comment