Time Poll Exposed

by Chris Bowers, Tue Sep 14, 2004 at 06:40:37 PM EDT

I never really thought there was anything wrong with the Time poll that showed Bush up ten, and that it was just an outlier. However, enterprising KevinO puts that to rest in a dailykos diary: Go to the following site, which has the full 20 page report:

http://www.srbi.com/condensed-data-2004-5.pdf

and go to page 19. And there you see STUNNING EVIDENCE that the survey methodology is weighted towards finding GOP leaning Democrats and Independents. It asks respondents who they voted for in 2000. BUSH 53% GORE 41%
Gee, that's not how the vote came out four years ago, did it?

In fact, if you adjust the numbers for this huge bias, Bush is actually TRAILING his performance of 2000, leading by 11% among an population that voted for him by 12% last year. That would seem to indicate a dead heat, or a slight Kerry lead.

It is one thing to make an argument about how Party ID should not be weighted since it is not necessarily a completely stable demographic determinant, but it is another thing entirely to obviously over-represent 2000 Bush voters in a poll. That's just plain unrepresentative crap.

Tags: General 2008 (all tags)

  • 6 Comments
  • 0 recs
  • Share
  • Permalink

Comments

6 Comments

Shadow_head_small
Don't you know when your being propagandized?
But the media, even reportes for the home town local newspaper attending C-SPAN events are now quoting the common wisdom, Bush is way ahead. So the ploy worked.

Bush and Rove are all about the BIG lie.

They know how to make it work for them, and the media has been complicit. As have naive Democrats who tend to give the media like Newsweek and Time credit before the fact for being imparital and rigourous in their methods.

They are not.

The media, its owners and its polling organizations are actively working for a specific outcome, and their reporters know who pays their salaries and what they should to do get ahead.

by leschwartz 2004-09-14 07:27PM | 0 recs
Shadow_head_small
the post facto epsitemological prism
Here's a grain of salt to take with your conclusion:  it's well documented that after any election, the number of people who claim to have voted for the winner is greater than the percentage who actually did vote for the winner.

People, memories, and convictions are much weaker than we would like to hope...

by globecanvas 2004-09-14 07:40PM | 0 recs
Shadow_head_small
Re: the post facto epsitemological prism
Agreed, this is no kind of expose.  The Bush-Gore gap means nothing and is actually about what we would expect given past results.  Saying that this poll or that poll is biased against us isn't helping anything.  We need to face the facts and get out there and work our butts off for Kerry.  It's pretty clear that Bush is ahead by a few points nationally and is probably ahead is most Bush 2000 states + WI (except maybe being tied in FL and certainly losing in NH).  Of course this isn't about doom & gloom for our guy, I'm just tired of the Dean-esque, "de-bunking" of polls.  This election still hinges on any external events and the debates, after which I'm betting someof "wrong direction" folks who are still sticking to Bush will take the plunge and support Kerry.  We can win, but we need to work on voter reg., and GOTV and quit seeing bias in every poll we don't like.
by asearchforreason 2004-09-14 07:54PM | 0 recs
Shadow_head_small
How well documented?
I have heard this claim, but where is the documentation?
by EvanstonDem 2004-09-14 08:17PM | 0 recs
Shadow_head_small
Re: How well documented?
You have not been paying attention, see the previous posts by Chris Bowers. Documentation provided in full.
by leschwartz 2004-09-14 08:57PM | 0 recs
Shadow_head_small
conventional wisdom
The main problem with polls is how the media love them -- it requires no research or other work to write a story about who is ahead in a poll.  Do they become a self-fulfilling prophecy?  Maybe yes, to some extent -- the idea that a voter says "well, I want to vote for Kerry but everyone else seems to be voting for Bush so maybe I should too" or "why bother voting for Kerry when Bush is going to win anyway".  But in Canada, during our election in June, the polls showed Martin and Harper neck and neck all the way - then Martin won 135 seats and Harper 95.  Things turned in just the last few days -- only the tracking polls picked up the change.
by CathiefromCanada 2004-09-14 10:06PM | 0 recs

Diaries

Advertise Blogads


----------- myDD - skin -----------