« General Election Cattle Call, June 10 | Main | Swing State Roundup »
Friday, June 11, 2004
The Empirical Nader Effect
Posted by DavidNYCWhen I see a poll that does a version with and without Nader, I often feel that there's something of a Nader effect - ie, that Ralph is distinctly drawing more votes away from John Kerry than from George Bush. Ed, a frequent commenter here and proprietor of his own blog, Unfutz, has actually crunched the numbers and turned a suspicion into cold, hard fact.
Nationally, says Ed, Nader draws 1.53% from Kerry. It doesn't sound like a lot, but in a very close election, such a margin can mean a great deal. At MyDD, Chris calculates the 2000 Nader effect (based on exit polls) at 0.65%. This means that right now, Nader is hurting Kerry almost a full point worse than he hurt Gore four years ago.
As Chris points out, Nader is likely polling far better now than he actually will on election day. Several polls have shown Nader pulling an implausible 8% in various states. What I'd love to see now is what kind of Nader effect polls in June of 2000 were showing.
Posted at 01:07 PM in General | Technorati
Comments
I'd like to see what if it's greater in swing states. I mean here in Washington and down in Oregon Nader got more than nationally, but with the Nader Trader web page and the fact that there was a greater incentive to vote for Gore in a swing state, the effect of voting for Nader might not be as great as the numbers suggest.
Also the Nader effect be tougher to quantify in exit polls than in telephone polls. I mean if someone asked you "ok now pretend that Nader wasn't on the ballot who would you have voted for" the answer might be different than if people just read Gore and Bush's names. And of course it doesn't include anyone who stayed home because Nader wasn't on their state's ballot (I know but stranger things have happened).
Posted by: Carl Ballard at June 12, 2004 02:55 AM | Permalink | Edit Comment | Delete Comment
I never see any mention of how well Nader is doing in getting on state ballots. A valiant effort by WV's Mountain Party at their convention awhile ago kept him off of the ballot in WV. So the first WV poll we've seen that had Nader polling at 3% and putting Kerry 4 points behind Bush is essentially bogus. To the extent that polls are valid indicators the 1 point deficit that resulted with Nader off the ballot was the only true measure. I've suspected that Nader experienced similar frustrations getting on the ballot in other states. Possilby fatal frustrations. If he had to get on the ballot for the primaries to be eligible to be on the ballot for the general election then this issue is already resolved. In how many states will Nader appear on the ballot in the general election?
Posted by: Bill Ambrose at June 12, 2004 09:32 AM | Permalink | Edit Comment | Delete Comment
My understanding is if the Greens endorse Nader (which could still happen) that brings with it ballot access in quite a few states.
Posted by: Ed Fitzgerald (unfutz) at June 12, 2004 10:12 PM | Permalink | Edit Comment | Delete Comment
Nader polled around 7-8% in Pennsylvania in 2000 but registered only 2% at the polls. If he had a 0.65% effect in 2000 I'd expect it to be much lower. Polls are simply unreliable for third party candidates for a variety of reasons - including the knowledge of those being polled that their votes don't really count.
If Nader gets nominated by the Greens he would have access to state ballots. Here in PA i've heard folks protest that they don't want to sign a petition that would give Greens Presidential Ballot access. But its the same petition that you use for local ballot access (that is how it works here) so you have little choice. And since most petitioning happens before the Party decides on a candidate (an issue Greens should resolve) you have to list a Presidential Candidate (usually a placeholder) in order that you don't undermine Party democracy.
Posted by: seamus at June 16, 2004 10:32 AM | Permalink | Edit Comment | Delete Comment
Hey Ed,
It is exactly things like these:
"Ed, a frequent commenter here and proprietor of his own blog, Unfutz, has actually crunched the numbers and turned a suspicion into cold, hard fact."
and
"This means that right now, Nader is hurting Kerry almost a full point worse than he hurt Gore four years ago."
(emphasis mine)
that really worry me. For those interested, Ed and I have been having a dicsussion of the statistical validity of his results (as well as a general discussion of the validity of comparing results across polls) over at myDD where I first saw reference to this analysis.
Short version of my take: the polling data is not specific enough to make the claim that Nader hurts Kerry.
Posted by: Scott Pauls at June 17, 2004 09:43 AM | Permalink | Edit Comment | Delete Comment
Scott,
I just put a comment in over at MyDD. I think this analysis gives a wrong impression bytrying to compare polling numbers with exit polls. In 2000 there was a huge difference between polling numbers and exit polls. The only way to do a comparative analysis at this point would be to compare polling numbers with polling numbers. Unfortunately even this will be highly unreliable since it still won't be able to guage the percentage of polled Nader supporters who will break for Kerry. That percentage will undoubtedly by higher than it was for Gore absent some stupid campaign maneuvers by Kerry (like making Ashcroft his VP).
Posted by: seamus at June 17, 2004 12:17 PM | Permalink | Edit Comment | Delete Comment