seems to be slightly higher than Nader's national impact during the previous election. In 2000, exit polls of Nader voters showed the following set of preferences:

46% would have picked Gore
31% would have sat out the election
23% would have favored Bush

Considering Nader's national vote total of 2,883,105, this exit poll estimates that if Nader had not been on the ballot, Gore would have finished with 50.07% of the national vote, while Bush would have finished with 48.91% of the national vote. A 1.16% national victory would have been an improvement of 0.65% for Gore over his actual victory of 0.51%, almost a full-point less than the impact Ed documents in 2004.

So I guess this means that 1.5 points is a pretty big deal. Chris does issue the caveat that third parties generally poll better early on (before the conventions) and close to Election Day as well, but then don't get results commensurate with their poll returns.

For myself, I'll be happier if the Nader Effect either decreases, or Kerry's margins become large enough to swallow it whole, but there is (so far, and we've still got 5 months to go) not much sign of a big breakout for him yet.

Update: Make sure to take a look at the objections to my conclusion of an observable "Nader Effect" from Scott Pauls, who (if I've Googled the right person) is a professor of mathematics, and therefore much more conversant with statistics (and their pitfalls) than I am. His remarks (and my reply and his response) are in the comments thread here, on MyDD.

Update (6/15): Well, it looks like someone had already beaten me to the punch. I've just now found on the website Don't Vote Ralph a study, made about a month ago, which examined the 37 national polls available to that time, and showed that:

Of the 37 polls reviewed, 32 show Nader hurting Kerry, 4 show no effect, and 1 shows Nader hurting Bush (and that by a scant 1%).

[...]

While the percentage swings to Bush are all single-digit, the consensus is overwhelming, directly discrediting Nader’s claims.

In addition to national polls, we found state and special-interest polls that similarly compared Bush and Kerry head-to-head and with Nader added to the mix. Here the results were even more striking. Among other things, these polls (the first six in the above table) show Nader flipping New Jersey and Pennsylvania from Kerry to Bush, and causing an 8% surge for Bush among the large Arab-American vote in four critical swing states. These results alone would almost certainly swing the election to Bush.

The implications of these findings could be enormous. The nation is very closely divided, and it is extremely likely that in some battleground states, these numbers would determine the outcome. When a few percent of voters in a few states will determine the next president, Nader’s independent candidacy could well tip the balance.