Scott McAdams (D): 28
Joe Miller (R): 43
Frederick Haase (L): 2
Lisa Murkowski (volunteered): 18
We saw two polls yesterday finding Lisa Murkowski either winning or in a dead heat, but those polls made no mention of Murkowski's unusual status as a write-in candidate, which should be viewed as a methodological problem. But now we have the first poll that seems to deal head-on with the write-in problem, and as an added bonus, it's Ivan Moore, probably Alaska's best-regarded local pollster. The result shows that Joe Miller might want to stop measuring the drapes.
Stay the same: 55
Write in Lisa Murkowski: 31
Not sure: 15
Scott McAdams (D): 14
Joe Miller (R): 36
Frederick Haase (L): 2
Lisa Murkowski (WI): 44
So here's what Moore did: his first round of questioning omitted Murkowski but permitted respondents to volunteer her; then he asked, as a follow up question "As you may know, Lisa Murkowski is running a write-in campaign for U.S. Senate. Knowing this, would your vote for U.S. Senate stay the same or would you write in Lisa Murkowski?" The second set of results are then re-computed based on prompted Lisa Murkowski votes. That still may not re-create the actual voting experience, where there isn't a voice reminding you that Lisa Murkowski is running. But this still seems an improvement from just listing her with the other candidates.
Here's Moore, with some circumspect analysis of what he tried:
Now, the reality of the situation is that neither of these results is going to be correct. The first should be perceived as a minimum for Lisa, the latter a maximum. The reality lies somewhere in between... the question is where? Personally, I think the second measure is what will happen in an ideal, impediment-less world, and should be adjusted downwards by what we'd reasonably expect the attritional effects of the write-in to be. I have always maintained these will be relatively minimal, maybe not much more than a few percent of people who somehow remain unaware come election day that Lisa is an option, or get her name wrong, or don't fill in the oval, or decide they can't be bothered to write a name."
There are also results from the (comparatively uneventful) gubernatorial and House races.