Ipsos/Reuters (8/13-15, likely voters, no trend lines):
Jack Conway (D): 40
Rand Paul (R-inc): 45
Undecided: 15
(MoE: ±4.7%)
Ipsos has dipped their thumb into the Kentucky pie for the first time this week, finding a 40-40 tie between Jack Conway and Rand Paul among registered voters, and a five-point Paul advantage among likely voters.
Jonathan Singer, formerly of MyDD, has some thoughts at his new blog, Polising, on Ipsos’ likely voter screen that are well worth considering. In screening their registered voter sample down to a pool of likely voters, Ipsos has whittled down their sample from 600 to 435 voters — suggesting a voter turnout at a level close to 73% of registered voters. That number is over 20 points higher than Kentucky voter turnout in the past three midterm elections, so it’s quite possible this sample has a few unlikely voters in its midst. Whom that would benefit, however, is up to debate.
i’d say paul by perhaps 8 is reasonable. conway can definitely come back
that Conway is still in this race. Even with the environment falling into complete and utter crap for Team Blue, Paul hasn’t really been able to get further ahead than 5-7 points. If the mood softens a bit and Conway runs a strong campaign he can still pull this one out, which would, I think, put him into the seat virtually for life unless he sees another year like this one someday.
I am a Republican and I would probably vote for Conway. I under no circumstances could support a candidate that does not believe we need a foreign policy.
You’re saying mid-term turnout in Kentucky is normally 50 some odd percent of registered voters? The number I’m used to seeing when people talk about “turnout” is a percentage of eligible adults. Most eligible adults are registered, but by no means all. Which would mean a “normal” Kentucky turnout would be somewhere around 60-70% of registered voters.
But you may actually mean what you said, in which case normal Kentucky mid-term turnout is pretty awful indeed.
of Republicans who’ve heard about the Aqua Buddha stuff at Baylor, 12% say it would make them MORE likely to vote for Paul. Good grief. http://blogs.reuters.com/front…
For instance, the drug issue in Eastern Kentucky, where he stated that it wasn’t a pressing issue. Conway has an opening there, he’s been the Attorney General and can counter Paul’s claim easily. This is another one where localizing is the best strategy.
Singer postulates that the poll posits that 73% of registered voters will vote, and since that’s absurdly high, some of the poll’s “likely” voters aren’t likely at all.
I don’t think that’s a reasonable conclusion.
The problem with Singer’s argument is that there’s no way to conclude that the registered voters who responded to the poll are representative of registered voters in Kentucky. There are lots of no-answers and refusals to take to the poll in the calling. It’s a reasonably hypothesis that registered-but-UNlikely voters are more likely, perhaps by a wide margin, to refuse to take the poll than “likely” voters. After all, if you know you’re not interested in the upcoming election, why would you take time out of your day for a phone survey on a subject you don’t care about?
I think it’s plenty reasonable to believe that 73% of registered voters who are willing to take the poll are “likely” to vote.