Yet another poll

by: Hopeful

Fri Oct 30, 2009 at 04:12:51 PM EDT

Well, I linked to conservative Rick Shaftan's Neighborhood Research poll when I liked the results, so I'll link to today's poll (via PolitickerNJ). His poll of "341 definite or very likely voters" has Corzine at 35%, Christie at 43%, and Daggett at 8%, with a 5.3% margin of error. The sample's almost too small to take seriously, but there it is. We do know from Rasmussen's polls that if you limit the electorate to "definite" voters you improve Christie's performance and from many polls that if you go to less likely voters you improve Corzine's, so this may be a combination of a small sample outlier with a restrictive voter screen.

What is worth checking out is the Shaftan poll analysis (PDF), which gives you a preview of the movement conservative's spin. Putting my cynical hat on, if Christie loses, the conservative Lonegan supporters can easily say "I told you so," and Shaftan has long been on record criticizing Christie, but what can he say if Christie wins?

The release tells us: Christie didn't run as enough of a conservative but Corzine's ads made him into a conservative like Steve Lonegan that the right will be energized to support. So too, Obama's visit will backfire (you should perhaps not ask how a "poll" shows next week's reaction.) And of course Corzine's lack of positive ads don't leave voters with enough reason to pick him if they didn't like Christie. (Okay, I've made that argument one myself, though I know there are some positive ads) In short, next week's election is a victory for Steve Lonegan and a defeat for Barack Obama. (By the way, Democrats' ads won the primary for Christie too!) Shaftan argues that Christie's attack ads have weakened Daggett with conservatives and helped him with liberals. It's plausible, but his sample of Daggett voters is 341 x 0.08 = 27 people. Can you draw conclusions about subsets of a poll of 27 people? No.

Hopeful :: Yet another poll