Monday, October 26, 2009

State of the Race: New Jersey Governor (10/26/09)

[Click to Enlarge]

A nine point Corzine lead? According to a new Suffolk survey of the Garden state, that is the case. However...
  • It was Suffolk's first poll in the state for this race.
  • The sample size is on the small end; only 400 people.
  • There were a lot of undecideds (14%). The last time there was anything in the double digits for undecideds was the September 9 Rasmussen release (10%). Let me add some context: that was "You Lie!" week.
  • Being that it was Suffolk's first poll in the race, they may have committed a grave error; weighting party identification to party registration. That's a no-no in New Jersey with unaffiliated registration.
Is the poll something to be dismissed? No, but it should certainly be treated as an outlier. At one end of the spectrum (an extreme application of the margin of error), if Corzine cedes five points to Christie, the Republican has a one point edge. If you were to do Monte Carlo simulations given the data in this poll, that particular outcome wouldn't come up very many times though.

2009 New Jersey Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Corzine
Christie
Daggett
Undecided
Suffolk [pdf]
Oct. 22-25, 2009
+/- 5%
400 likely voters
42
33
7
14

In FHQ's opinion, though, there is one saving grace to this survey: the Daggett result. Sure, perhaps that number should be written off as an outlier given the independent's polling numbers of late. But I think there's something to this and we all should have been tipped off to the possibility a few weeks ago when Fairleigh Dickinson released a poll with a split sample where Daggett's name was used against Corzine and Christie with one subsample and Gary Steele, another candidate for governor, was used in the other subsample. I was too busy that day bemoaning the fact that we couldn't really trust the results of such a small subsample. All the while, though, there was an interesting trend underlying that poll and now this Suffolk one: Daggett has been afforded a privileged position in most of the polls in which he has been included. The independent has been the only non-Corzine/Christie candidate to appear in many of these polls and as such has been adopted by disaffected survey respondents as the anti-Corzine/Christie; a refuge for all wanting to avoid casting a ballot for either of the two major party candidates.

When other third party names are included Daggett's support slips. Sure, he bested Gary Steele when the two subsamples of that Fairleigh Dickinson poll were compared, but the fact that Steele got over 10% was indicative of the fact that there was a fairly sizable portion of the likely electorate that did not particularly care for either Corzine or Christie.

In today's Suffolk poll, all 12 gubernatorial candidates that will appear on the ballot next Tuesday were named. Again, Daggett lost his position as the only Corzine/Christie alternative and was knocked down to earth to some extent because of it. FHQ has already discussed Daggett's ballot problem and this poll may have picked up on what Daggett faces in the ballot box next week: that instead of one alternative to Corzine and Christie, there are several. Daggett's problem now appears to be that survey respondents recognized in him an alternative to Corzine and Christie and didn't necessarily recognize him.

We won't know one way or the other until next week or until we get some more data along these lines (with hopefully a bigger sample).

[Click to Enlarge]

But what about the top two contenders vying for a chance to call Drumthwacket their home for the next four years? I don't think that it comes as a surprise to anyone that posting a 33 in this poll negatively affected Chris Christie's average. It did. In fact, it brought the former US attorney's average closer to the 43% mark; again uncharted territory. Meanwhile, Corzine inched ever so slightly further up, but the incumbent Democrat still pretty much sits on the 39% line. So, while Christie maintains a four point lead here, that is shrinking quickly. That's because he is doing far more falling than Corzine is jumping, though.


Recent Posts:
State of the Race: Virginia Governor (10/26/09)

Wrong! Wrong! A Thousand Times Wrong! One Bit of Misinformation from the Democratic Change Commission's Meeting This Past Weekend

New 2012 Presidential General Election Trendlines. Now Time Adjusted!

3 comments:

Simple thoughts said...

you show pick up the polls taken in the NY-23 race

Josh Putnam said...

NY23 is certainly shaping up to be a closer race than Virginia is.

Honestly, I've kept an eye on the updates in the polling in that race and tweet the new ones as they come in. But I'm wary of the total number of polls there. As the Georgia Senate runoff demonstrated last year, FHQ's averages don't do well with just a handful of polls.

I will say this: I'll look into it.

Josh Putnam said...

I should probably provide a link to that GA Sen projection if I'm going to cite it.

Projection

Results

The same problem exists in both the Georgia case and for NY 23: no one knows what turnout is going to be like (other than lowish), so it is difficult for the polls to identify who the likely voters are.