Pollster.com

Articles and Analysis

 

NJ: Christie 40 Corzine 39 (SurveyUSA 10/12-14)


SurveyUSA
10/12-14/09; 611 likely voters, 4% margin of error
Mode: IVR
(SurveyUSA release)

New Jersey

2009 Governor
Christie 40%, Corzine 39%, Daggett 18% (chart)

 

Comments
Stillow:

3 jersey polls today....+4 christie, +3 corzine, +1 christie....all over the place....the partisans are fudging right now....don't expect real numbers til the last week.

____________________

tjampel:

so it's MOE; some of you expect Daggett's supporters to move to Christie in the booth. I expect some to just stay home and GOTV to decide this election, with the base loyalty determining things. I expect a very low turnout; there's just no love for these two out there.

____________________

Xenobion:

This horse race is so boring. More Palin/Edwards polls please.

____________________

lat:

Stillow,

It's ok buddy. I know you are upset that Christie will lose in NJ and your party will fail again, but you have my sympathies. Hey, look at it this way you are most likely going to win in VA so there is your consolation prize.

____________________

jmartin4s:

This poll had men and woman at a 50% 50% split. I reran the number at 51% 49% with woman having the edge which is fair since the CNN exit polls had 54% women 46% men. The new numbers are
39% Christie
39% Corzine
18% Daggett

____________________

Scooby The Doo:

This race is a dead heat. It's really pathetic to see the much hyped Christie about to sink into the Abyss after having a lead people predicted (on this board) that Corzine couldn't overcome.

Just remember that it's New Jersey and the DEMS have a natural advantage and Corzine has more money so he can do what he pleases.

____________________

Rasmus Pianowski:

jmartin,

this is an off-year election. Turnout dynamics are way different here, and as long you're no expert in this and don't know what exactly you're doing there, it's nothing more than fudging the numbers.

____________________

jmartin4s:

Fudging numbers is a little bit of an overstatement. All I did was make it go from 50 men, 50 woman, to 49 men 51 woman. Every other poll had more woman than men. Seriously, do your research before you criticize.

____________________

Cyril Washbrook:

@Rasmus: prove to me why jmartin4s' analysis is wrong. It's not enough to say "oh, it's an off-year election, the demographics are different". You need to present a case as to why they'd be different and how they'd be different. jmartin4s' hypothesis is perfectly reasonable: there is no trend of women simply not turning out in off-year elections.

____________________

Rasmus Pianowski:

Well, I might have overreacted a bit because there were hundreds of people during the elections that were like 'oh, my candidate is behind by 10 points, but- don't worry- there are 5% too many blacks, and it's totally unreasonable to think that pro-lifers would vote this way, so let's assume it would be different- and look, all of a sudden it's just a 3 point deficit'.

Polling is a science. Saying 'Oh, I think it should be different, let's simply change the numbers' is not.

The CNN exit polls have a sampling error of several percent. Just because they say that the electorate was 54-46 women it doesn't have to be true. There are several instances in which exit polls were WAY off- for example the percentages of black voters in the South in the 2004 presidential elections, which they overestimated drastically.

This poll assumes that the electorate will include 50% women. The population of New Jersey has 51% women. Off-year electorates tend to skew Conservative, women are more liberal than men.

To assume that the electorate will include 50% women is reasonable, just as reasonable to assume that it will include 51% women (as you do), 52% (as PPP does) or 54% (as the 2008 Exit polls suggest).

But to simply change the numbers of a poll is just invalid. You don't need polls then at all, you could simply make up the numbers. 'Let's say the electorate will be 54% women, and they will break 55-40-10 for Corzine. Sounds great'.

This is basically what you're doing, and it's worth nothing.

'Fudging numbers is a little bit of an overstatement.'

Merriam-Webster says:

1 a : to devise as a substitute : fake b : falsify

Exactly this is what you did there.

____________________

jmartin4s:

You are being rediculous. What I did is what a lot of mathematicians and staticians do and many intelligent people, professors, and researchers would take extreme offense to your comment. This summer there was a poll from Quinnipiac with Specter vs Tommey and Nate Silver at fivethirtyeight.com reran the numbers using the exit polls to see if there was any difference. I'm sure you wouldn't criticize him one bit for doing. There is nothing fake or false about what I did and it shows you have very little intelligence to say that just because you disagree with my analysis and as shown above Im not the only one who thinks that. I suggest you stop bashing other people's opinions and analysis when you comment on this site.

____________________

John:

Electorates normally do have a larger number females than males, in 2006 of the 24 governor races which had exit polls 22 of them had more females than males. (NY tied and Nebraska had more males than females). While it is true that individual exit polls can be off on demographics, NC in 2004 for example, it not is highly unlikely that they all are. (it's also probably not due to sampling error but that is a different matter).

Pollsters generally set the number of females they have in the polls as traditionally females have a higher tendency to respond. Looking at SUSA this poll was 50/50 and their last poll 51/49 (males/female) so it looks like they are not setting the number or perhaps only have an upper limit on the numbers of females in their electores before they will step in. Whether this is a 'good' policy is an interesting question.

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR