Joe Lieberman.....46%
Ned Lamont......42.1%
Alan Schlesinger....3%
MoE +/- 3%
This is an online Zogby poll, so the methodology is, as usual, suspect. Still, maybe there's a trend here to analyze. The last Zogby poll showed a lead of 10% for Lieberman.
Then again, a recent Public Opinion Strategies (R) poll showed Lieberman leading by double-digits, as did the last Q-poll in August. Two mid-August polls by Rasmussen and American Research Group showed a much, much tighter race. How much Schlesinger can draw from Lieberman may prove to be crucial.
Bottom line? Who knows? It's probably closer than 16%, but maybe not quite as close as 4%. Here's the trend:
Source"Zogby/WSJ Battleground States Poll." Connecticut Senate. 11 September, 2006.
12 comments:
Don't put different polls on one trend line - its much, much better to use a scatter graph with one trend average line...
Here is what I mean...
np. It used to be mystery pollster, but he teamed up with the guy who has been doing the graphs on his own and voila! graphs + commentary...
Yes, Gabe, trend lines matter. But it's quite rare that a candidate leads every poll before a general and then loses.
Let's see just one- from anybody- where Ned is ahead before we call this election close.
What did my comment have to do with your response?
Just curious.
Did I bring up either candidate or simply state then one graphical method of representing polling is far superior to another?
Actually, I was agreeing with you Gabe. But adding in another thing to look at as well.
I'm not spinning for Joe, BlueCoat- I just challenge anyone to find a candidate who, since Tom Dewey, has led every poll and lost.
Frankly, I'm just a poll junkie with a stats background- and I'm interested in the intersection of numbers, reality and strategy.
As far as any axe-grinding goes, why would the "since Tom Dewey" question evoke such a nervous response?
Might I suggest any on-line based polls are likely to greatly overstate Lamont strength and understate Lieberman strength.
Joe's got the older non-wired set, not the netizens
Anon 9:52 - While they are online based polls, they are still selecting people - I don't think they are more likely to reflect the will of the netizens, just more likely to be wildly inaccurate. Rember that the last WSJ/Zogby Interactive poll was one of the worst for Lamont. A drunk monkey could do better with darts than Zogby does.
If you aren;t online (and most less affluent older voters aren;t) Zocby can't find you
It would be like using telephone polling in the 1948 election
Then explain the outlier from Zogby interactive that had lamont down 15 points?
They are just plain inaccurate...
My educated guess is their sample is whack and they are weighing like crazy to fix it....(and maybe changing the weighing from poll to poll) sorry, you can;t turn chicken droppings into chicken salad.....so the low effective sample size causes the needle to oscillate wildly
Post a Comment