FL/OH/PA: Quinnipiac: None of us are swing states

[tokar]

You cant keep trashing every pollster until you get a result you like...jeez.

Quinnipiac does not poll every state, it polls a limited number of states.  So to say that they suck in all these states starts limiting their abilities.

They poll NY, NJ, PA, CT, FL and OH.  And thats it.  No more, no less.

There is no reason to poll NY, CT and NJ because they are ultra safe dem states.  So that leaves them with PA, OH and FL.  OH might be a bit excessive, but PA is right in the mix of all the polls which show Obama+10 or more, as is Florida.  Florida has been back and forth, showing Obama up by a bit, then McCain a bit.  So to just throw out the Florida poll is dumb.  Its just part of the mix of polls which show Florida a slight Obama advantage.
But PA, come on now...every poll is showing Obama+10.  The last poorish poll for Obama was a MONTH ago.  Again, no one trusts that the margin of victory will be +10...the point of the polls are to show the trend and the trend is Strong Obama.

[tokar]

Quinnipiac is a very good poll in some states, but not in any of these three.

What states are they good in then?I remember hearing in '06 during the senate and governors races how Quinnipiac was the best in PA. Are you saying this because its a result you don't like?

That is exactly my point.

If they aren't good in PA, then what states ARE they good in?

If people aren't aware, Quinnipiac has historically polled only SIX STATES.

Connecticut
Florida
New Jersey
New York
Ohio
Pennsylvania

However, this election cycle they have stepped out of their bounds and polled Colorado, Minnesota, Michigan and Wisconsin a bit.

But based on their total polling history, I would say the states which they focus MOST on are Ohio, Pennsylvania and Florida.

Here is the general election count so far (Number - State):
13 - Pennsylvania
12 - Florida
12 - Ohio
6 - Colorado
6 - New Jersey
5 - Minnesota
5 - New York
5 - Wisconsin
2 - Connecticut

In the Democratic Primary season here are the polling numbers:
18 - Pennsylvania
15 - Florida
15 - Ohio
9 - New Jersey
8 - New York
4 - Connecticut

Here are the 2004 polling numbers:
11 - Pennsylvania
9 - New Jersey
7 - Florida
5 - Connecticut
4 - New York


I think they started polling Ohio after it was close.  But as history shows, they are a 6 state pollster.  OH-PA-NY-NJ-CT-FL.  New York, Connecticut and New Jersey are rarely worth polling since no one actually things they have a chance to go republican.
So that leaves you with Ohio, Pennsylvania and Florida.  Seriously, if they are just focusing on THREE states, how bad can they be?

They did a good job in their polling in 2004, and ditto for their 2008 primary polling.

Personally, when I think of polling outlets which do best in PA I put Quinnipiac at the top of my list.

[tokar]

Quinnipiac is a very good poll in some states, but not in any of these three.

What states are they good in then?I remember hearing in '06 during the senate and governors races how Quinnipiac was the best in PA. Are you saying this because its a result you don't like?

No, Tokar, it's because Q isn't hugely accurate outside of NY, New England, and is fair in NJ.

Q begins to have problems outside of the area; it's a uni poll after that.

Accuracy...what type of accuracy.

Accuracy in margin?  What pollster is accurate enough to be considered "trusted"?  This is why, when a pollster DOES get the margin right, they have a coming out party, with press releases, pat-on-the-back ceremonies, blah blah blah.  If pollsters routinely got the margin of victory right you wouldn't see this kind of crap and we wouldn't have 20 or more polling outlets.
Accuracy in trend?  Pollsters are very reliable when it comes to trends.  I've gone through this with you before, and I showed you that Quinnipiac has a very good history when it comes to trends.

As I've try to explain to you time and time again, polls are meant to be used more to gauge the temperature of a race, not predict the margin of victory.  People may THINK that pollsters are trying to predict the margin (there are a few people on here like that *cough*), but I don't think that way.  I look at polls just to gauge the race and every...single...PA...poll...in...the...last...two...weeks shows Obama outperforming McCain by a large margin and pulling above 50%.

Am I going to harp on Quinnipiac or any other pollster for not getting the trend right?  Nope.  But I will pat them on the back for getting the trend exactly right when, on November 4th, PA goes for Obama.  There has not been a SINGLE poll showing McCain ahead since late april, well before Obama clinched the nomination.  Theres something to be said about that.

[tokar]

Quinnipiac is a very good poll in some states, but not in any of these three.

What states are they good in then?I remember hearing in '06 during the senate and governors races how Quinnipiac was the best in PA. Are you saying this because its a result you don't like?

No, Tokar, it's because Q isn't hugely accurate outside of NY, New England, and is fair in NJ.

Q begins to have problems outside of the area; it's a uni poll after that.

Accuracy...what type of accuracy.

Accuracy in margin?  What pollster is accurate enough to be considered "trusted"?  This is why, when a pollster DOES get the margin right, they have a coming out party, with press releases, pat-on-the-back ceremonies, blah blah blah.  If pollsters routinely got the margin of victory right you wouldn't see this kind of crap and we wouldn't have 20 or more polling outlets.
Accuracy in trend?  Pollsters are very reliable when it comes to trends.  I've gone through this with you before, and I showed you that Quinnipiac has a very good history when it comes to trends.

No, in this case Q isn't accurate either way outside of that NY/New England area, where they usually poll.  It isn't particularly good at showing trends either.

A poll is basically suppose to be a snapshot of the electorate.  Sometimes they show trends, especially trackers, but usually they don't.  That is the reason I look to tracking polls for trends.

Can you stop saying "New England area"?  You obviously have not looked at Quinnipiac's polling record.  The only state that is trully in the "New England area" that Quinnipiac polls is Connecticut.


I thought I showed you this before...I guess not.

Lets go through it again.

PA DEM Primary 2008 - strong polling for Clinton.  Result: Strong Clinton
OH DEM Primary 2008 - strong polling for Clinton.  Result: Strong Clinton
NY DEM Primary 2008 - strong polling for Clinton.  Result: Strong Clinton
FL DEM Primary 2008 - strong polling for Clinton.  Result: Strong Clinton
NJ DEM Primary 2008 - strong polling for Clinton.  Result: Strong Clinton
CT DEM Primary 2008 - strong polling for Clinton.  Result: Strong Clinton

They polled NO OTHER STATES during the 2008 Democratic Primary.

NY GOP Primary 2008 - strong polling for McCain.  Result: Strong McCain
NJ GOP Primary 2008 - strong polling for McCain.  Result: Strong McCain
FL GOP Primary 2008 - weak polling for McCain.  Result: Weak McCain

All other polls in GOP Primary were too early, from 2007, when Guiliani was cruising.  They polled no other states for the GOP Primary following February 5th.


Lets go back to 2006 Senate:

CT Senate 2006 - strong polling for Independent.  Result: Strong Independent
PA Senate 2006 - strong polling for Democrat.  Result: Strong Democrat
NY Senate 2006 - strong polling for Democrat.  Result: Strong Democrat
OH Senate 2006 - strong polling for Democrat.  Result: Strong Democrat
NJ Senate 2006 - weak polling for Democrat.  Result: Strong Democrat
FL Seante 2006 - strong polling for Democrat.  Result: Strong Democrat

They polled no other states in the 2006 Senate race.  Only in NJ did they go wrong.  They had it as a rather weak hold, but it turned out to be a 9+ point win for the democrat.


Now lets go back to 2004:
PA 2004 - weak polling for Kerry.  Result: Weak Kerry
FL 2004 - weak polling for Bush.  Result: Weak Bush
NJ 2004 - weak polling for Kerry.  Result: Weak Kerry
CT 2004 - strong polling for Kerry.  Result: Strong Kerry
NY 2004 - strong polling for Kerry.  Result: Strong Kerry

They polled no other state in 2004.



So I am not sure where you are getting YOUR data, but the data available on THIS SITE shows that Quinnipiac has a very good track record.

[tokar]

Actually, Tokar, Q was 6 points off for Casey in 2006 and about 7.5 off the margin.  Muhlenburg, their general poll, was more off actually with the same gap as Zogby.

Keystone (which I thing is F & M) came closest.

I think I made pretty clear in my post that I was not trying to show that the margins were accurate, rather the TRENDS were correct.

You are fool to take these margins in the polls as "predictors" for the actual margin of victory.

Again, I don't think pollsters are trying to accurately predict the margin of victory, rather they are making polls to gauge the temperature of the race.  And right now the temperature says STRONG OBAMA, and every pollster is showing this exact trend.

If pollsters were out there to accurately predict the exact margin of victory they would do all their polling internally and then on the last day they would release a poll.  The fact that pollsters release polls like every 3 days should be a pretty clear indication that polls are meant to be used as a temperature gauge.

If the actual margin ends up being +7, +8 or even as high as +15, it pretty much shows the pollsters were all correct in identifying PA is STRONG OBAMA.

Again, it would be one thing if Quinnipiac was polling PA at +20 Obama or +20 McCain right now.  But with ALL...OTHER...POLLING...OUTLETS polling PA at +10 Obama I think Quinnipiac is doing a good job of gauging the race as it is.

[tokar]

Let's let Tokar have his pro-Obama numbers. 

Considering they are in line with the rest of pollsters, I will take them.

[tokar]

Let's let Tokar have his pro-Obama numbers. 

Considering they are in line with the rest of pollsters, I will take them.

Actually, the rest of pollsters have shown a decline for Obama in PA.  Some of them are well out of the MOE.

Its not even worth arguing with you.  You obviously can't read polls or trends.  If you seriously can not read that the polls have been, and still are "strong Obama" then you are hopeless.