Defeat defeatism!
Posted on 09/28/2006 11:28:36 AM PDT by okstate
The latest Keystone Poll finds incumbent Republican Congressman Jim Gerlach locked in a tight battle for re-election with challenger Democrat Lois Murphy. The poll of 431-registered voters conducted by the Center for Opinion Research at Franklin & Marshall College finds Gerlach with a 45-to-38 percent edge.
When narrowed to likely voters, Gerlach's advantage is down to 44-to-41 percent.
Gerlach scored a two point win over Murphy in 2004. Poll respondents list the war in Iraq, the war on terror and healthcare as the top issues. The sixth district covers portions of Berks, Montgomery and Chester counties.
The likely voters sample was only 349 people in size, which is extremely small -- you would generally want to see about 500-800. A GOP poll taken a week ago had Gerlach up 11. Constituent Dynamics had Murphy up 5 at the end of August, and a Dem poll had Murphy up one in late July.
Gerlach bump!
And if Gerlach, considered a bellweather, is ahead by 5 (and again, ALL these polls are failing to include appropriate numbers of GOP), then it likely means most of the GOP House people are pretty safe.
Extremely relatable to PA-09, I'm sure.
"If you think the Dems will have a better ground game, you're sadly mistaken. What I'm hearing is that the Dems have NOT been contacted, that the campaigns are STILL using the old Union thugs/moveon idiots/paid people while we have volunteers. ALL the info I got, save one small demographic, showed not only a motivated voting group, but a superior ground game."
I do not doubt that we have a superior ground game. The point is that the LV/RV split in polling is reversed this year. All that means is that RIGHT NOW Democrats are more motivated to vote as a whole than are Republicans. And in terms of this poll all it really means is that Dems are more motivated to vote in PA-06, which, incidentially, is NOT located in Ohio.
"And if Gerlach, considered a bellweather, is ahead by 5 (and again, ALL these polls are failing to include appropriate numbers of GOP), then it likely means most of the GOP House people are pretty safe."
This is definitely positive news for House GOP members on the whole, especially when there aren't any special circumstances involved (by special circumstances I mean districts like AZ-08, TX-22, and CO-07). It bodes extremely well for Weldon and Fitzpatrick. I don't think it says much about the three troubled incumbents in Indiana, though.
"Extremely relatable to PA-09, I'm sure."
Meant PA-06.
But if you think that the situation in OH has nothing to do with PA-09, you're sadly mistaken. You've bought this polling data hook, line, and sinker. But have you actually gone out and talked to any of the committe chairmen? The county people? In your state? Any where?
The RVs are in our favor, and LVs have ALWAYS been in our favor. Don't know how far back you'd have to go to find higher % Dem turnout. Won't happen in 06, either.
How about the headline "Gerlach Expected To Exceed 2004 Election Margin."
From the story: "When narrowed to likely voters, Gerlach's advantage is down to 44-to-41 percent.
"Gerlach scored a two point win over Murphy in 2004."
Not downplaying the race, just pointing out my point of view
The 3 Philadelphia suburb seats (PA-6,7,8) are kinda depressing because the demographic trends are so badly against us. We can probably hang on for another year, but I doubt for too much longer.
Incidentally, it may end up that the most endangered PA Repub incumbent is in a district with a very favorable partisan breakdown (PA-10). But that's due to local mistress-ish issues.
I live in PA 06. I think Gerlach is looking pretty good right now. The GOP poll had him up 11. This one has him up 8. I don't buy the fact that the Democratic voters are more motivated than the GOP voters this year, so I'm not sure I believe that the gap narrows when you're looking at likely voters.
I know I'm motivated. Lois Murphy is an out-and-out leftist. She's not a Tim Holden Democrat, or even an Ed Rendell Democrat. She's a Hillary Clinton/Nancy Pelosi Commie. If she wasn't so far to the left, Gerlach might be in trouble. But I don't think this district is so left that we'll put Lois the Lefty in there. Gerlach is a good conservative.
Not in most of the polls I've seen.
Honestly I haven't seen that many that actually show an RV vs LV results breakdown. A notable example was the five Senate polls Gallup released about a month ago. Most other companies don't break it down that way (Survey USA, Rasmussen, and others report only likely voters). I just find it interesting in a case like Gerlach's district that the split is reversed from normalcy. It shows me that Gerlach will more than likely get a lot of Republicans to "come home" to him by November and that he's in better shape than this poll would suggest.
the Crosstabs blog on Redstate has talked a lot about this reversed LV/RV phenomenon this year. www.crosstabs.org
I think it has a lot to do with the media. The continued stories in spring and summer that the Rats were going to take the House and maybe the Senate made a lot of self-identified Republicans say they were less likely to vote in November. But now that the media is easing up a bit on those stories we may see the LV/RV disconnect reverse and go back to its typical state of affairs.
In talking to people in my office. Even Dems don't like Liberal Lois Murphy that much. Everyone that has come in, says that her campaign is just ALL negative and they will be voting for Jim.
Yes, that's why I specifically mentioned the small sample size of about 350 here. But you are incorrect about an 800-1200 sample size putting him up 8-12%. There is an equally likely chance of that as there is that a big sample size would put Murphy up 4-8%.
And there are diminished returns for accuracy when sample sizes get too big. Sure you can poll 1200 people instead of 600, but the results are not going to be twice as accurate. Instead the 600 people will have an error of about 4 percent or so, and the 1200 will have an error of about 2.8 percent or so. The cost of interviewing those extra people is not worth the tiny increase in accuracy.
I am going to have to call "bunko" on this one. In my 12+ of being politically active I have never seen this to be the case. And I mean never.
Polling data used by registered voters usually favors democrats because it takes little to get registered. Likely voters come from asking people if they have voted in past elections, not asking if they plan on voting. People with established voting history are more likely to vote GOP> Sure when you ask that to a bong brained college student he will say yes, but on election day they are always too lazy to get off the couch.
More correctly 600 would have a 10% error (90% confidence) and 1200 would have 5% error (95% confidence) provided that the sample is scientific and polled voters based upon the demographic characteristics of the population.
If the method is not scientific and based upon innacurate demographics the %error starts to increase exponentially and the sample is deemed useless.
Ummm.... did you read my comment? I am saying that this year is different for many reasons, one of which is the reverse in the LV/RV split (in some polls). Yes, I know that Democrats NEVER do better in LVs than in RVs, but for some reason they are this year in some isolated races. I don't know what it means, but I have some guesses. Check out crosstabs.org, a blog on Redstate, for more information and analysis about these trends.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.