Pollster.com

This is bad for dems, how did their advantage deteriorate so fast? They had a like an 8-10 point lead only like a couple of months ago.

This result is clearly an outlier. There is no plausible explanation for why a stable Democratic advantage of 6-10 points in primary identification, 10-14 including leaners, would have been erased. Other recent national polls show a continued and substantial Democratic advantage. Moreover, this is the same survey in which Gallup found a shift in public attitudes on abortion including, for the first time, a "pro-life" majority. But that result is clearly suspect given the skewed party identification of the sample.

Not bad news for Dems. Just a Gallup sample screw-up. If anything it simply explains the huge shift in their abortion numbers. The clear warning was there when they said in the original write-up that the movement came entirely from Republicans. Fox, CBS and Ipsos didn't find anything like parity in their recent national polls.

There's typically always mroe Dems than GOP....but I would not be surprised if these #
's are accurate, Rasmussen also shows the gap narrowing. Not to pick a fight on this, but there's no debate Obama is deploying socialist methoology....that could be having an affect on some conservative Dems who could swap parties or some indy's.Look at a state like PA, during the primary season last year the state reported it had 400,000 more dems than GOP registered, that has narrowed to 200,000 from the last state I read.
I think what happened to Bush could veyr well happen to Obama, which will in turn have an affect on party ID. A presidents personal likeability and favorability eventually gives ay to poor policies. Bush was highly likable to most people, but his policies were not, in the end it all caught up to him. Obama has the same risk, perosnally likable to most, but if you look at the polling, his policies are not liked by majoirties. Thus while Bush may have direct more Dems to pop up, Obama may end up directing more GOP to show up.
A good indiciation if this is true will be the 2010 congressionals, which I belive GOp will pick up many in the House and I think they also pick up seats i nthe Senate. Unless obama stops engaging i nthe same debt spending policies of Bush, he will lose support and as a result do the same to his party Bush did to his. Party id changes all the time, it always has and it always will...and aitting president has much to do with it.

Fox and CBS polled after the Gallup dates and Ipsos just before. The Dem ID lead in each case was 12, 11 and 10 respectively. Respectfully, Stillow, I disagree.

conspiracy

No problem...that is what its all about. Agree to disagree and see what happens. There's no question the GOP is at one of its lower points, so I think everyone agrees there are more registered Dems than GOP. I persoanlly see very little difference between the parties these days, but that is just me. 2010 will be a key indicator. typically the party in the white house does not do well in congressional elections, the people tend to favor balance to keep things in order, Bush was the first in a long time to buck that trend and actually pick up steas in 2002 and keep many 2004 when he was still pretty popular. If history goes back to norm, GOP should see pickups in 2010, I think we all see what one party control gets us...nothing but a mess!

I've posted an analysis of the Gallup PID numbers:

http://www.pollster.com/blogs/new_gallup_has_pid_tied_yep_it.php

And how nice to see a civil disagreement. Thank you Stillow and conspiracy! Would that others could follow your example-- we don't have to agree to be civil.

Charles

Not being a pro at this statistical stuff, I have a question. When you actually have an outlier, it is dismissed out of hand or can it still mean something? Even if it is an outlier, is this a ray of sunshine for the GOP or a reason for concern for Dems? Surely this is not good news for Dems, outlier or not.

Yep, I'm a straw clutching conservative.

So much for my typing. The question was is it dismissed out of hand (not it is!).

Bigmike - I'd say dismiss it unless there is confirmation from another source.

Stillow - Anything is possible. I'm sure the GOP will pick up some House seats but I think the Senate playing field is too stacked against them to make any gains. Though I suppose stranger things have happened.

I just wrote about this on DailyKos - the poll asks TWO questions. The first shows OVERWHELMING support for legal abortion. The second shows increasing self-identification as "Pro-Life." What that says to me, and the poll on my blog and some comments also reveal, is that pro-life is a personal stance that does not directly correlate to "anti-legal abortion." And that is the big news from this poll that the MSM is overlooking.

I am having trouble with this outlier concept. In the scientific world, if you get bad results there is typically a reason. Instrument problems, an programming error, or something.

So Gallup conducts a poll, presumably using generally accepted methods, and gets the wrong answer. Other than comparing to other polls which have potentially the same kinds of issues, is there any way of knowing just how wrong gallup is?

I guess my question would be is this as exact a science as it claims to be? If "outliers happen," just how good are the MOE values?

Big Mike-- Great question and an important point.

First, a single outlier in a sample doesn't mean the entire sample is equally "off" for all questions. It is easy to seize on a single item-- PID in this case-- and use that to discredit the entire poll. But most survey questions are only modestly correlated with each other so one outlier doesn't mean all the questions will also be outliers.

Second, it is certainly possible to provide objective analysis of the extent to which a poll is an outlier. See my post on this poll as an example of that. Of course once can easily grab at one problem and use it rhetorically to dismiss other results, but statistically that's not justified.

Third, you ask "is this as exact a science as it claims to be? If "outliers happen," just how good are the MOE values?" First, the MOE is meaningful in the same way quality control in engineering is meaningful. I know the properties of some product won't be exact, but I can specify the error tolerance with pretty good precision. The problem in surveys is that there are sources of error beyond that of sampling that are hard to account for with precision. It is like measuring the life of a lightbulb (something Quality Control does routinely and with good precision) but then using the bulbs with a wide variety of different voltages-- if you don't account for those differences in conditions, you manufacturing QC estimates of lifespan won't account for all the variation in actual bulb life. In surveys we know very well how to measure MOE from sampling but the effects of question wording and non-response (most centrally) we have to estimate empirically rather than from the well developed theory of sampling alone. So if you look at my post on this poll you'll see that the EMPIRICAL Confidence Intervals of the second chart are about +/- 5% rather than the +/-3% that the sample size alone would imply. This shows how much extra variation we empirically observe beyond what would be expected from sampling alone.

Finally, the thing about outliers is we find out pretty soon if they really mean anything. if this poll is really an outlier (as I think the evidence shows) then the next Gallup poll will be back inside the MOE in my figures and we'll quit worrying about it. If it really isn't an outlier, but an early reading of a change in public opinion, then we'll soon lean that too as other polls also start to show this uptick in Rep identification. But we can only say at this point that it is out of line with previous data. The future will either confirm that or show that it was signaling a change.

And as "a straw clutching conservative" you might take solace from this-- Rep ID is likely to be nearing a minimum (baring a stunning success by the administration). Therefore we might expect some Rep gains before too long. Midterms are almost always a good year for the out-party. This poll, outlier or not, doesn't really affect those two facts.

dmitcha--- good point. Abortion attitudes are very complex and conditional. It is easy to find seemingly contradictory views within a single survey, let alone across surveys. Generally, when asked specifically about making abortion illegal or reversing Roe, 70+% are pro-choice. When asked if abortion should be legal in ALL circumstances, only 20% or so agree with that (and only 20% or so say ILLEGAL in ALL circumstances.) Much hangs on how the in-between categories are described, or if they are described at all. Ambiguity of what "some or most circumstances" means makes it very hard to know what the public has in mind when it seems to support greater restrictions.

Charles

I took a deep breath when I started reading "stillo's" comments. I don't know the answer to any of this but it was a breath of fresh air to hear a debate that is not so octane and built on thoughful dialog. I found myslef actually thinking about the substance of what was being said instead of think here's another crazy from the right (although I am quickly finding out (there are just as many crazies on the left).

I just want to add my own kudos for the tone of the dialogue among the previous posters. It is a bit of fresh air, particularly when compared to the partisan echo-chamber effect that increasingly seems to permeate some of the other blogs, like at 538.

I think that party ID really only matters a few months prior to an election, and only if the party in the minority has a clear advantage (5+) over the party in the majority. Otherwise I think it's a wash.

I would wager that the Republicans would pick up a few net house seats, and the Senate might change by a seat.

I think the GOP actually picks up 3-5 in the Senate and 10+ in the House. Dems will retain majorities in both houses, but the GOP may pick up just enoughin the House to stop some of Pelosi's agenda when you count the bluedogs.
Keep in mind Obama will not be on the ballot...and usually the party not in power is mor eenrgzied to change things up. So typically the party not in power gets higher turn out.
A lot of this though will depend on how the ecommy is in late 2010...I still say the major concern for Obama and the Dems is a temporary recovery of the economy, but then an inflationary recession sets in due to the overspending. That would really hurt Obama and the Dems and you could very well see a conservative uprising like in 1994.
The problem for me as a fiscal conservtive is that I have seen both parties move to the left on spending. Bush moved the GOP to the left on spending and entitlements and what it did was force the Dems further to the left to present contrast. That is why I see not a lot of differences between the two parties, the GOP on fiscal matters has become a center left party and the Dems have become much more liberal on fiscal matter sthan they used to be. That also explains the internal struggle in the GOP right now also....the fiscal conservatives are tryign to take back the party after 8 years of Bush, but are being met with resistance from much more left leaning fiscal GOP.
But since most will vote there pocket books on election day, it will all depend o nthe economy in 2010....we knew the economywould recover no matter what since it works in cycles, the problem Obama has is what will the impact be from all his g'ment spending...will it truly be what he says, or it will do what I think it will do and take us right into a inflation based recession.
The margin of error is not very big for Obama, he is president based on 5 or 6 states he won by less than 5 points....FL, NC, IN, VA, etc, etc. Those states can very easily swap in 2012....
Both parties are drunk with spending, eventually that is going to catch up to all of us.....and its gonna hurt.

I agree with your House assessment Stillow, but could you tell me where the Republicans will pick up 3-5 seatsin the Senate? They may be lucky if they only lose 2 seats....

CO, CT, NY, DE and NV are very promising. The Dems should not think Reid is safe, he is actually fairly unpopular in NV, the GOP just needs a high profile challenger.
I think the GOP holds NH and OH...and with Crist running in FL, that seat is almost a lock for the GOP to keep. I could be wrong, that is just what I see.

How does this party ID poll mesh with this one, but the same organization?

http://www.gallup.com/poll/118528/GOP-Losses-Span-Nearly-Demographic-Groups.aspx

@Stillow:

Most of the high profile GOPers in NV either passed on the race or have ethics issues. Reid is unpopular, but he's not going anywhere.

NY and CO will stay Democratic, but I think there's some issue with DE and CT. But we'll have to see how this develops over the next few months.

Those are some pretty liberal states that you're assuming. The Republican brand does not sell well in those places. I think it would be jus as fair to say places like Kentucky will go Dem if NY will go Repub. CT and DE could be on the radar for the Repubs, but NY and NV at this point are tough to see switching, even CT could be a long shot. I could be wrong, but it's just as plausible to say the Dems will win Kentucky, Ohio and Missouri taking away any gains the Repubs make in CT, DE and CO.

I assume Dodd will be the nominee in CT, if he is then he will lose that race to a moderate republican.
DE, if Castle runs who is pretty popular, he will probably win in DE, espeically if Biden's son is his opponent.
CO is a tru toss up.
NY, if Pitaki runs who again is a very popular moderate republican, he woud probably win. Early polling shows he would do very well there.
NV, this is an outside shot in my view, if Heller runs he will carry northern Nevada heavily....and probably spit the Henderson area...it would depend on how much of Las Vegas he can peel away from Reid. Reid's approval rating in NV is below 40 the last I read, that is never good for an incumbant. Even a potential outsider like a wealthy unknown businessman could give him a serious challenge.
Then I put thos eearly favorable leaning GOP together with the fact hisotircally the opposing party does well in congressional elections....and the inherant desire for balance of power among the people....I just see GOP pickups. In 1992 no one presidcted a GOP uprising, but it happened....
Plus if Senator Byrd is forced to retire due to illness, that seat could also swap as WV has bcome quite conservative the past decade.

Could happen, but all the stars would have to allign. Of course you never know.

I guess it's different here in PA. Moderate Republicans as of now tend not to win the primaries. So if it would work in those other states I could see your point, but it depends on how conservative the base is.

@Stillow:

I just don't see Gillabrand loosing in NY. Pataki is somewhat okay on paper, but Gillabrand has a whole lotta cash and heavy institutional support.

As for NV, it's certainly possible that a self-funder could come through for the Republicans, but I just don't see anyone right now.

Also, the Governor of WV is a Democrat, so he'd likely appoint another Democrat, and with the advantages of incumbency, well, I just don't see an opening for the GOP.