Pollster.com

Articles and Analysis

 

Morning Status Update for Monday 10/27

Topics: Status Update

All of the speculation about polls this week will center around two: Are we seeing any meaningful trends in the margins by which Barack Obama leads John McCain and which poll is coming closest to measuring the level of Obama's lead?The question of accuracy of the size of Obama's margin is difficult for us resolve with certainty, but since pollsters generally keep their methodology consistent from poll to poll, our ability to check for trends is very strong. Trends are what I concentrate on in these morning summaries, and for the last two weeks we have seen little evidence of an erosion in Barack Obama's lead. If anything Obama's position today -- in both the national and state level surveys -- is slightly better today than it was in early October.

While the polls we logged yesterday and over the weekend are relatively few in number (as compared to the coming flood for the rest of this week), today's summary is no exception. The changes indicate mostly random fluctuation in the battleground states, although we do see changes in status in two states -- Georgia and Arizona -- that few would have considered "battlegrounds" in any sense until very recently.

Here is a list of new polls logged yesterday (see yesterday's update for the Saturday releases):

081027 new polls.png

The highlights:

Two new surveys showing a deadlocked race in Missouri by Mason-Dixon and Research2000 confirm that state's status as a toss-up of toss-ups.

A new Mason-Dixon survey in Georgia, showing McCain leading by just six percentage points (49% to 43%), helped nudge Georgia into our toss-up category, at least for the moment. Two recent surveys (Insider Advantage and Democracy Corps) have shown a margin of less than two points; four more last week including the latest from Mason-Dixon show McCain leading by 5 or 6 percentage points. The combination helps narrow McCain's lead on our trend estimate to 3.8 percentage points (49.1% to 45.3%), just enough (given the smaller than average sample sizes in Georgia) to shift the state to toss-up status, at least for the moment.

In Arizona, we saw two new surveys over the weekend, one sponsored by a Democratic aligned interest group and a second conducted by non-partisan research and public relations firms. Both included Ralph Nader and Bob Barr as choices and both yielded a combined 4-5% for these third party candidates. Perhaps as a result, John McCain's margins on these two polls were surprisingly narrow (2 and 4 points).

Since these are the only two new polls in Arizona in October, and since the "nose" of our trendlines tend to be more sensitive when recent polls are sparse, they help shift Arizona from "strong" to "lean" Republican.

081027 trends

Once again, the net impact on the battleground states with new polls shows the sort of random pattern consistent with no trend in either direction. With yesterday's new polls, the trend estimates shift slightly in Obama's direction in four states and in McCain's direction in three. And as noted in last night's update on the daily national trackers, their changes yesterday were also small and mostly random.

 

Comments
NdChristie:

Mark -

I've been reading these for a while now and, while they're quite comprehensive, I think what's called "no trend" can be a bit misleading, as it seems to suggest that all the recent data is negligible noise when it really makes a strong and obvious pattern : consistency. There's no good jargon for a steady graph, but there are plenty of implications from it that seem to be dismissed : if the GOP doesn't get busy quick, we won't even need to find a trend.

____________________

gil pittsburgh:

I agree, there has been remarkable consistency.

I think we'll see some actual tightening this week (as opposed to the imaginary Drudge kind) this week as those undecideds maybe make up their mind.

____________________

mj1:

Would not hardening of the race be a better term than tightening? The problem for McCain is that there are not enough undecideds left to change the dynamic here. Also, there is no guarantee that all undecideds are going to McCain. Many, if not most will go to Obama.

____________________

JimGray:

I also agree. Besides, this layout confirms most other electoral maps online- Slate, NYT, RCP, etc.
The polls in my home state, MI, lagged only a couple of days behind my predictions of when it would turn light blue and then solid blue. What comes as a surprise though is now apparent shift to navy blue.

____________________

Obamanuck:

It seems to me people are not seeing the forest for the trees. Looking at the forest, it looks more like a movement than a trend, and that movement has been: blue/Kerry states remaining blue and getting bluer, and red/Bush states going from red to pink to yellow, with some very surprising ones going all the way to light and then to dark blue.

There is not a single blue/Kerry state that is not still blue; all toss-ups are formerly red/Bush states, and several former red/Bush states are now some shade of blue. That we are seeing not only Virginia and North Carolina but also Montana, North Dakota and now perhaps South Dakota, plus Georgia and even Arizona slipping from dark red some other shade or colour is surely an indication of a "trend" -- McCain is in a hole, and it is trending deeper.

____________________

RS:

I wonder if Mark B.'s comment of "noise" was more in response to media stories (pushed, no doubt, by Zogby? and the media's own pollsters) that the race is tightening. Mark B.'s considered response, of course, is (I am paraphrasing here) "bull!" :-)

Basing trend-reads on results from an individual pollster for that state (e.g. R2K for MO and UNH for NH) is somewhat misleading, because the gap between polls can be a full month - many lifetimes in politics!

As for a previous poster who mentioned Slate - "Election Scorecard" is sourced from Pollster.com.

____________________

gil pittsburgh:

mj1:

Yeah, I think hardening might be a better term - and you're right there really isn't enough in that group to swing it to McCain - because he obviously won't get 100% of undecideds and there's probably 2-3% of those that will go Nader or Barr or whoever LaRouche is running this year.

____________________

NdChristie:

Actually, while I think the graphs are consistent and that mccain would need to hit hard and fast, i don't think the nation is obama country yet for certain.

what makes me nervous is that undecided is an honestly open distinction, something we can count on to not be counted on, while I can't help but feel that a lot of obama "supporters" will turn at the last minute when it comes time to vote for a minority, or just keep themselves home. I will be stunned if even some of the deep blue states don't see some tight races just from this withering. I may have a bleak outlook on racism in America, but I think we dems need more than the graphs are showing before I breathe easy.

____________________

wjbill49:

I know most are superstitious but doesn't seem that we are now playing for how big a mandate? Will it be 375EV's? Will McCain get to 200 EV's or 150?

____________________

Keith in KY:

I have several questions since everyone here seems to be very knowledgeable about the polling.

1. I have seen estimates that in some early voting states over 50% of the electorate is expected to have already voted by November 4th, an amount that could equal 30% or more of voters nationwide.

How is this being accounted for in interpreting the polls? What are the potential impacts? Do we know how early voters are the same or different from election day voters?

2. I have only heard of one pollster (don't remember which one) who is able to contact cell-phone only people. To what degree is everyone just ignoring this polling problem because we don't have a good solution to it? How could it impact interpretation of the polls?


3. Apparently a lot of exit polling is still being planned for this year. To account for early voting, apparently a lot of phone calls will be conducted to augment the actual exit polls.

What reason do we have to believe that the phone surveys will be as reliable as the actual exit polls? How is the cell-phone only bias being accounted for?

____________________

RS:

@Keith in KY:
1. Michael McDonald's site may be of interest:
elections.gmu.edu/early_vote_2008.html [lists early voting statistics, not a pollster, though.]
Here's PPP's take on early voting:
publicpolicypolling.blogspot.com/2008/10/early-voting-in-north-carolina.html

2. Here's Mark B.'s post on cell-phone voters:
www.pollster.com/blogs/more_cell_phone_data_from_gall.php

As for exit polls... Don't wait for them, GOTV :-)

____________________

Jerdog:

@wjbill

Not counting ones chickens is not being superstitious. Also please don't use the term "mandate." It was inappropriate for Bush to use it and will remain so.

@Keith on KY

Exit polling is a quick snapshot of peplle who have just voted. I dont see how phoning people can get this.

____________________

RWCOLE:

The tracking polls out this morning seem to be moving just a bit for McCain (except for hotline)---can't tell if there's anything behind it until we see the others but worth watching.

____________________

UpstateProgressive:

Mark Blumenthal may not quite spell it out, but his implication is clear: if the race is stable and consistent and there are no real trends either way, then that's very bad for McCain.

Obama has now established a margin more than sufficient for victory and perhaps sufficient for an electoral landslide. As others have said, the horserace isn't just rock steady, the trends have now hardened. There are very few voters who haven't made up their mind, and in fact, a large number have already cast their ballots, in states with early voting. If undecideds split their votes, which seems likely, Obama wins easily. If they trend toward Obama, it's a landslide. But since Obama is at or over 50% in most polls, McCain needs to win back some Obama voters to have a chance. and with most already firmly decided, that's going to be hard.

Team Obama has to guard against over-confidence, obvious, and turnout is going to remain very important, but the absence of a change in trend is a terrible trend for McCain with 8 days left. With the economy the central issue, and no likelihood of a turnaround in the markets in the next week, how is McCain going to find a way to come back?

____________________

Anandakos:

MJ1,

I'm an Obama supporter but I have to take strong exception to your assumption that most undecideds will go for the Senator.

Anyone who is undecided at this point is either being coy or is uninformed which implies pretty disinterested.

If you were uninformed and disinterested would you vote for a relatively young black man with a Muslim-sounding African name? I don't think so.

Prejudice runs deep among LIV's I expect. I'd give McCain 80% of the undecideds, and hope I'm right to give Obama even 20.

____________________

RWCOLE:

It appears that there is a strong tendancy for McCain to gain ground over the weekend and for Obama to gain during the week- has anyone examined this carefully?

____________________

KMartDad08:

Anadakos,

I keep hearing that the undecideds will break for McCain, and you offer an interesting rationale for that. However, to the extent that someone is going to vote against Obama for racial reasons, they were going to do that all along. I can't see that being the deciding factor at the end of the race. I can see people making up their minds based upon perceived "inexperience", but McCain's "experience" has been to toe the line on the core Republican philosophies that got us into this mess. With more bad economic news expected to break this week, I would think Obama will get a healthy share of undecideds.

____________________

bee91:

Given this data, what can we expect from the news channels Nov-4? Of course they will spin it to create suspense. Some suspense is built-in given the time zones, East Coast blue states, central red, etc. Is data available from 2004 regarding the rollout of results?

____________________

Michael Ott:

While I find the margin very interesting, I find the two numbers independently significant as well. I see how the margin is significant; the center of gravity shifts in each battleground state between +3 for Obama to +3 for McCain. But what about the top line numbers? I would be interested in seeing a chart of Obama's numbers in each state as a function of the margin. Or at least an indicator of the number of not-accounted-for voters in each poll.

____________________

I think there may be a "x" factor that pollsters have not discussed--as far as I can determine. For simplicity sake, I call it the "depressed Republican" factor. Here in Albuquerque on Saturday, around 1500 people attended the McCain rally and ~45,000 attended to Obama rally. This seems to suggest a statistically significant lack of enthusiasm by McCain voters. Can anyone tell me if the polls take into account the likelihood that polled voters may be too demoralized to actually vote?

____________________

Pedro,
Mishkin has an article about "The Obama Effect" and mentions "Effects" from past elections. I like how it is broken down into a "social effect" and an "enthusiasm effect". Obama's "enthusiasm effect" does seem significant. The post-election analysis will probably spend a lot of time debating how significant.
Other columns on pollster.com talk about how pollsters try to determine likely and unlikely voters.

____________________

Keith in KY:

@RS

Thanks, good information. I think I had read Mark B's original column. I guess I'm glad to see that the big polling org's ARE able to incorporate cell phone only households in their data for completeness.

Neither of these sources mentioned whether the polling numbers we're looking at include people who have already voted or if they are only looking at people who have yet to vote. Anybody know?

@jerdog That's what I'm thinking. It seems to me that exit polling has pretty much reached the end of its useful life. As the percentage of early voters continues to increase, it seems like the "exit poll" will need to morph more into a "tracking poll" where day by day the pollsters find people who voted that day and ask them the things they normally would have asked outside the polls.

____________________

Gary Kilbride:

KMartDad08, you make a valid point. I keep hearing Chuck Todd and others asserting McCain will win 70% of undecideds, and that both camps expect it. I would be surprised. That is asking a small portion of the electorate to vote dramatically against the overall trend of the race, against the candidate favorable/unfavorable percentages, and against the current approval numbers of the two parties. In 2004 I was posting that Kerry would not win the typical avalanche or undecideds to the challenger, and this time I doubt it will be one sided in either direction.

However, I believe there is a slight tightening of the race. If you look at Mark's table, the 5 national polls combine for +11 points red since the last sample, and +6 points blue. I noticed the same thing yesterday. I always focus on the national polls, believing the state results will naturally fall in line with the national margin. Per usual, the state polls are older and less frequent, two of the margins compared to September samples.

____________________

electiontime1974:

McCain vs. Obama: Which Candidate Can You Afford? How Could the Next President Affect Your Financial Plan?

Voyant:

http://www.planwithvoyant.com

I am loving this tool. I created a better plan than my financial planner gave me. I fired him. I’m done with paper financial plans.

____________________

Cash6115:

Michael Ott, I have similar thoughts, and hopefully some answers. From the state data, Obama is currently pulling in 282 electoral votes from states where his average is above 50%. The closest states in this are CO, NH, & NM where he has a 6.5-7.0% lead over McCain. The 282 does not include RI's 4 EVs because of a large undecided due to lack of polling. If we switch RI for those three closest states were at 268 EVs. Morale of the story is Obama has to lose people who have committed for McCain to have a chance, according to the polls.

Of the states where neither have a polling average greater that 50%, FL,GA,IN,MO,NV,NC,& OH have Undecided/Other totals of 4.4-6.4%. No states have lower and MT is the next closest at 7.9% then SD(9.7),AZ(13.2),ND(13.8), and RI(20.9). Here I think that the Mountain West States could be anywhere based on a lack of pushing respondents and AZ & RI are more than likely going the way we think they will, but suffer the same problem.

____________________

Michael Ott:

Cash6115 Thank you very much!

____________________



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