Showing posts with label Bob McDonnell. Show all posts
Showing posts with label Bob McDonnell. Show all posts

Friday, September 25, 2009

State of the Race: Virginia Governor (9/25/09)

[Click to Enlarge]

The story in the Virginia governors race currently is what will Bob McDonnell's campaign do to counteract the effect the former state attorney general's master's thesis has had on the dynamics of his contest with Democrat, Creigh Deeds. The answer this week has been to fight back, taking the thesis issue on directly, and attempt to raise the specter of tax and spend Democrats in Deeds' transportation agenda. And given the post-debate back-and-forth over taxes and the transportation plan, that is probably a wise move on the part of the McDonnell campaign.

Still, polling continues to show a tight race to become the Old Dominion's next governor. That was highlighted again by the release of a new Insider Advantage poll that showed McDonnell with a lead of a little less than four points. And as was the case in the Washington Post poll that came out earlier in the week, this one seems to be showing a shift of women into the Deeds' column. Deeds lead among women here by just under three points.

2009 Virginia Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Deeds
McDonnell
Undecided
Insider Advantage
Sept. 24, 2009
+/- 3.8%
602 likely voters
44.2
47.9
7.3

The thing to keep in mind, though, how much room McDonnell and Deeds have to grow their support. Deeds is getting nearly 90% of Democrats (about a third of the respondents) while McDonnell has a shade under 80% of Republicans on his side. The Republican is also winning among independents by over 20 points. If Deeds is to catch up he will likely have to keep pounding McDonnell on the thesis and making the case to independents that the Republican is too extreme. But 20 points is a lot to make up.

And we are just now entering a period where the tax issue may be settling in as well. There was a lag between when the thesis revelation took place and when it seemingly showed up in the polling. If the tax/transportation issue has gained any traction, that may manifest itself in next week's polling.

As for FHQ's averages of the race, the margin is now down to seven points and closing.

[Click to Enlarge]



Recent Posts:
State of the Race: New Jersey Governor (9/25/09)

2012 Presidential Trial Heats: PPP (Sept. '09)

PPP 2012 Presidential Poll: Huckabee Still Does Best, but All GOP Candidates Drop Off Against Obama

Sunday, September 20, 2009

State of the Race: Virginia Governor (9/20/09)

[Click to Enlarge]

FHQ is averse to taking one poll out of context, but the evidence is mounting in the polling of the Virginia governors race that the contest is tightening as September comes to a close. And in the clearest indication yet that Bob McDonnell's thesis is having an impact on the race, the Washington Post's new survey shows the gap not only closing, but that independent women -- the group of potentially undecided, or at least heretofore undecided, voters most likely to be affected by the thesis revelation -- are moving away from the Republican and helping Creigh Deeds draw closer. In last month's poll, the Post found McDonnell ahead by 28 points among that subsample. Now however, Deeds leads narrowly by three points; a dead heat given the margin of error in the poll (no to mention a larger margin among the subsample).

2009 Virginia Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Deeds
McDonnell
Undecided
Washington Post
Sept. 14-17, 2009
+/- 3%
1003 likely voters
47
51
2

And while Deeds is closing the margin between himself and McDonnell, it is important to note that McDonnell is still above the 50% mark in this poll and is only down three points from the Post's survey a month ago. There has been some damage inflicted, yes, but to this point it isn't the type of damage that will take down the Republican's campaign. As the Post points out, the enthusiasm remains higher on the Republican side of this race. Still, if the thesis continues to have the same dampening effect on the former state attorney general it could spell real trouble. However, FHQ contends that the magnitude of the thesis will decay over time. As more voters move from the undecided to decided column, the less chance there is for the thesis to prove a decisive role in their vote choice.

Does Deeds have some momentum? Yes, but Bob McDonnell is still very well positioned (above 50%) in this race.

[Click to Enlarge]

In our averages, the gap between Deeds and McDonnell is under eight points, and as such, Virginia (in the figure at the top) is now not as red as it has been recently.

...but it is still decidedly reddish at the moment.


Recent Posts:
About that New Jersey Governors Poll, Part II

Huckabee Takes 2009 Value Voters Straw Poll

Friday Afternoon Open Thread: The Americano

Friday, September 18, 2009

State of the Race: Virginia Governor (9/18/09)

[Click to Enlarge]

Another day, another poll in Virginia. And while the new Daily Kos/Research 2000 survey of likely voters in the Old Dominion does not show as close a race as yesterday's Rasmussen poll, it seems a rather accurate depiction of where the race is currently: McDonnell is around the 50% mark and Deeds support has rebounded some after a summer swoon followed his initial post-primary boost.

2009 Virginia Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Deeds
McDonnell
Undecided
Daily Kos/Research 2000
Sept. 14-16, 2009
+/- 4%
600 likely voters
43
50
7

And the underpinnings of the poll are consistent with that "feels accurate" idea in mind. In the survey, McDonnell is doing about ten points better among Republicans than Deeds is doing among Democrats (89-80, respectively), but Deeds still clings to a small advantage among women (and while that margin is not statistically significant, it is likely a better gauge of the true breakdown among women than other recent polls that have shown that sub-race all over the map. The bottom line is that there just isn't that much of a gender gap in this race.). Most importantly, though, McDonnell continues to best Deeds with independents. The margin in the race may have decreased, then, but McDonnell's position near 50% has not changed. The movement is all with Deeds at this point as the Democratic state senator has made a move into the mid-40s in the most recent two polls.

In FHQ's averages, though, Deeds still comes in below that point while McDonnell is just under 50%.

[Click to Enlarge]



Recent Posts:
State of the Race: Virginia Governor (9/17/09)

What if Obama Won the Electoral College 1265-599?*

GOP Temporary Delegate Selection Committee to Meet on Sept. 28

Thursday, September 17, 2009

State of the Race: Virginia Governor (9/17/09)

[Click to Enlarge]

It must be the thesis, right?

Maybe, but if the thesis is driving the margins in the polling of the Virginia gubernatorial race lower, we should expect that when a "not this optimistic" poll from the Daily Kos and Research 2000 comes out tomorrow (?) commentators to spin the less favorable results as a function of Deeds' tax comments* at today's debate. [Nevermind the fact that the poll was in and out of the field after those comments were made.] Yes, that's an unfair depiction of the media, but such is life when poll gazing.

2009 Virginia Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Deeds
McDonnell
Undecided
Rasmussen
Sept. 16, 2009
+/- 4.5%
500 likely voters
46
48
5

The undeniable fact is that once the idea of the thesis and what it meant was internalized, the poll numbers began to dip for Bob McDonnell. But let's not lose sight of the fact that the Republican is still in a good position in the Rasmussen poll that was released today. The candidate who should be getting the focus is Creigh Deeds. The Democratic state senator is at his highest point in terms of polling since he led McDonnell in a poll conducted (by Rasmussen) the day after Deeds' victory in the Democratic primary. Taken on its face, then, this result is something of an outlier compared to the recent polling the race.

Are things closer than they were pre-thesis? Yes.

Are they within the margin of error close? That's debatable.

What's clear is that the thesis seems to have closed the gap to some degree. But does Deeds have the momentum? We'll have to see. Internally, I mocked the idea of a Thursday debate, but it was well timed if there happened to be a gaffe of some sort for either candidate (but more so for Deeds since he didn't have a thesis-type revelation already out there). Fridays are typically days to bury some bit of news before the weekend.

The race is closer, but the fundamentals of the race still favor McDonnell overall. The Republican is inching closer to the 50% mark even as Deeds is closing the gap. And that is something not to lose sight of as we head down the home stretch in this one.
[Click to Enlarge]

Surely the good folks at Rasmussen don't follow little ol' FHQ, but it sure sounds like they are being a wee bit defensive/opportunistic with their comment in the write up of the poll above.
"All of those figures include “leaners.” Leaners are those who initially indicate no preference for either of the candidates but answer a follow-up question and say they are leaning towards a particular candidate. Premium Members can review the data without leaners and complete demographic crosstabs."
In other words, complain if you wish, but if you want the data, pay up. Duly noted.

...said the employed political scientist still trying to shake the cheap graduate student mindset.

FHQ will say this: I'm glad to see Rasmussen come out with some "information" on how they collect their leaners data. [Hint, hint; nudge, nudge.]

*You absolutely have to love the title to that National Review blog entry.

Recent Posts:
What if Obama Won the Electoral College 1265-599?*

GOP Temporary Delegate Selection Committee to Meet on Sept. 28

Early Voting in New Jersey and Virginia?

Tuesday, September 15, 2009

State of the Race: Virginia Governor (9/15/09)

[Click to Enlarge]

Was it the thesis?

That will be the question asked all day Wednesday when this new Clarus Research Group poll sees the light of day in the morning. Regardless of whether it was Bob McDonnell's thesis or something else, the margin between the former Virginia Attorney General and Democratic state senator Creigh Deeds is as narrow as it has been in any poll since mid-July.

2009 Virginia Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Deeds
McDonnell
Undecided
Clarus Research Group
Sept. 10-14, 2009
+/- 4%
600 registered voters
37
42
20

Notably Deeds has pulled even with McDonnell among women (a figure that has been all over the place in recent polling in the Old Dominion) and has a slight edge (five points) among voters in vote-rich Northern Virginia. Solid numbers for the Democrat.

However, there are some aspects of this poll that stand out.
1) 20% undecided? In September? That is a very high number this late in the race. Now, it could just simply be that some respondents in this sample are reconsidering their position given the news in the race of late, but are hesitant to shift over to Deeds completely. There has not been a total of undecideds that high since the Daily Kos/Research 2000 poll a week before the Democratic primary; in other words, in a time when there was uncertainty surrounding Deeds' viability or even candidacy.

2) Registered voters? Again, at this late stage, most polling firms have switched to a likely voters only model (and others are even including leaners. cough! Rasmussen. cough!). It is curious, then, that Clarus has stuck with the registered sample and not the narrower likely voter sample. We saw just yesterday in the Monmouth poll in New Jersey how large a discrepancy can exist between the two samples. Chris Christie led by eight points among likely voters, yet Jon Corzine edged the former US attorney out among registered voters by a point (something the DNC wanted to point out today.). That's a nine point swing. A similar nine point swing in Virginia would put McDonnell up 14 (Not that that would be the case here. I'm just trying to illustrate what we're actually looking at in this case.) and that would be in line with the Survey USA poll that was released the week after the thesis revelation occurred.

Of course, FHQ did state last week, that we would have to give it a week or so to see how much impact the bombardment of thesis stories would have on the race. That much time is now behind us, and the race looks closer.

...with some caveats.

As for FHQ's graduated weighted average of the race, McDonnell has been pulled back under the 50% mark but continues to hold a sizable advantage over Deeds. Is the momentum on Deeds' side now, though? Maybe. But recall that there were a series of polls that had Corzine within a handful of points of Christie about two weeks ago. Those polls now seem a part of the distant past, however.

[Click to Enlarge]



Recent Posts:
State of the Race: New Jersey Governor (9/15/09)

"Wait till her fat keister is sitting at this desk"?

State of the Race: New Jersey Governor (9/14/09)

Saturday, September 5, 2009

State of the Race: Virginia Governor (9/5/09)

[Click to Enlarge]

How about now?

Have the ideas proffered by Bob McDonnell in his thesis and covered widely this week sunk in among likely voters? More importantly, has it had any impact? The Rasmussen poll that was in the field on the Monday after the news broke showed very little, if any movement away from McDonnell or toward Deeds since the previous survey of the state from the polling firm. The same is true of the Survey USA poll that was released on Friday night (before a holiday weekend!?!).

2009 Virginia Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Deeds
McDonnell
Undecided
Survey USA
Sept. 1-3, 2009
+/- 4%
611 likely voters
42
54
4

Since the firm's last poll in the race (in late July), there has essentially been no movement. Sure, it appears as if there has been a slight shift toward Deeds -- the gap has closed in the interim period from 15 to 12 points -- but that movement is within the margin of error.

So the thesis has had no impact?

Well, it looks that way doesn't it? But let's look a bit more closely at the toplines of this poll. What segment of the likely voting population would be expected to be most affected by the news of the thesis. Other than Democrats, the most obvious answer is women (Yes, there is a fairly broad overlap between the two.). The underlying "a woman's place is in the home" theme that peppers the thesis would hypothetically motivate women (on average) to want to vote against Bob McDonnell or say they would.

But what do we see in the latest Survey USA poll? First of all, women in this sample actually prefer McDonnell to Deeds by a margin of seven points. And secondly, that has increased since the firm's last, pre-thesis poll in July. Now, this is a finding that runs counterintuitively to what we know of electoral politics. Typically, women side with the Democratic candidate. That varies from race to race, but more often than not that gender gap is present to some extent. Before we dismiss this poll -- something FHQ wouldn't do anyway -- let's look at some other recent polls and how the samples broke along gender lines.

This isn't an exhaustive examination, but let's look at this trend in the polls conducted during and since August. The polls for which we have data in that period are the Public Policy Polling survey conducted last weekend and the Daily Kos/Research 2000 poll done earlier in August. The Washington Post poll, though it released a ton of data, did not include the gender breakdown and the two Rasmussen polls* are excluded. For good measure, let's throw in that July Survey USA poll and the earlier August poll from PPP.
PPP (early Aug.):
McDonnell: 50
Deeds: 38

SUSA (July):
McDonnell: 49
Deeds: 44

Kos/R2K (Aug.):
McDonnell: 45
Deeds: 46

PPP (Aug.):
McDonnell: 40
Deeds: 49

SUSA (Sept.):
McDonnell: 52
Deeds: 45
Honestly, those results are all over the place; especially those PPP numbers. They represent the two extremes, +12 McDonnell (in the early August poll that showed McCain voters outnumbering Obama voters by 11 points in the sample) to +9 Deeds in their latest poll. Granted across all of these polls, we're talking about subsamples, but on gender; not on something like liberal Republicans who voted for Obama, for example. It's a healthy subsample in other words. If we simply average the spreads in these five polls (not accounting for any kind of decay in the polls' weights over time) we find McDonnell ahead by about 3 points among women. At least that gives us some baseline for comparison. It provides us with enough information to say, "Well, it looks like this recent Survey USA poll shows a but more support for the Republican than average."

The bottom line is that this is still a somewhat striking result given what we know of the gender gap. However, at least we have some context now.

Gender wasn't what Survey USA thought was interesting in this particular poll. To the firm it was about the 2008 vote choice of the respondents in the sample. "Of those who voted for Barack Obama in 2008 and who are judged by SurveyUSA to be likely to vote in November 2009, 13% cross-over for McDonnell, twice the number of McCain voters who cross-over for Deeds." Yeah, that catches your eye, too. I'd really like to see those cross-overs isolated (talk about a subsample) just so we can see what the characteristics of a typical Obama-McDonnell or McCain-Deeds respondent are.

That may be asking too much.


[Click to Enlarge]

For the moment, though, it looks as if McDonnell has yet to be affected by the thesis revelation. It still feels too early to me for there to have been an effect. [This coming week's polls will be indicative.] That hunch is conditioned by the presence of the story to some extent. If it continues to get play in the media, that obviously increases the chances that it will potentially have an effect (exposure theory). Of course, the more polls released that show no impact, the better for McDonnell. That will serve to shift the narrative away from his previous writings.

*Yeah, I'm still too cheap to pony up for the Rasmussen subscription that would provide me with a more in-depth series of crosstabs for these polls. If you're reading this, have a Rasmussen subscription and would like to share that gender information with our readers, just drop a comment in the comments section below.


Recent Posts:
The 2012 Presidential Candidates on Twitter (Aug. 2009)

State of the Race: Virginia Governor (9/2/09)

State of the Race: New Jersey Governor (9/1/09)

Wednesday, September 2, 2009

State of the Race: Virginia Governor (9/2/09)

[Click to Enlarge]

It is too early.

Anxious observers of both partisan stripes are waiting for the inevitable (or perhaps not so inevitable) negative impact Bob McDonnell's Regent University thesis will have on the Virginia Republican's fortunes in the 2009 gubernatorial race in the Old Dominion. A day after the beginning of the political week began, when this story broke, is just too early for this to have sunk in and have had any measurable influence. Does that mean McDonnell is headed for automatic defeat in this contest? No. However, this does represent an opening for the Deeds campaign. It affords the Deeds campaign the opportunity to energize their base and close the enthusiasm gap in a way that to this point has been lacking (not from lack of trying). As I said recently, this isn't the type of revelation that is necessary going to hurt McDonnell among conservatives, but it does potentially put some independents in play.

Is it to that point yet? Will it get to that point? Maybe. Maybe not. But it will take some time. Polls next week will be a good starting barometer of the state of things post-thesis.

2009 Virginia Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Deeds
McDonnell
Undecided
Rasmussen
Sept. 1, 2009
+/- 4.5%
500 likely voters
39
49
9

In the meantime, Rasmussen released a new poll in the Virginia race today that did not really show that much movement from a month ago. In fact, Deeds dropped by a handful of percentage points among the sans-leaners version of the survey results while McDonnell stood pat. [Again, it is FHQ's policy to take that version -- the one where undecideds are placed in the camp of one candidate or another -- over the other.] Is this status quo result a telltale sign that McDonnell is safe on this? No, I don't think so (for the reasons I've mentioned already).

That said, the numbers didn't make that much of an impact on FHQ's graduated weighted average. Yes, Deeds fell a bit and McDonnell held steady versus yesterday's update, but we are not talking about some fundamental shift here. This was a status quo sustaining survey.

The bottom line? McDonnell leads, but an unknown has been introduced into the equation; one that could have an effect on the former Virginia attorney general's numbers.

[Click to Enlarge]


Recent Posts:
State of the Race: New Jersey Governor (9/1/09)

State of the Race: Virginia Governor (9/1/09)

2012 Presidential Trial Heats (Clarus Research Group): Obama vs. Gingrich/Huckabee/Palin/Romney

Tuesday, September 1, 2009

State of the Race: Virginia Governor (9/1/09)


[Click to Enlarge]

Let's take a quick and dirty look at the Virginia governors race, where today Public Policy Polling released a new poll. Again, this race is in wait-and-see mode at the moment. While the PPP poll was in the field for part of the post-thesis revelation period over the weekend, the full backlash/reaction to it could not have been fully incorporated yet.

2009 Virginia Gubernatorial Race Polling
Poll
Date
Margin of Error
Sample
Deeds
McDonnell
Undecided
Public Policy Polling [pdf]
Aug. 28-31, 2009
+/- 4%
596 likely voters
42
49
9

Even still, there are some interesting numbers to take note of in this poll. For starters, McDonnell's 14 point edge in the PPP poll a month ago has been halved and coincidentally, that matches the decrease in McCain voters in the sample from July to August. So not only is McDonnell's lead down to seven (actually Deeds gained more than McDonnell lost), but the composition of this month's sample is down from +11 for McCain in July to +4 in August. And recall, this is a state Obama turned blue in 2008, winning by seven points. The extent to which McCain voters dominate these polls then really drives home the presence of an enthusiasm gap in this race. Republicans have the edge there, but as I stated yesterday, this thesis matter has the potential to help close that gap by activating Democrats and moving independents (who by the way are favoring McDonnell by a 2:1 margin in this poll) into, if not the Deeds column, then into the undecided column.

For the time being, though, this race is in neutral at least according to FHQ's graduated weighted average. The poll did little to move the needle and in fact pulled the average into pretty close alignment with it. The next polls will tell the tale on where this race will be going for the next little bit or the rest of the race.

[Click to Enlarge]


Recent Posts:
2012 Presidential Trial Heats (Clarus Research Group): Obama vs. Gingrich/Huckabee/Palin/Romney

Have the Races Changed in New Jersey and Virginia?

2012 GOP Presidential Primary Poll (Clarus Research Group): Romney Jumps

Monday, August 31, 2009

Have the Races Changed in New Jersey and Virginia?

[They haven't in the polls with few exceptions.]

As we get ready to usher in the fall campaign once Labor Day passes next week, both gubernatorial races this year are at something of a crossroads. In both cases, the Democrat is trailing and in both cases, the Republican has been faced with some negative news recently. In New Jersey, Chris Christie is weathering the storm of links to the Bush administration as well as a series of personal issues. [Well, that makes it sound like a drug problem or something of that ilk. Speeding tickets and loans to subordinates aren't drug problems, but they don't come free of any negatives.] Further south, in Virginia, Republican Bob McDonnell is being haunted by his own words from twenty years ago, in the form of a Regent University master's thesis.

My first reaction to the McDonnell news was that if Democrats wanted to use the thesis effectively, they would have to be selective with it and not bombard people on the chance that it could trigger a backlash. After all, the Virginia electorate that has been picked up in polls since the June primary has tilted to the Republican end of the spectrum. This news isn't necessarily going to change their minds; it might energize them further. The power of this story, though, is in its narrative capacity. Democrats and the Deeds campaign have been pushing the "this isn't the real McDonnell" card since the spring, but they now have something to hang that on and refer back to ad infinitum. Anytime McDonnell says something that in anyway links back to the thesis, Democrats will pull it into the "that's all part of his blueprint" narrative. And while there is the chance that that also risks a backlash, it is a gamble that could also play well with apathetic Democrats and independents, who have to this point remained on the sidelines in this race*. But to say that that isn't a tightrope act for Democrats and Deeds is a mistake. The degree to which Deeds can find that balance between effectively playing this (and making trouble for McDonnell) or seeming desperate will determine the direction of this race down the home stretch.

[That said, I still want an answer to Jonathan Martin's tweet yesterday (I'm paraphrasing): If the thesis is so damaging, how did oppo researchers miss it in the 2005 attorney general race or any other office McDonnell has run for?]

In New Jersey, the narrative has progressed to its second step (beyond the Bush connection). Instead of the debate being on moderate/not moderate turf, as in Virginia, this one seems to revolve around Christie's judgment (the bent of the news does lately, anyway). So, while Christie maintains a lead in the polls, you can certainly see that the stories have shifted in this race from "throw the bum out"/anti-incumbent stories to "does Christie have the requisite judgment to be governor?" stories. And that isn't a good change in the prevailing winds for the challenger.

Does that mean trouble for the Republicans currently leading the polls in both states? Possibly, but it could end up costing the Democrats if their actions are perceived as desperate. The latter is more likely in Virginia simply because of the underlying partisan composition of New Jersey. Corzine just simply has a cushion (even while trailing) that Deeds does not in Virginia.

A couple of other caveats:

1) Timing. Is this too early for a closing narrative to emerge? If either becomes established, will it be stale by time the waning days of October roll around? The press may be looking for something new by then.

2) Decision-making: I need to check on this in the gubernatorial context, but it is true that in presidential races, most likely voters have made up their minds by around Labor Day. James Campbell would also contend (and has shown) that the polling around that time of the race is also the best indicator of the November election results. I'll have to check on both, but I thought I'd throw both ideas out there. Obviously the presidential race is a bit more high-profile, which hypothetically would mean that voters may wait a little longer in a gubernatorial race.

In any event, these items merit tracking over the next few weeks.

*The Public Policy Polling survey due out of Virginia purportedly shows heightened interest from Democrats versus a month ago (and this poll was in the field before the thesis news broke for the most part). However, McDonnell still has a 2:1 advantage among independents (again, pre-thesis).


Recent Posts:
2012 GOP Presidential Primary Poll (Clarus Research Group): Romney Jumps

Defense Authorization Bill Amendment Could Affect Primary Timing

Nevermind: Democratic Change Commission Meeting Postponed

Tuesday, August 25, 2009

New Jersey and Virginia: A Diagnostic Comparison of the State of the Race(s)

Part of the reason FHQ wanted to examine the New Jersey and Virginia gubernatorial races this year was not the races themselves as much as what they represent: an opportunity to test and try out a few things in terms of how we process new polling information as it comes in. To this point, though, we have essentially leaned on the graduated weighted averaging formula used to a fair amount of success in the presidential race last year. There's nothing wrong with that formula. It was far more simplistic than some of the alternatives out there and only missed North Carolina and Indiana in the electoral college categorization (Even then, North Carolina was essentially a tie in the average. But I digress...). The governors races in New Jersey and Virginia, then, are being utilized with an eye toward 2012 and the electoral college.

First of all, one feature that would have been nice last year (for every state or at the very least the swing states) is a graph similar to the ones FHQ has appended to each New Jersey and Virginia polling update. As I said recently, though, the lines on those charts seem to be floating in space without some baseline for comparison (the actual raw polling data, for instance). But that got me to thinking: The graduated weighted average is constructed to give the most recent poll the most weight, but also to incorporate past polling data in a way that guards against a shock to the system; an anomaly. And all that really is is a poor man's regression line. My question then was, How does the graduated weighted average stack up against a simple regression projection based on the polling data we have at our disposal? Sure, I could take an "everything but the kitchen sink" approach and add seemingly relevant variables to my heart's delight, but let's see how a simple bivariate regression as a start. Remember, Virginia and New Jersey are test cases for the 2012 electoral college model.

So all I did was regress the time in the campaign so far (measured as the number of days in the campaign*) on each candidate's share of support in the polls conducted over that period. All that basically does is provide us with a trendline based on the hypothesis that over time there will be some changes to a candidate's level of support. Yes, that is ambiguous, so let me be a bit more specific. Most clearly, we can hypothesize that over time, the undecided share will decrease and in this particular instance, that the Republican share will increase. Indeed, in both cases, the time component explained a surprising amount of the variation in the undecided share across polls as well as both Chris Christie's and Bob McDonnell's support (between 30 and 50%).

But the two Democratic models performed far worse. In both cases, less than 10% of the changes in Deeds' and Corzine's shares were accounted for in the time series. Why? Well, in neither case is there much change to speak of. There's more change in the Deeds case than for Corzine, but not by much.

Fine, what does any of this mean and what does it leave us with graphically? Good questions. I'll take the second one first and then use the two figures below to illustrate the former. Graphically, as you can see below it leaves us with a bit of a mess. Nine separate lines are a lot to take in. However, there is a wealth of information in these two figures. The most volatile lines are the raw polling datum (referred to there as actual) while the smoother two lines around with they hover (and are based upon) are the graduated weighted average (average) and the regression projection (predicted).

The raw data are nice, but let's focus on the other two lines, as this post is supposed to be about comparing two different projections of the state of each of these races.

[Click to Enlarge]

In the New Jersey example, we see that the graduated weighted average and the regression line track each other almost exactly in the case of Jon Corzine. Again, there isn't too terribly much change, relatively speaking, in the Corzine numbers and that keeps the lines closer together. Where there is more volatility, there is more divergence between the two measures. This is most clear among the undecideds. The graduated weighted average projects the level of voters yet to be had by either campaign at a consistently higher level in New Jersey since June than does the regression measure of the same concept. The same sort of phenomenon can be seen in Chris Christie's numbers. However, in this case the graduated weighted average comes in below where the regression finds the Republican candidate's support across these polls. And on the whole, the difference between the two measures appears to be growing over time. If the regression prediction is the more reliable measure (and that is an arguable point given the simplicity of the model), then the graduated weighted average is losing predictive power over time in regards to Chris Christie's share of support in this race. That isn't really the best trajectory to be on if you are attempting to use polling information as a means of forecasting the results of an election.

But New Jersey is just one case. How do things look further south in the Old Dominion?

[Click to Enlarge]

Things are a bit more muddled in the Virginia example and that is largely attributable to the differences across the two races.

First of all, there are far fewer polls that have been conducted in the Virginia race. Still, given the window of time that is being considered in each race, each state is averaging a poll every seven or eight days. Regardless, fewer polls overall in the Virginia case translates to more volatility in the graduated weighted average.

But also, there have been different dynamics at work in both races. In New Jersey, Jon Corzine has been stuck in a holding pattern in the polls while Chris Christie has, on the whole, gained over time. The Virginia case is quite different. The polls showed a close race early, but over time that has yielded a seemingly comfortable McDonnell lead.

The smaller window of time in Virginia means that there is less time for past polling results to have decayed and less new polls to have outweighed them. As a consequence, the graduated weighted average is stuck to some degree; overvaluing some of those past results that were more Deeds-heavy. Well wait, what that really means is that this graduated weighted averaging methodology is bias against Republicans. It happens to be in this case. But what the average is really biased against is rapid change. And in 2009, both Republican candidates are the ones who are moving in the polls, at least as compared to their Democratic counterparts. Which brings us to the crux of the matter: The issue with the average has always been whether the past polls are over or undervalued. In this comparison, it seems as if those past polls are still being overvalued, potentially at the expense of gleaning the true state of each race.

But let's return to those Virginia results for a moment. With the above caveats in mind, we would expect the average to underperform the regression in the McDonnell model while the two lines would remain rather close to each other (while slightly overperforming) for Deeds. All that means is that the status quo from poll to poll is protected more in the case of the average than with the progression of the regression trendline.

What does this mean for the graduated weighted average? I'm not apt to scrap it just yet. This exercise is helpful in determining the usefulness of the measure in settings other than the electoral college (and even for the electoral college, truth be told). Again, FHQ even examining these races in the first place is a function of tweaking the measure with 2012 on the distant horizon. Better to do it while that is distant and not on top of us with or after the midterms next year.

As for what this means for what you'll see in subsequent iterations of these polling updates, you'll continue to see the Actual vs. Average trend and will likely see occasional (and perhaps more advanced) regression model predictions. So, be on the lookout for that.

*For New Jersey, that means the number of days since the first of the year (as Christie was the clear cut Republican frontrunner to challenge Corzine) and in Virginia, the time since the Washington Post's endorsement of Creigh Deeds in the Democratic primary race (It was at that point that Deeds was really first seen as a legitimate candidate in the race -- primary or general election.).


Recent Posts:
Don't Forget Your Change Commission Reform Suggestions: Deadline Today

2012 Presidential Race: August PPP Trial Heats In-Depth

PPP Poll: 2012 Trial Heats (Obama v. Gingrich/Huckabee/Palin/Romney) August Edition

Wednesday, August 19, 2009

State of the Race: Virginia Governor (8/19/09)

As was the case with the New Jersey race for governor, FHQ has fallen behind in Virginia as well. Let's see what we missed (and what impact it has had):

Virginia Gubernatorial Polls (July 31-Aug. 18)
Poll
Date
Deeds
McDonnell
Undecided
Washington Post
Aug. 11-14, 2009
39
54
7
Rasmussen
Aug. 5-9, 2009
41
49
7
Daily Kos/Research 2000
Aug. 3-5, 2009
43
51
6
Public Policy Polling [pdf]
Aug. 4, 2009
37
51
12

FHQ had put on its Missouri cap at the end of July with the release of the Survey USA poll showing Republican Bob McDonnell cresting above the 50% mark and Creigh Deeds double digits behind. Well, in the time since that poll -- time FHQ was otherwise indisposed -- our "Show Me" attitude has turned a 180 and arrived at "Oh, I see." All four polls since have shown McDonnell around 50% and Deeds further back in the upper 30s to lower 40s range.

[Click to Enlarge]

Together, these polls, in addition to the earlier ones FHQ is considering in its graduated weighted averages of this race, put McDonnell ahead by right at ten points at the moment. And while the gap in the New Jersey race might be ever so slightly contracting, the race in the Old Dominion is headed in the other direction entirely. If you're a Democrat from Virginia or not, you have to be wondering where the Bush administration's loyalty rating of Bob McDonnell is (or at least a Virginia version of what may be a game changer in New Jersey*).

I read the word balance used in connection with this race recently (I'll have to track down the link.) and it is being used in a way I don't know that I've ever encountered. The idea proposed was that Virginians might be opting for McDonnell to balance out the recent Democratic shift in the state at the national level. Typically, we talk of balance in terms of ticket balancing; choosing an executive (on either the federal or state level) from one party and a legislative representative from the other as a means putting a check on the power of government. But this is balancing based not on an interbranch basis but in terms of levels within the federalist system. Nationally, then, Virginia has shifted toward the left; electing a pair of Democratic senators since 2006 and Barack Obama as president in 2008. In a state where Republicans control the House and Democrats control the Senate by the narrowest of margins, Virginians could opt to send McDonnell into office with a newly GOP-controlled Senate, pitting the GOP state government against the Democratic folks sent by the state to act in the federal government. This is a new idea to me. One I'll have to look at a bit more closely (or at least search for some research within the political science literature). Certainly, this is an altogether different type of sophisticated voting.

[Click to Enlarge]

The trendline at least is more interesting to look at than the New Jersey version. Instead of a rather static picture, the Virginia race offers a distinct trend in McDonnell's direction across the polls conducted since the Washington Post's primary endorsement of Creigh Deeds. Deeds' dilemma is trying to figure out a way to reverse the current trajectory of the race. With Labor Day on the horizon, it is getting down to crunch time.

*It remains to be seen whether the Bush/Christie link will inflict any sustained damage on the Christie campaign, but it is out there as an issue now. There will need to be more polls conducted to show whether that revelation is having any real long term influence on the race. The "Show Me" attitude is now shifting northward to the Garden state.

NOTE: For once, it is nice to have said that something is coming tomorrow and be able to put it out early. Too often, I end up saying something is coming tomorrow only to have it come out the day after that.


Recent Posts:
State of the Race: New Jersey Governor (8/19/09)

Marist 2012 Presidential Poll: Palin Lags Well Behind Obama but Holds Her Own in the GOP Primary Race

Which Republican is the Biggest Threat in 2012?

Friday, July 31, 2009

State of the Race: Virginia Governor (7/31/09)

[Click to Enlarge]

And now the race in the Old Dominion.

Survey USA weighed in on the race for the first time since about a week before the June 9 Democratic primary that handed Creigh Deeds the gubernatorial nomination. While the results were solid for the Democrat (against Republican Bob McDonnell) in that poll, they are not in this recent poll. Like New Jersey, this is another poll this week that shows both Republicans in better positions than both their Democratic counterparts. Let's look at the toplines and then I'll comment:

Bob McDonnell: 55%
Creigh Deeds: 40%
Undecided: 5%

Margin of error: +/- 4.3 points
Sample: 526 likely voters
Conducted: July 27-28, 2009

For starters, there's no doubt that McDonnell has a lead in this race, but this result is well outside of the margins we've seen in any other poll (other than the Daily Kos poll around the same time as the first Survey USA poll) in the time since Deeds got the Washington Post's endorsement in the Democratic primary race in May (when Deeds became viable as a general election candidate). This is definitely a Republican-laden sample. I don't know that the 38/32 Republican to Democratic split in the sample is all that uncharacteristic of Virginia. It is probable that such a sample could be drawn there, but it still feels a bit out of the ordinary. My point is driven home by the "who'd you vote for -- McCain or Obama" question. The split there was 52/43 in favor of McCain -- in a state Obama won 53-46 last November.

You may also notice that McDonnell has jumped out to a fairly sizable lead in FHQ's weighted average of the race. Meanwhile, Chris Christie has maintained about a ten point advantage over Jon Corzine in the New Jersey race despite the fact that Christie has been above that margin in quite a number of polls recently. The difference between races is the number of polls conducted. There have been far more polls in New Jersey than in Virginia and that translates into more volatility in the Virginia numbers. The Virginia race, then, is more vulnerable to outliers like the one this Survey USA poll represents.

Regardless, in these two races, the polls this week have had a Republican flavor with both candidates stretching their leads over their Democratic competitors.

[Click to Enlarge]



Recent Posts:
State of the Race: New Jersey (7/31/09)

Last Athens Post

FOX Poll: 2012 GOP Primary--The Romney/Huckabee Dead Heat Continues

Wednesday, July 15, 2009

State of the Race: Virginia Governor (7/15/09)

[Click to Enlarge]

Rasmussen released the results from a new survey of the Virginia gubernatorial race earlier today and unlike the New Jersey case did not make a distinction between leaners and non-leaners. Oh well. What the poll did indicate was the Bob McDonnell continues to hold a small lead (44-41) over Creigh Deeds in the race for the top statewide slot in the Old Dominion. Oddly enough, that brings FHQ's graduated weighted average margin in the race to right at three points.

The noticeable thing about the polling conducted since the June 9 primary is that Rasmussen has tended to give Deeds a higher share of support relative to the other handful of polls released. In fact, Rasmussen's poll the day after the primary is the only poll (other than the Anzalone Lizst poll that FHQ is not incorporating into its averages) to show Deeds in the lead. This new poll, however, seems to be in line with the other recent polls that been conducted. As Christian Heinze at GOP12 pointed out today, though, Deeds has the higher "partisan ceiling" simply because he's pulling in fewer Democrats than McDonnell is Republicans at this point.

Deeds had a good June in terms of fundraising thanks to some assistance from the DGA and some labor organizations, but trails McDonnell overall in cash-on-hand. The extent to which the Democrat can unite blue Virginians and catch up in the money game will go a long way toward pulling him even with or pushing him past Bob McDonnell as the summer days fade into to fall. But we're not quite to August yet; still a ways to go.

[Click to Enlarge]


Recent Posts:
North Carolina in 2012: Obama - 49, Palin - 42

The Paths of Presidential Primary Frontloading

State of the Race: New Jersey (7/14/09)

Wednesday, July 8, 2009

State of the Race: Virginia Governor (7/8/09)

[Click to Enlarge]

The Deeds' post-primary victory boost seems to have peaked. After handily defeating his Democratic primary opponents, Virginia state senator, Creigh Deeds, jumped past his Republican counterpart, Bob McDonnell, in the polling of the Virginia gubernatorial race, but has remained virtually stationary in the (scant) polling conducted since that immediate, after-primary period. In fact, other than the Rasmussen poll that was in the field the day after the primary, Deeds has been at either 43 or 44 points in every poll since the May 22 Washington Post endorsement (other than the Daily Kos poll that was in the field just prior to, during and after the June 2 primary). [As you can see in the graph below, McDonnell has ever so slightly stretched his lead from the last update and has pulled ahead since Deeds surged to the lead following his win in the primary.]

The real story, then, seems to be McDonnell's approaching that 50% mark FHQ has been discussing in the New Jersey race. The new Public Policy Polling poll out for the race has McDonnell ahead 49-43 and that keeps Deeds just within the 4 point margin of error for the survey. Sure, this is just one poll, but McDonnell's ability to stay at or around that mark, puts the onus on Deeds and the DNC to turn this race around. Regardless, this poll doesn't particularly change the long-term outlook. The race still "feels" like a toss up with a slight lean toward McDonnell at this point. And in that case, turnout becomes highly important. If Democrats can remain as enthused as they were throughout 2008, that likely helps, but if not, this becomes McDonnell's and the Republican Party's race to lose.

In any event, there is quite a bit of time between now and November. This one should continue to be a tight one.

[Click to Enlarge]



Recent Posts:
2012 GOP Primary Polling (June 2009 -- Rasmussen)

And Another Thing About Those Winner-Take-All Primaries

Happy July 4th! No More 'Politics as Usual' Palin Edition

Thursday, June 18, 2009

State of the Race: Virginia Governor (6/18/09)

[Click to Enlarge]

Of all the lines I'd imagine penning, this would probably not have been at the top of my list, but here goes:

That's more like it. An independent poll in the Virginia governor's race from Daily Kos.

When compared with the Anzalone Liszt poll conducted on behalf of the Democratic Governors Association that was released a couple of days ago, though, the Kos/R2K poll withstands the independence test better. A new poll, then, means FHQ must recalculate the average standing of the candidates in the governor's race in the Commonwealth of Virginia. And what do you know? The deadheat from a week ago is still a deadheat today. But Bob McDonnell edges past Creigh Deeds here (and the map above becomes ever so slightly more reddish purple than it was).

Our typical caveats remain, however. There are still but five polls total in this race (since the point at which Deeds became viable following the Washington Post endorsement), and that means we are still likely to see some "wild" fluctuations. With the "Deeds leads" poll not receiving the full weight of the most recent poll, the full body of evidence outside of the most recent poll continues to favor McDonnell overall. That should be said with a note of caution, though. This is still a close race; not the virtual tie from a week ago, but still very much within the margin of error.

That said, I'll add one more helpful graphic now that we have a couple of averages in the books. As we saw recently in the 2012 presidential primary poll tracking, two points on a graph aren't terribly exciting. But you have to start somewhere. To keep track of the ebbs and flows of the FHQ averages over the course of the campaign is something that I think will prove instructive for us all in both the New Jersey and Virginia contexts. [And this is something that would have been helpful during the presidential race a year ago on a state by state basis. When I have more time on my hands for combing back through that data, that may be something I'll add here in the future.] The graph really tells the tale -- at least as it is scaled with the inclusion of the undecideds. This is just a close race.

It should continue to be fun to track.

[Click to Enlarge]


Recent Posts:
How Not to Emerge as a 2012 GOP Darkhorse

A Week Later, Deeds Still Leads, but...

Republicans and Democrats to Work Together to Prevent Frontloading/National Primary?

Tuesday, June 16, 2009

A Week Later, Deeds Still Leads, but...

A new poll is out from Anzalone Liszt Research on the Virginia governor's race. Taegan Goddard over at Political Wire has the results up showing Deeds ahead of current Virginia attorney general, Bob McDonnell, by a 42%-38% count. That spread is in line with the Rasmussen poll conducted in the aftermath of last week's Democratic primary, but the drawback here is that this poll was conducted by the Democratic polling firm on behalf of the Democratic Governors Association. No, that doesn't compromise the results necessarily, but the numbers certainly have to be taken with a grain of salt.

Our traditional action in this case is to hold out any polling from partisan firms from our polling averages. That was done during the 2008 election cycle and will continue to be employeed in FHQ's tracking of the New Jersey and Virginia gubernatorial races.

In other words, interesting results, but...

UPDATE: Pollster's got the results up now with a link to a pdf of the full poll.


Recent Posts:
Republicans and Democrats to Work Together to Prevent Frontloading/National Primary?

Monday Reading: GOP Behind the Eight Ball?

State of the Race: Virginia (6/11/09)