Wednesday, July 25, 2012

The Electoral College Map (7/25/12)

Polling releases picked up on Wednesday as compared to Tuesday. Michigan was back and on the move. There were also new polls out of Georgia and New York (a couple of Strong states) as well as Pennsylvania.

New State Polls (7/25/12)
State
Poll
Date
Margin of Error
Sample
Obama
Romney
Undecided
Poll Margin
FHQ Margin
Georgia
7/??
+/- ?.?%
591 ?? voters
41
50
--
+9
+10.05
Michigan
7/21-7/23
+/- 4.1%
579 likely voters
53
39
8
+14
+5.20
Michigan
7/23
+/- 3.4%
825 likely voters
44
45
--
+1
--
New York
7/17-7/23
+/- 2.3%
1779 registered voters
55
32
10
+23
+24.15
Pennsylvania
7/21-7/23
+/- 3.6%
758 likely voters
49
43
8
+6
+6.65

Polling Quick Hits:
Georgia:
If the pendulum is swinging back toward the Republican candidate in 2012, the new Insider Advantage poll of the Peach state was further evidence of that. Toward the tail end of the 2008 race, Georgia tightened, while still remaining about 5 points outside of Obama's grasp. A state that in 2008 was on the periphery of being if not competitive, then perhaps worthy of some resource expenditure, has inched away from the Democrat and toward his Republican challenger in 2012. In the light polling that has been done in Georgia, Romney has maintained a comfortable 10 point advantage. The terrain on which this election fight will be waged is not near or even threatening to be near the the line between lean and strong Republican states. Instead, in 2012, the battle will be fought over states bluer in hue. Like...


Changes (July 25)
StateBeforeAfter
MichiganToss Up ObamaLean Obama
Michigan:
There is nothing quite like a day with the release of two widely divergent polls in one state. Such was the case in the Great Lakes state today. [Incidentally, it is even more noteworthy when the average of the two polls essentially confirms the margin in a Rasmussen poll released just yesterday.] In actuality, the +14 (Obama) from PPP and the +1 (Romney) from Mitchell Research do serve as bookends on an array of polls that in FHQ's weighted average now place Michigan in the Lean Obama category. Of course, news late yesterday from PPP that today's poll would show a double digit lead for the president in the state forced some quick calculations on a preliminary average. Alone the PPP survey would have pushed Michigan's average beyond the current weighted average of Nevada. With the Mitchell survey included the impact was muted; inching Michigan just over into a new category.

New York:
Is anyone surprised? No, New York has remained firmly within the president's tally of states, but these polls are helpful in determining the relative order of states. The more information we are privy to the better consumers we will all be.

Pennsylvania:
Just south in the Keystone state, Obama held a six point advantage over Mitt Romney in the latest PPP poll there. The survey represents a small improvement (+2) for Romney since the last time PPP was in the field, but is well within the range of polls conducted in the last two months.

--
Tangent:
Both PPP surveys released today used and reported likely voter screened results. One question that continues to pop into my head regarding how the various polling firms transition from registered voter to likely voter screens in their reporting. Most tend to be one or the other at this point. Rasmussen has been issuing likely voter results for a while now as Quinnipiac or NBC/Marist have continued to utilize just registered voters. PPP, on the other hand, has not been consistent one way or the other. Pennsylvania and Michigan were likely voter surveys this week while New Mexico last week was a registered voter poll. FHQ is not suggesting anything is amiss here, but rather that I don't know of (or haven't see) the reasoning for the variation across states. My hunch is that it has more to do with gathering a large enough likely voter sample to be representative than anything else. But that's just  a guess.



Today's polls did slightly shake up the map and the Electoral College Spectrum. Michigan, as mentioned before, slipped into the Lean Obama category, but holds its spot though not color on the Spectrum. New York flip-flopped with Hawaii (...again), and Georgia (thankfully) switched places with West Virginia. That the Mountain state has been the state closest to shifting into Lean Romney territory masked the fact that Georgia is likely closer. Remember that states where no polling has been conducted are measured using the election results from the last three presidential cycles. Gore's 2000 finish in West Virginia pulled the average down and provided some insulation between Georgia and its place straddling the line between categories (inconsequential though it may be in this cycle).

The Electoral College Spectrum1
RI-4
(7)2
NJ-14
(160)
NH-4
(257)
IN-11
(159)
ND-3
(55)
HI-4
(11)
CT-7
(167)
OH-183
(275/281)
MT-3
(148)
MS-6
(52)
NY-29
(40)
MN-10
(177)
CO-9
(284/263)
GA-16
(145)
AL-9
(46)
VT-3
(43)
WA-12
(189)
VA-13
(297/254)
WV-5
(129)
KY-8
(37)
MD-10
(53)
NM-5
(194)
IA-6
(303/241)
SC-9
(124)
KS-6
(29)
CA-55
(108)
OR-7
(201)
FL-29
(332/235)
LA-8
(115)
AK-3
(23)
MA-11
(119)
PA-20
(221)
NC-15
(206)
NE-5
(107)
OK-7
(20)
IL-20
(139)
WI-10
(231)
MO-10
(191)
AR-6
(102)
ID-4
(13)
DE-3
(142)
NV-6
(237)
AZ-11
(181)
TX-38
(96)
WY-3
(9)
ME-4
(146)
MI-16
(253)
TN-11
(170)
SD-3
(58)
UT-6
(6)
1 Follow the link for a detailed explanation on how to read the Electoral College Spectrum.

2 The numbers in the parentheses refer to the number of electoral votes a candidate would have if he won all the states ranked prior to that state. If, for example, Romney won all the states up to and including Ohio (all Obama's toss up states plus Ohio), he would have 272 electoral votes. Romney's numbers are only totaled through the states he would need in order to get to 270. In those cases, Obama's number is on the left and Romney's is on the right in italics.

3 Ohio
 is the state where Obama crosses the 270 electoral vote threshold to win the presidential election. That line is referred to as the victory line.

The states remained the same on the Watch List today, but Michigan's move into the Lean category now means that poll watchers should keep an eye on polls out of the Great Lake state that show a small enough margin to pull Michigan back into Toss up territory.

The Watch List1
State
Switch
Georgia
from Strong Romney
to Lean Romney
Michigan
from Lean Obama
to Toss Up Obama
Missouri
from Toss Up Romney
to Lean Romney
Nevada
from Lean Obama
to Toss Up Obama
New Hampshire
from Toss Up Obama
to Lean Obama
New Mexico
from Strong Obama
to Lean Obama
North Carolina
from Toss Up Romney
to Toss Up Obama
Washington
from Strong Obama
to Lean Obama
West Virginia
from Strong Romney
to Lean Romney
1 Weighted Average within a fraction of a point of changing categories.

Please see:



Are you following FHQ on TwitterGoogle+ and Facebook? Click on the links to join in.

Tuesday, July 24, 2012

The Electoral College Map (7/24/12)

Tuesday was a slow day on the state-level polling front. The only addition was a Rasmussen poll out of Michigan. Elway released new numbers in the Washington state gubernatorial race in for a handful of initiatives in the Evergreen state.

...but there were no presidential numbers. Given that the margin in that survey of the governor's race favored Democrat, Jay Inslee, for the first time, that may serve as some indication as to the underlying partisan makeup of this particular sample. On its face, that would favor the president in a state he carried by 17 points, but we will hold off on speculating further until or if the Seattle-based firm releases any presidential numbers.

New State Polls (7/24/12)
State
Poll
Date
Margin of Error
Sample
Obama
Romney
Undecided
Poll Margin
FHQ Margin
Michigan
7/23
+/- 4.5%
500 likely voters
48
42
4
+6
+4.90

Polling Quick Hits:
Michigan:
On its surface, the Rasmussen poll appears to behave well within the expectations in Michigan. Yet, while most polls have recently shown a tighter race (<5%) in the Great Lakes state, Rasmussen has so far not followed suit. The six point spread in the poll conducted this past Monday is only two points closer than the June survey of the state by Rasmussen. And that two points was shaved off of Obama's share while Romney remained steady at 42%. Nevertheless, none of this did much to sway the FHQ weighted average in Michigan. There was a very slight shift toward Obama, pushing the state within a tenth of an averaged percentage point of moving into the Lean Obama category.



As for changes on the map or in the Electoral College Spectrum, it was close but not quite. With just one poll -- one confirming the current state of the race there -- there just was not that much chance of a change anyway. Michigan is to the Toss Up/Lean line as Washington, New Mexico and Minnesota have been to the Strong/Lean line recently: new information has translated into movement, but not category-changing movement.

...yet.

The Electoral College Spectrum1
RI-4
(7)2
NJ-14
(160)
NH-4
(257)
IN-11
(159)
ND-3
(55)
NY-29
(36)
CT-7
(167)
OH-183
(275/281)
MT-3
(148)
MS-6
(52)
HI-4
(40)
MN-10
(177)
CO-9
(284/263)
WV-5
(145)
AL-9
(46)
VT-3
(43)
WA-12
(189)
VA-13
(297/254)
GA-16
(140)
KY-8
(37)
MD-10
(53)
NM-5
(194)
IA-6
(303/241)
SC-9
(124)
KS-6
(29)
CA-55
(108)
OR-7
(201)
FL-29
(332/235)
LA-8
(115)
AK-3
(23)
MA-11
(119)
PA-20
(221)
NC-15
(206)
NE-5
(107)
OK-7
(20)
IL-20
(139)
WI-10
(231)
MO-10
(191)
AR-6
(102)
ID-4
(13)
DE-3
(142)
NV-6
(237)
AZ-11
(181)
TX-38
(96)
WY-3
(9)
ME-4
(146)
MI-16
(253)
TN-11
(170)
SD-3
(58)
UT-6
(6)
1 Follow the link for a detailed explanation on how to read the Electoral College Spectrum.

2 The numbers in the parentheses refer to the number of electoral votes a candidate would have if he won all the states ranked prior to that state. If, for example, Romney won all the states up to and including Ohio (all Obama's toss up states plus Ohio), he would have 272 electoral votes. Romney's numbers are only totaled through the states he would need in order to get to 270. In those cases, Obama's number is on the left and Romney's is on the right in italics.

3 Ohio
 is the state where Obama crosses the 270 electoral vote threshold to win the presidential election. That line is referred to as the victory line.

The good news is that Public Policy Polling was in the field with a new survey in Michigan over the weekend. The 14 point gap -- the largest of any poll conducted in Michigan in 2012 -- the firm found there in May will likely be fodder for "Obama's lead slips" stories when those results are released either Wednesday or Thursday. All of that, however, will depend on what the results are.

FHQ would be surprised if they find a +14 margin in favor of Obama this time around. But there I go again: speculating on polls that have yet to be released. [UPDATE:] The results to be released Wednesday in Michigan apparently still show Obama up by double digits.

The Watch List1
State
Switch
Georgia
from Strong Romney
to Lean Romney
Michigan
from Toss Up Obama
to Lean Obama
Missouri
from Toss Up Romney
to Lean Romney
Nevada
from Lean Obama
to Toss Up Obama
New Hampshire
from Toss Up Obama
to Lean Obama
New Mexico
from Strong Obama
to Lean Obama
North Carolina
from Toss Up Romney
to Toss Up Obama
Washington
from Strong Obama
to Lean Obama
West Virginia
from Strong Romney
to Lean Romney
1 Weighted Average within a fraction of a point of changing categories.

Please see:
Frequently Asked Questions about FHQ graduated weighted average methodology.



Are you following FHQ on TwitterGoogle+ and Facebook? Click on the links to join in.

Monday, July 23, 2012

The Electoral College Map (7/23/12)

At the dawn of another new campaign week, there are a couple of new polls in traditionally strong blue states. One typical, the other closer than expected.

New State Polls (7/23/12)
State
Poll
Date
Margin of Error
Sample
Obama
Romney
Undecided
Poll Margin
FHQ Margin
California
7/16-7/17
+/- 3.4%
812 likely voters
51.9
32.6
10.3
+19.3
+19.36
Minnesota
7/17-7/19
+/- 4.3%
552 likely voters
46
40
7
+6
+11.06

Poll Quick Hits:
California:
I don't know that there is all that much to say about a survey of the Golden state showing the Democratic nominee up by nearly 20%. There isn't much. The Pepperdine poll appears to lowball both candidates' shares in the previous polls in California while still capturing a margin between the two major party candidates that is consistent with much of the survey work conducted in the state prior to July.

Minnesota:
The Survey USA poll in the Land of 10,000 Lakes was more revealing, though perhaps not to the extent that some may suggest, than the California poll above. On its surface, it is tempting to lump Minnesota in with neighboring and more competitive states like Iowa and Wisconsin -- and that may ultimately prove to be the case -- but we just don't have a solid enough body of evidence to reach that conclusion as of yet. On the one hand, Iowa was closer than Minnesota was closer than Wisconsin in the 2008 election, but on the other, the six polls in Iowa and five in Minnesota do not necessarily provide the confidence in the current FHQ weighted average that the 22 polls in Wisconsin to this point do. What has been made clear in the last week or so is that there is a group of strong Obama states that probably deserve more polling in the coming weeks. Minnesota, New Mexico and Washington are all hovering around the +10 Obama range. The averages have been brought closer due to the influence of new polling information -- albeit in a shallow, intra-state pool -- and additional data from other polling firms may help to build a more robust picture of either a tightening race or one that in those states keeps them on solid ground for the president. If it is the former, the Romney campaign may be able to play offense to some extent and force Obama to reconsider continued resource expenditures in, say, North Carolina (post-Charlotte convention). The Tarheel state is endangered on the president's board as the lone red state within range for the Democrat.



On this Monday 15 weeks before election day, the map remains unchanged in terms of state classifications. However, Minnesota jumped up a couple of spots on the Electoral College Spectrum to join the aforementioned Washington/New Mexico group at the periphery of the Lean Obama category. There is some separation after those three states and Connecticut and New Jersey, but there has been limited polling in the Nutmeg state to this juncture in the general election race. This is why the Watch List is helpful. It affords the opportunity to see which states are close to switching categories.

The Electoral College Spectrum1
RI-4
(7)2
NJ-14
(160)
NH-4
(257)
IN-11
(159)
ND-3
(55)
NY-29
(36)
CT-7
(167)
OH-183
(275/281)
MT-3
(148)
MS-6
(52)
HI-4
(40)
MN-10
(177)
CO-9
(284/263)
WV-5
(145)
AL-9
(46)
VT-3
(43)
WA-12
(189)
VA-13
(297/254)
GA-16
(140)
KY-8
(37)
MD-10
(53)
NM-5
(194)
IA-6
(303/241)
SC-9
(124)
KS-6
(29)
CA-55
(108)
OR-7
(201)
FL-29
(332/235)
LA-8
(115)
AK-3
(23)
MA-11
(119)
PA-20
(221)
NC-15
(206)
NE-5
(107)
OK-7
(20)
IL-20
(139)
WI-10
(231)
MO-10
(191)
AR-6
(102)
ID-4
(13)
DE-3
(142)
NV-6
(237)
AZ-11
(181)
TX-38
(96)
WY-3
(9)
ME-4
(146)
MI-16
(253)
TN-11
(170)
SD-3
(58)
UT-6
(6)
1 Follow the link for a detailed explanation on how to read the Electoral College Spectrum.

2 The numbers in the parentheses refer to the number of electoral votes a candidate would have if he won all the states ranked prior to that state. If, for example, Romney won all the states up to and including Ohio (all Obama's toss up states plus Ohio), he would have 272 electoral votes. Romney's numbers are only totaled through the states he would need in order to get to 270. In those cases, Obama's number is on the left and Romney's is on the right in italics.

3 Ohio
 is the state where Obama crosses the 270 electoral vote threshold to win the presidential election. That line is referred to as the victory line.

And even though Minnesota is not on the Watch List, it is not all that far away either. Therein lies one potential problem with the methodology FHQ utilizes. Again, if you will go back to the initial 2012 electoral college post, I described the tradeoff between a responsive average and one that bounces around too much, prone to outlier polls. In that post I discussed that it would likely be after the conventions before FHQ began to even further discount past polls. In 2008 around a similar point, the discount on the past surveys was doubled. That obviously gives more weight to the most recent poll. And that will be an appropriate action after the 2012 conventions. But now is probably a bit too soon.

Yet, just as an exercise, let's have a look at the impact such a move would have on the averages of a couple of states: Florida and Minnesota. By increasing the discount on the past polls in Florida and placing more relative weight on the most recent poll (+5 Obama, Survey USA poll from late last week), the average increases from +1.21 to +1.44 in the president's favor. A similar move in Minnesota, alternatively, has a greater impact, reducing the margin from +11.07 to +9.68. Substantively, that may not be a big change in the big picture, but it would shift Minnesota from a Strong Obama state to a Lean Obama state by the FHQ measure.

The difference? Well, the weighting of various polls matters, but so too do the number of polls conducted in a state. The far greater number of polls in Florida gives a better indication that the Sunshine state does ever so slightly favor the president at the moment. The scant polling in Minnesota, however, leaves an incomplete picture or at least one that is more vulnerable to outlier polls. The point here is not to say that either way is right. Rather, the idea is to be transparent about what is reflected in the figures and the methodology behind them.

The Watch List1
State
Switch
Georgia
from Strong Romney
to Lean Romney
Michigan
from Toss Up Obama
to Lean Obama
Missouri
from Toss Up Romney
to Lean Romney
Nevada
from Lean Obama
to Toss Up Obama
New Hampshire
from Toss Up Obama
to Lean Obama
New Mexico
from Strong Obama
to Lean Obama
North Carolina
from Toss Up Romney
to Toss Up Obama
Washington
from Strong Obama
to Lean Obama
West Virginia
from Strong Romney
to Lean Romney
1 Weighted Average within a fraction of a point of changing categories.




Are you following FHQ on TwitterGoogle+ and Facebook? Click on the links to join in.