01 October 2012

Battleground Report 09/30/2012 - T-minus 37 - DELUXE VERSION

Sunday, September 30, 2012, was exactly 37 days before the General Election on November 6, 2012.

Here the corresponding BATTLEGROUND REPORT from 37 days before the GE 2008: BATTLEGROUND REPORT from 09/28/2008. That report from 2008 was a DELUXE VERSION, and so is this one:
-it will compare 09/30 to 09/29, as expected.

-but will do a complete pass in review of the last week and compare 09/30 to 09/24. Explanation as we get there.

All of the 2012 polling for all states, including national figures, including specialty polls, including the newly published RED STATE REPORT and of course, including the BATTLEGROUND STATE REPORT, is HERE for 09/30 at GoogleDocs.

The BATTLEGROUND REPORTS run also with the national numbers. Here is the NATIONAL POLLING OUTLOOK thread, with links to every day going back to September 13, 2012. Want to see the numbers for a specific day? Go to that link.

The LEGEND for the following EXCEL table is HERE. In fact, you need to read it in order to develop an "eye" for the table:
StateEV% of ECPolls30.09.1229.09.12Shift20082004SWING (H-I)End polling 2008Compare E – KCompare E – GE008
GA162,97%0 / 213,5013,500,005,2016,6011,403,859,658,30
NE50,93%0 / 1011,0011,000,0014,9333,2218,2919,00-8,00-3,93
MO101,86%0 / 49,129,120,000,137,207,070,468,668,99
IN112,04%1 / 38,008,000,00-1,0320,6821,711,186,829,03
TN30,56%0 / 17,007,000,0015,0614,270,7914,00-7,00-8,06
MT30,56%0 / 27,007,000,002,3820,5018,121,645,364,62
AZ112,04%1 / 46,757,67-0,928,4810,471,994,432,32-1,73
----
--------------------
NE-0210,19%0 / 10,000,000,001,2121,6920,484,00-4,00-1,21
NC152,79%1 / 61,830,831,000,3312,4312,760,621,211,50
FL295,39%0 / 162,352,350,002,815,017,821,790,56-0,46
CO91,67%0 / 93,083,080,008,954,6713,626,62-3,54-5,87
IA61,12%1 / 74,004,000,009,530,6710,2012,75-8,75-5,53
VA132,42%2 / 134,314,310,006,308,2014,504,93-0,62-1,99
NV61,12%1 / 74,434,430,0012,492,5915,086,83-2,40-8,06
NH40,74%1 / 54,604,600,009,611,378,2410,43-5,83-5,01
----
--------------------
USA538100,00%4 / 104,634,74-0,117,262,469,727,54-2,91-2,63
----
--------------------
----
--------------------
OH183,35%2 / 115,455,220,234,582,116,692,303,150,87
WI101,86%0 / 86,196,190,0013,900,3813,5211,34-5,15-7,71
MN101,86%0 / 38,008,000,0010,243,486,7611,34-3,34-2,24
MI162,97%2 / 7 8,748,740,0016,443,4213,0214,44-5,70-7,70
PA203,72%2 / 108,798,790,0010,312,507,8111,34-2,55-1,52
OR71,30%0 / 19,009,000,0016,354,1612,1915,57-6,57-7,35
NM50,93%2 / 39,339,330,0015,130,7915,929,260,07-5,80


I was expecting a relatively slow weekend of polling before the first debate and was not disappointed, for the pollsters are surely waiting until after the first debate to do a flurry of polls, and in order to save costs, they are waiting. Seems pretty reasonable to me.

On Sunday, 4 polls came in for 3 states: NC, IA and OH.

Sunday, September 30, 2012, was exactly 37 days before the General Election on November 6, 2012.

Here the corresponding BATTLEGROUND REPORT from 37 days before the GE 2008: BATTLEGROUND REPORT from 09/28/2008. That report from 2008 was a DELUXE VERSION, and so is this one:
-it will compare 09/30 to 09/29, as expected.

-but will do a complete pass in review of the last week and compare 09/30 to 09/24. Explanation as we get there.

All of the 2012 polling for all states, including national figures, including specialty polls, including the newly published RED STATE REPORT and of course, including the BATTLEGROUND STATE REPORT, is HERE for 09/30 at GoogleDocs.

The BATTLEGROUND REPORTS run also with the national numbers. Here is the NATIONAL POLLING OUTLOOK thread, with links to every day going back to September 13, 2012. Want to see the numbers for a specific day? Go to that link.

The LEGEND for the following EXCEL table is HERE. In fact, you need to read it in order to develop an "eye" for the table:
StateEV% of ECPolls30.09.1229.09.12Shift20082004SWING (H-I)End polling 2008Compare E – KCompare E – GE008
GA162,97%0 / 213,5013,500,005,2016,6011,403,859,658,30
NE50,93%0 / 1011,0011,000,0014,9333,2218,2919,00-8,00-3,93
MO101,86%0 / 49,129,120,000,137,207,070,468,668,99
IN112,04%1 / 38,008,000,00-1,0320,6821,711,186,829,03
TN30,56%0 / 17,007,000,0015,0614,270,7914,00-7,00-8,06
MT30,56%0 / 27,007,000,002,3820,5018,121,645,364,62
AZ112,04%1 / 46,757,67-0,928,4810,471,994,432,32-1,73
----
--------------------
NE-0210,19%0 / 10,000,000,001,2121,6920,484,00-4,00-1,21
NC152,79%1 / 61,830,831,000,3312,4312,760,621,211,50
FL295,39%0 / 162,352,350,002,815,017,821,790,56-0,46
CO91,67%0 / 93,083,080,008,954,6713,626,62-3,54-5,87
IA61,12%1 / 74,004,000,009,530,6710,2012,75-8,75-5,53
VA132,42%2 / 134,314,310,006,308,2014,504,93-0,62-1,99
NV61,12%1 / 74,434,430,0012,492,5915,086,83-2,40-8,06
NH40,74%1 / 54,604,600,009,611,378,2410,43-5,83-5,01
----
--------------------
USA538100,00%4 / 104,634,74-0,117,262,469,727,54-2,91-2,63
----
--------------------
----
--------------------
OH183,35%2 / 115,455,220,234,582,116,692,303,150,87
WI101,86%0 / 86,196,190,0013,900,3813,5211,34-5,15-7,71
MN101,86%0 / 38,008,000,0010,243,486,7611,34-3,34-2,24
MI162,97%2 / 7 8,748,740,0016,443,4213,0214,44-5,70-7,70
PA203,72%2 / 108,798,790,0010,312,507,8111,34-2,55-1,52
OR71,30%0 / 19,009,000,0016,354,1612,1915,57-6,57-7,35
NM50,93%2 / 39,339,330,0015,130,7915,929,260,07-5,80


I was expecting a relatively slow weekend of polling before the first debate and was not disappointed, for the pollsters are surely waiting until after the first debate to do a flurry of polls, and in order to save costs, they are waiting. Seems pretty reasonable to me.

On Sunday, 4 polls came in for 3 states: NC, IA and OH.

It would seem like a paradox, but in spite of a TIED poll (48/48) from PPP (D) coming in last night for NORTH CAROLINA, the polling average actually went UP from Obama +0.83 to Obama +1.83

Why? Because, according to rules, which I enforce to the tee, any poll older than 14 days from the most recent poll falls out of the statistic, and no repeater polls are allowed. Therefore, the Rasmussen poll from 09/14, which is now 15 days old, and showed +6 for Romney, automatically fell out of the statistic. It was the  only of the 6 polls in the average for 09/29 to have a Romney lead. With that poll now to the wayside, a tied poll still means that within the 6 polls in the average for 09/30 (one poll came in, one poll left, net sum averaged polls: 6) the average moved toward Obama. And for this reason, the new average shows a jump for him, when in reality, the newest poll shows a tie. This is not abnormal in polling statistics. These jumps can be even more extreme if the time frame is smaller. For this reason, I keep the time-frame to 2 weeks until 21 days before the election, and then for the hotly contested states that will have maybe 4 or 5 polls coming in a day, I reduce the time frame to one week. And then, in the hotly contested states, right before the election, only the final polls, one per pollster, within 3-4 days of the election, will be in the statistic. RCP, TPM, HuffPo, WAPO and a slew of others use practically this same method, at least close to election day. So, my method is neither out of the ordinary nor is it unique.

On this same corresponding day in 2008, McCain was leading in North Carolina with a +2.70 average. This means that Obama is currently 4.53 points ahead of his poll average from the same time-frame in 2008.

---------------------------------------------------------

A new poll came in for IOWA: Selzer, which is absolutely the "Gold Standard" for the Jawhawk State. It shows Obama at +4. No polls fell out of the statistic, which means the polling "gene pool" in Iowa grew from 6 to 7 polls. On 09/27, Obama was also at +4.0 here and no new polls came in until the Selzer poll, which did not change the margin. The topline and bottom-line shifted ever so, but their shifts were identical, leaving the margin identical. Iowa, like Ohio is close to the National margin. It is just 0.63% under the current national margin of Obama +4.63, while Ohio is +0.82% above the national margin. It is very interesting how these two states, whose demographics are not all that similar to each other, are paralleling each other in polling data.

At this time in 2008, IOWA was not in the battlegrounds.

2 polls came in for Ohio. The PPP (D) poll shows Obama +4, which is pretty close to where the Ohio margin average has been for months now, +/-0.50. The second poll, a mail-in poll from the Columbus Dispatch, shows Obama at +9. This makes for a large broadband of polls for Ohio. Right now, there are 11 polls in the Ohio statistic and no polls fell out between 09/29 and 09/30. The poll values are all for Obama: +0.90, +3, +3.10, +4, +4, +5, +6, +7, +8, +9, +10. That is a huge disparity over two weeks, but it should be noted that all of  the high-end polls except the FOX poll, came in in the last 6 days. The history of the Columbus Dispatch poll is  so interesting that I have created an entire thread to this poll:

http://www.politicalhotwire.com/current-news-events/76308-columbus-dispatch-poll-oh-obama-9-a.html

I highly recommend this thread for further reading about the Columbus Dispatch poll and its polling accuracy since at least 1984.

Those two polls move Obama's margin from +5.22 on 09/29 to +5.45 on 09/30. On this corresponding day in 2008, McCain was leading in Ohio with a +1.40 average. This means that Obama is currently 7.85 points ahead of his poll average from the same time-frame in 2008.  It should be noted that the Romney campaign, which has allocated 2.5 million of it's ad campaign in the battleground states for this week, more than 1/2 of the total sum, has not allocated money for ads in Ohio this week. That is indeed unsual, to say the least.

----------------------------------------------------------------
On September 28, 2008, McCain was ahead in: WV (+5.30), MO (+3.80) NC (+2.70), IN (+2.20), NV (+1.70), OH (+1.40) and FL (+1.30).  Obama was ahead in VA (+0.40), NH (+1.30), MN (+4.00), WI (+4.10), PA (+4.30), CO (+5.00), MI (+5.60) and NM (+5.80). The statistics from 09/28/2008 were absolutely unchanged over 09/26/2008. There were 5 tossup states and 12 battleground states strictly within the +5 boundary. That is 5 more statistical battleground states than right now in 2012.

-----------------------------------------------------------------------------------------

Now comes the DELUXE part:

Here is the battleground table again, with only the battleground (or possibly battleground states) that have had polling in the last 6 day, and you can see the average for each of those days at one, a sort of "pass-in-review":
ADEEEEEEFGHIJKLM
StatePolls30.09.1229.09.1228.09.1227.09.1226.09.1225.09.1224.09.12Shift20082004SWING (H-I)End polling 2008Compare E – KCompare E – GE008
IN1 / 38,008,008,008,006,006,006,002,00-1,0320,6821,711,186,829,03
AZ1 / 46,757,677,677,677,337,337,33-0,588,4810,471,994,432,32-1,73
--------------------------------
NC1 / 61,830,830,830,830,600,600,601,230,3312,4312,760,621,211,50
FL0 / 162,352,352,352,352,351,911,830,522,815,017,821,790,56-0,46
CO0 / 93,083,083,083,083,082,782,780,308,954,6713,626,62-3,54-5,87
IA1 / 74,004,004,004,005,004,504,50-0,509,530,6710,2012,75-8,75-5,53
VA2 / 134,314,314,314,093,453,453,450,866,308,2014,504,93-0,62-1,99
NV1 / 74,434,434,434,434,834,835,00-0,5712,492,5915,086,83-2,40-8,06
NH1 / 54,604,604,604,002,752,752,751,859,611,378,2410,43-5,83-5,01
--------------------------------
USA4 / 104,634,744,534,744,384,193,960,677,262,469,727,54-2,91-2,63
--------------------------------
--------------------------------
OH2 / 115,455,225,224,554,554,003,891,564,582,116,692,303,150,87
MI2 / 7  8,748,748,748,618,618,618,610,1316,443,4213,0214,44-5,70-7,70
PA2 / 108,798,798,798,668,668,138,130,6610,312,507,8111,34-2,55-1,52
NM2 / 39,339,339,339,339,339,339,330,0015,130,7915,929,260,07-5,80


Column G, highlighted in pumpkin-orange, is the fun one: it shows the shift in margin in each state from 09/24-09/30.

There are only three negative shifts, one for Romney and two for Obama - over a week's time: AZ, IA and NV. Also interesting that all three of these negative shifts are exactly between -0.5 and -0.6, which is a narrow range, but could be decisive in a real tossup race. However, none of these three states are statistical tossups.

NM has not shifted at all, but it did get a new poll. See: IA, 09/30 over 09/29. Same story.

All other states where Obama is leading have shifted more toward him, most notably, NH and OH.

So, those are the cold, hard numbers for 09/30/2012, the DELUXE version.


--------------------------------------------------------------------------------
Facit: in the BATTLEGROUNDS, President Obama is on decidedly and measurably better footing right now than he was exactly four years ago against John McCain. If these margin shifts continue, then a number of states will be leaving the battleground list in the next three weeks.

No comments:

Post a Comment

Constructive comments and critique are always welcome. Please keep it polite and respectful.