04 November 2008

PC 12 - Distillation: Obama 356 / McCain 182 - however, with adjustment: Obama 375 / McCain 163

This is the complete statistical analysis on the polling for the FINAL POLL CONVERGENCE, No. 12, for the time frame of 10/26 -11/04/08 (AM), or 10 days. A few last second polls from 11/04 may come on board – if they change the statistics in a major way, I will report it. If not, then not.

In this time-frame there were a whopping 264 polls from 46 out of 51 “states”.

Critical information: This is the third time in this election cycle and the third time in a row that all three of the so-called “TRIFECTA” states (FL, PA, OH) are in the Obama column. Further, VA is still in the Obama column but with a leaner margin than in the last poll convergence. And Obama is showing a small but resilient lead in NC. All date points to a decisive Obama win today. If Georgia is too close to call or the networks declare VA pretty quickly for Obama, then we will know for sure that McCain has lost.

The exact polling results from POLL CONVERGENCE NO. 12, alphabetical, by state, with hyperlinks, can be found here.

 First, a look at the number of polls and how the candidates did, numerically speaking:

10/26-11/03
total
Obama
Obama %
McCain
McCain %
TIE
TIE %
PC 12
264
174
65.90%
79
29.92%
11
4.18%

Second, the frequency of polls from 10/15-10/25, per state, in descending order:

No. States for PC 12 No. for calculation
29 PA
21 OH
20 FL,
19 NC
18 VA
13 MO
12 NH
11 CO, IN
10 PA
9 AZ, GA, FL, OH, VA
8 NV CO, NC
7 MN, AZ, GA, IN, IA, MI, MN, MO, NH, NJ, OR, WV, WI
6
MI, NJ
AK, ME, MT, NV, NM
5
IA, MT, OR, WI
KY, WA
4
AK, KY, WV
AR,CA, IL, NY, RI, SC, SD, TX
3
AR, CA, ME, NM, WA
AL, CT, DE, LA, MD, MA, ND
2
AL, IL, LA, MS, NY, OK, SC, SD, UT
HI, KS, MS, OK, TN, UT, VT, WY
1
CT, DE, ID, KS, MA, ND, TN, TX, VT, WY
DC, ID, NE
0
DC, HI, MD, NE, RI


How to read the table above: the column for „States for PC12“ lists the state names next to the number of times those states were polled in this time frame. The column for „No. For Calculation“ lists the states next to the number of polls in the current calculation.

If the number of times a state has been polled for PC12 is higher than the gene-poll for the end-average, this means logically that all polls from PC11 have fallen completely out of the statistic. It is also no surprise that the bitterest of battleground states were polled the most.


Based only the statistics, I currently project the Electoral College as follows:

Obama: 28 states, 341 EV
McCain: 21 states, 171 EV
Tossup: 2 states (NC, MO), 26 EV

EV Margin: Obama + 170

Based on 99% statistic and 1% gut feeling, here my altered projection, without tossups:

Obama: 30 states, 375 EV (IN, MO, MT to Obama)
McCain: 20 states, 163 EV (ND to McCain)

EV Margin: Obama + 212


Here the table of all 51 "states".

Here's how it works (please read carefully):

Blue = democratic value
Red = republican value

The table runs in descending order of  Obama's winning margin averages, down to tossups (1, IN), and then in ascending order of McCain's winning margin averages. This is in the column “10/16-10/25. If nothing is in this column, it means that the state has not yet been polled in this month, in which case the values for “10/03-10/15” apply.

Values in bold= more than one poll for that state for PC 10.
Values in normal print= only one poll for that state for PC 10.
Values in normal italic = states with no new polls for PC 10

The column to the right of it, 10/03-10/15 , reflects the margins from the last poll data convergence (September 2008).

The next column, Diff. 1” is of course the calculated difference between these two margin averages. In this way, improvement or shortcoming can be measured.

The 2004” column reflects the winning margin from the 2004 GE.

The Diff. 2” column reflects the calculated difference between Obama's or McCain's margin average to Kerry's or Bush's averages from 2004.

1948-2004” is the average for the 56 year time period from 1948-2004 (15 cycles). This is interesting for purposes of education, but has no effect on the current statistics.


State
EV
10/26-11/03
10/16-10/25
Diff. 1
2004
Diff. 2
1948-2004

DC

3
69.00
69.00
0.00
79.84
-10.84
69.56
HI
4
36.00
36.00
0.00
8.74
+27.26
10.01

NY

31
29.00
27.75
+1.25
18.29
+10.71
5.42
VT
3
27.00
33.00
-5.00
20.14
+6.86
6.93
IL
21
24.90
23.53
+1.37
10.34
+14.56
0.09

CA

55
24.00
21.00
+3.00
9.95
+14.05
0.96
MA
12
21.33
23.67
+2.34
25.16
-3.83
14.75

CT

7
21.33
17.67
+3.66
10.37
+10.96
0.03

DE

3
20.70
16.05
+4.65
7.59
+13.11
0.05
RI
4
18.10
18.10
0.00
20.75
-2.65
16.23
MD
10
17.30
17.30
0.00
12.98
+4.32
2.85
ME
4
17.15
14.16
+2.99
9.00
+8.15
4.67

NJ

15
16.00
14.58
+1.42
6.68
+9.32
3.43

OR

7
15.57
14.60
+0.97
4.16
+11.41
1.58

WA

11
14.60
15.33
-0.73
7.18
+7.42
0.97
MI
17
14.44
13.88
+0.56
3.42
+11.02
0.42

IA

7
12.75
12.32
+0.43
0.67
+13.42
3.84
MN
10
12.00
11.33
+0.67
3.48
+8.52
6.15

WI

10
11.34
11.93
-0.59
0.38
+10.96
1.83

NH

4
10.43
7.61
+2.82
1.37
+9.06
11.34

NM

5
9.26
6.12
+3.14
0.79
+10.05
4.10

PA

21
7.30
11.93
-4.63
2.50
+4.80
0.25
NV
5
6.83
1.36
+5.47
2.59
+9.42
9.94

CO

9
6.62
4.75
+1.87
4.67
+11.29
9.41
VA
13
4.93
5.43
-0.50
8.20
+13.13
10.14

OH

20
2.31
5.67
-3.36
2.11
+4.44
5.15

FL

27
1.79
2.05
-0.36
5.01
+6.80
8.65

ND

3
1.34
1.67
-0.33
27.36
+28.70
18.3

--

--


--
--
--
--

--

--


--
--
--
--
NC
15
0.62
1.21
-0.59
12.44
+13.06
4.13

MO

11
0.46
1.39
+1.85
7.20
-6.74
0.21

--

--


--
--
--
--

--

--


--
--
--
--
IN
11
1.18
0.88
+0.30
20.68
-19.50
13.29
MT
3
1.66
3.25
-1.59
20.50
-18.84
9.54
GA
15
3.85
4.85
1.00
16.60
-12.75
3.17
AZ
10
4.43
19.00
-14.57
10.47
-6.04
14.60
SD
3
7.50
8.00
-0.50
21.47
-13.97
13.31

WV

5
7.66
5.82
+1.84
12.86
-5.20
3.81
LA
9
10.43
16.00
-5.57
14.51
-4.08
3.64
MS
6
10.50
9.00
+1.50
19.69
-9.19
9.66

AR

6
10.75
15.00
-4.25
9.76
+0.99
3.53
KY
8
12.40
12.60
-0.20
19.86
-7.36
4.32
SC
8
12.55
15.05
-2.50
17.08
-4.53
7.47
TX
34
13.00
13.66
-0.66
22.86
-9.86
5.49
TN
11
14.00
13.33
+0.67
14.27
-0.27
4.23

AK

3
14.58
17.66
-3.08
25.55
-10.97
15.80

KS

6
17.00
12.50
+4.50
25.38
-8.38
19.84
NE
5
19.00
19.00
0.00
33.22
-14.22
24.44

AL

9
19.33
23.50
-4.17
25.62
-6.29
10.25
ID
4
23.00
37.00
-15.00
38.12
-16.12
24.44
WY
3
23.00
22.00
+1.00
39.79
-16.79
20.97
UT
5
24.00
36.00
-12.00
45.54
-21.54
26.59
OK
7
30.15
27.55
+2.60
31.23
-1.08
16.03

Here a table comparing the margin spreads (numerically):

Margin
BO 11/03
EV
BO 10/25
EV
BO 11/03
EV
JM 10/25
EV
+30 and up
2
7
3
10
1
7
2
9
+20 to +30
7
132
4
119
3
12
3
19
+10 to +20
11
99
13
126
11
105
10
100
+5 to +10
4
40
4
42
2
8
3
14
+2 to +5
2
33
2
36
2
25
2
18
+1 to +2
2
30
4
34
2
14
--
--
Subtotal

341
30
367

171
20
160
Under +1
1
15
--
--
1
11
1
11
TOTAL

356
30
367

182
21
171

There was no new polling for RI and MD, so the statistical probability is VERY HIGH that both of these state will come in for Obama at +20 or more.

For the margin predictions, here first the national average for 11/03-11/04:

Pollster
Date
Obama
McCain
Margin
Gallup FINAL
11/04
53
42
+11
11/04
52
46
+6
11/04
51
46
+5
11/04
54
43
+11
GWU/Battleground
11/04
49
44
+5
Marist
11/04
52
43
+9
Harris
11/04
52
44
+8
ARG
11/04
53
45
+8





11/03
50
45
+5
FOX
11/03
50
43
+7
CBS
11/03
51
42
+9
IPSOS
11/03
53
46
+7
IBD/TIPP
11/03
48
43
+5
Pew
11/03
52
46
+6
NBC/WSJ
11/03
51
43
+8
ABC/WAPO
11/03
53
44
+9
DEMCORPS
11/03
51
44
+7





Average:

52.00
44.13
+7.87
Average 2:

51.47
44.06
+7.41
Average 3:

51.47
44.13
+7.34
MEAN:

51.65
44.11
+7.54
-GE 2004



+2.46
ADJ. MEAN



+10.00

Average 1 = only polls from 11/04
Average 2 = polls from 11/03 ans 11/04 (17 Endpolls)
Average 3 = average without the highest and the lowest individual margin.
Mean = the mean of these averages.

BTW, a national average of +7.54, assuming 2% for third parties, would result in the following national popular vote:

Obama: 52.77%
McCain:45.23%
Other: 2%

And now, the margin predictions for ELECTION DAY, based on the current national average being incorporated as 1/8th of the equation. These adjusted margin predictions are most critical for the so-called battleground states, for the smaller the margin average, the larger the effect of the national average as a component of the whole thing. Obviously, if you are winning a state with an average of 30+, the national average has little pull on how you are doing. But in battlegrounds like IN, MO, NC and so forth, this is critical, so I am only showing the part of the list where a switch from one candidate in the pure statistic to another in the adjusted statistic in possible (just as I did in every BATTLEGROUND STATE REPORT).


State
EV
Margin
* .875
Natl.
* .125
Adj.
NV
5
6.83
5.98
10.00
+1.25
+7.23

CO

9
6.62
5.79
10.00
+1.25
+7.04
VA
13
4.93
4.31
10.00
+1.25
+5.56

OH

20
2.31
2.02
10.00
+1.25
+3.27

FL

27
1.79
1.56
10.00
+1.25
+2.81

ND

3
1.34
1.17
10.00
+1.25
+2.42
NC
15
0.62
0.54
10.00
+1.25
+1.79

MO

11
0.46
0.40
10.00
-1.25
+0.85
IN
11
1.18
1.03
10.00
-1.25
+0.22
MT
3
1.66
1.45
10.00
-1.25
+0.20
GA
15
3.85
3.37
10.00
-1.25
+2.12
AZ
10
4.43
3.88
10.00
-1.25
+2.63

The two states shaded in grey (MO, IN) are examples of a switch by inclusion of the national polling influence. There are other factors as well, most of all the ground organization by the Obama campaign that has obviously outstripped the ground organization of the McCain campaign.

But also the MT statistic, with the national polling calculated in, makes the thing really too close to tell and for this reason I do believe that despite the raw statistics on their own, Obama will win IN, MO and MT today.

However, ND, which looks good on paper, I believe will be picked up by John McCain. There were far too many undecideds in those polls, and statistically, the chances are great that those undecideds will break for McCain.

No comments:

Post a Comment

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