This is an update to the Generic poll model using two projection methods:
1) calculating the 120-poll linear trend and adding an undecided voter
allocation adjustment (UVA) of 60% to the Democratic final trend.
2) averaging the final 10 Generic polls and adding the UVA adjustment.
Key results:
The projected Democratic share was 56.4% in both methods.
The projected shares matched to within .05%.
The actual Democratic vote was 51.3%.
The probability of a 5.1% discrepancy occurring by chance is near ZERO.
Therefore, there were errors in the 120 Generic polls or...
there were errors in the vote count.
If you believe that the polls were off, take a trip to Sarasota.
It's beautiful this time of the year.
Try to find some time to check out the voting machines in FL-13.
And see if you can find those 18,000 missing votes.
________________________________________________
Assuming the Democrats captured 60% of the undecided vote,
1-The 120-Generic poll trendine projected a 56.43% share
2-The Final 10 Generic poll average projected a 56.38% share
1) Projection based on 120-Generic Poll trendline:
Dem = 46.98+ .0419x
GOP = 38.06+ .0047x
Substituting x=120 and allocating 60% (UVA) to the Democrats:
........Trend + UVA = Projection
Dem= 52.01 + 4.42= 56.43%
Rep = 38.62 + 2.95= 41.57%
Assuming a 1.5% MoE and 60% UVA, we calculate the probability that the
Democratic vote share would decline 5.13% from the Generic Poll (56.43%) to
the actual vote (51.3%).
Using the Excel Normal Distribution function:
Probability = 1.0E-11 = NORMDIST(0.513,0.5643,0.015/1.96,TRUE)
or 1 in 99,128,359,468
SENSITIVITY ANALYSIS I
Probabilities of Democratic vote discrepancies from the final point on the
120-Generic Poll trend line for various MoE and UVA assumptions.
UVA 50% 55% 60% 65% 70% 75%
Dev 4.4% 4.8% 5.1% 5.5% 5.9% 6.2%
Vote 55.7% 56.1% 56.4% 56.8% 57.2% 57.5%
Margin 13.4% 14.1% 14.9% 15.6% 16.3% 17.1%
The Probability is 1 in X, where X is given in the table below.
(t=trillion, b=billion, m=million, k=thousand)
UVA 50% 55% 60% 65% 70% 75%
MoE Probability of Vote Discrepancy
1.25% 358b 24t 2252t nc nc nc
1.50% 213m 4b 99b 3t 114t 4503t
1.75% 2.3m 21m 220m 2.7b 40b 704b
2.00% 120k 650k 4.1m 28m 226m 2b
2.25% 15k 60k 255k 1.2m 6.2m 36m
2.50% 3.5k 111k 35k 124k 476k 2m
2.75% 1.2k 2.9k 7.9k 23k 69k 228k
3.00% .5k 1.1k 2.5k 6k 16k 43k
________________________________________________
2- Projection based on the final 10 Generic Polls
Undecided: 7.3%
Assumption: 60% of undecided vote (UVA) allocated to Democrats
..............Dem GOP Margin
Avg Date 52.0 38.7 13.3
Harr 1023 47 33 14
AP 1030 56 37 19
CBS 1101 52 33 19
Nwk 1103 54 38 16
TIME 1103 55 40 15
Pew 1104 47 43 4
ABC 1104 51 45 6
USA 1106 51 44 7
CNN 1106 58 38 20
FOX 1106 49 36 13
..............Dem GOP Margin
Average....... 52.0% 38.7% 13.3%
UVA........ 4.4% 2.9% 0.9%
----- ----- ----- -----
Projection. 56.4% 41.6% 14.8%
Vote....... 51.3% 46.4% 4.9%
----- ----- ----- -----
Discrepancy -5.1% 4.8% -9.9%
________________________________________________
SENSITIVITY ANALYSIS II:
Probability of vote discrepancies from the projected 56.4% Democratic vote,
based on the final 10 Generic Polls, for various MoE and UVA assumptions.
Base Case:
Undecided vote: 7.3%
UVA to Dems: 56.4% (equals 10-poll average)
Average Dem vote: 52.01%
UVA adjustment: 4.42%
--------------- -----
Projected Vote: 56.43%
UVA 50% 56.4% 60% 65% 70% 75%
Dev 4.4% 4.8% 5.1% 5.5% 5.8% 6.2%
Vote 55.7% 56.1% 56.4% 56.8% 57.1% 57.5%
Margin 13.3% 14.2% 14.8% 15.5% 16.2% 17.0%
Probability is 1 in X, where X is given in the table
(t=trillion, b=billion, m=million, k=thousand)
UVA 50% 56.4% 60% 65% 70% 75%
MoE Probability of Vote Discrepancy
1.25% 219b 36t 1286t nc nc nc
1.50% 151m 5.4b 62b 1.8t 63t 3002t
1.75% 1.8m 25m 157m 1.9b 26b 428b
2.00% 99k 766k 3m 21m 161m 1.4b
2.25% 13k 68k 207k 950k 4.8m 27m
2.50% 3k 12k 29k 102k 381k 1.5m
2.75% 1k 3.2k 6.8k 19k 58k 186k
3.00% 446 1.2k 2.2k 5.3k 14k 37k