TruthIsAll
06-03-2008, 11:06 AM
2004 Pre-election Polling Analysis
TruthIsAll
http://www.geocities.com/electionmodel/2004PreElectionNationalPolls.htm
This analysis will confirm that the 2004 national pre-election registered voter (RV ) polls were closer to the True Vote than likely voter (LV polls). It will also show that the LV polls, after adjustments, matched the RVs – and the unadjusted exit polls.
Real Clear Politics (RCP) is the 2004 data source of 133 national polls (102 were LV and 31 RV).
The final 15 RCP polls all sampled likely voters (LV).
Kerry led the pre-election polls in August, fell behind in September and rebounded in October.
When undecided voters were allocated, Kerry was projected to win.
Kerry did much better in the unadjusted RV average (45.0-45.1) than in LV (45.6-48.1).
Kerry did much better in the adjusted RV projection average (50.1-46.8) than in LV (49.0-49.2).
Kerry led by 47.3-46.9 in the Election Model final 18 poll average.
Kerry led by 50.9-48.1 in the Election Model projection.
Bush led by 49.0-47.0 in the RCP final 15 LV poll average.
Bush led by 49.8-49.3 in the 15 poll projection.
There are plausible reasons for Kerry doing better in the RV polls.
1- RV polls include newly registered voters.
2. The majority of new voters (57-61%) were Democrats.
3- A 17m net increase in the recorded vote (105.4m in 2000 to 122.3m in 2004).
4- Approximately 5m 2000 voters died and 5m did not vote (95% turnout).
5- Therefore, we can estimate that 95m 2000 voters returned to vote in 2004.
6- There were 125.7m total votes CAST in 2004 (3.4m uncounted).
7- Therefore, there were approximately 30.3m (125.7-95.4) new voters in 2004.
8- Kerry won approximately 18m (59.4%) new voters; Bush 12m.
The final 15 RCP poll national average had Bush leading by 49-47%.
The final Gallup poll was an identical 49-47%.
Gallup assigned 90% of undecideds to Kerry, so the final was tied at 49%.
First Key Confirmation:
Kerry and Bush split 47-47m returning (LV) voters; there were 1.3m Other voters.
Allocating 18m new voters to Kerry, he won by 65-59m (51.8-46.9%).
Kerry also had 51.8% in the unadjusted aggregate state exit poll (WPE method).
Comparing the True vote to the pre-election national polls:
The final 15 LV poll average, adjusted for undecided voters, was 49-49%.
But the LV polls did not sample newly registered (RV) voters.
The Election Model utilized 18 final national polls (9 RV and 9 LV).
Allocating 75% of undecideds to Kerry, he led by 50.9-48.1%.
Therefore, based on the 122.3m recorded vote, he won by 62.3-58.8m
Kerry's projected share was within 0.9% of the True vote (51.8%).
Second Key Election Model Confirmation:
But the projection did not include 3.4m uncounted votes (2.74% of 125.7 cast).
The base case assumption was that Kerry would capture 75% of undecideds.
(2.5m to Kerry, 0.9m to Bush).
The adjusted vote is Kerry 64.8-Bush 59.7m (51.6-47.5%).
That is within 0.2% of the calculated True vote (51.8%).
In summary, the final 15 pre-election LV poll average, adjusted for new (RV) and undecided voters, matched the Election Model projection.
TruthIsAll
http://www.geocities.com/electionmodel/2004PreElectionNationalPolls.htm
This analysis will confirm that the 2004 national pre-election registered voter (RV ) polls were closer to the True Vote than likely voter (LV polls). It will also show that the LV polls, after adjustments, matched the RVs – and the unadjusted exit polls.
Real Clear Politics (RCP) is the 2004 data source of 133 national polls (102 were LV and 31 RV).
The final 15 RCP polls all sampled likely voters (LV).
Kerry led the pre-election polls in August, fell behind in September and rebounded in October.
When undecided voters were allocated, Kerry was projected to win.
Kerry did much better in the unadjusted RV average (45.0-45.1) than in LV (45.6-48.1).
Kerry did much better in the adjusted RV projection average (50.1-46.8) than in LV (49.0-49.2).
Kerry led by 47.3-46.9 in the Election Model final 18 poll average.
Kerry led by 50.9-48.1 in the Election Model projection.
Bush led by 49.0-47.0 in the RCP final 15 LV poll average.
Bush led by 49.8-49.3 in the 15 poll projection.
There are plausible reasons for Kerry doing better in the RV polls.
1- RV polls include newly registered voters.
2. The majority of new voters (57-61%) were Democrats.
3- A 17m net increase in the recorded vote (105.4m in 2000 to 122.3m in 2004).
4- Approximately 5m 2000 voters died and 5m did not vote (95% turnout).
5- Therefore, we can estimate that 95m 2000 voters returned to vote in 2004.
6- There were 125.7m total votes CAST in 2004 (3.4m uncounted).
7- Therefore, there were approximately 30.3m (125.7-95.4) new voters in 2004.
8- Kerry won approximately 18m (59.4%) new voters; Bush 12m.
The final 15 RCP poll national average had Bush leading by 49-47%.
The final Gallup poll was an identical 49-47%.
Gallup assigned 90% of undecideds to Kerry, so the final was tied at 49%.
First Key Confirmation:
Kerry and Bush split 47-47m returning (LV) voters; there were 1.3m Other voters.
Allocating 18m new voters to Kerry, he won by 65-59m (51.8-46.9%).
Kerry also had 51.8% in the unadjusted aggregate state exit poll (WPE method).
Comparing the True vote to the pre-election national polls:
The final 15 LV poll average, adjusted for undecided voters, was 49-49%.
But the LV polls did not sample newly registered (RV) voters.
The Election Model utilized 18 final national polls (9 RV and 9 LV).
Allocating 75% of undecideds to Kerry, he led by 50.9-48.1%.
Therefore, based on the 122.3m recorded vote, he won by 62.3-58.8m
Kerry's projected share was within 0.9% of the True vote (51.8%).
Second Key Election Model Confirmation:
But the projection did not include 3.4m uncounted votes (2.74% of 125.7 cast).
The base case assumption was that Kerry would capture 75% of undecideds.
(2.5m to Kerry, 0.9m to Bush).
The adjusted vote is Kerry 64.8-Bush 59.7m (51.6-47.5%).
That is within 0.2% of the calculated True vote (51.8%).
In summary, the final 15 pre-election LV poll average, adjusted for new (RV) and undecided voters, matched the Election Model projection.