AdamT
DP Veteran
- Joined
- Jul 26, 2011
- Messages
- 17,773
- Reaction score
- 5,746
- Gender
- Male
- Political Leaning
- Undisclosed
It's generally accepted that aggregate polls are more reliable than any individual polls when it comes to predicting elections. By combining the polls in various ways aggregators minimize the effect of outliers and try to achieve something like a consensus view of the race, from a statistical standpoint. I know of six different aggregator sites following the elections and it occurred to me that it might be interesting to aggregate the aggregators. Perhaps, similar to individual polls, this kind of averaging will produce an even more accurate result? Don't know, but here is the result as of 1:20 p.m. on November 5:
------------------Obama-------Romney-----------------
Electoral-vote.com: 294 - 220
Princeton Election Consortium: 303 - 235
Votamatic: 332 - 206
Pollster.com: 281 - 191
FiveThirtyEight: 307 - 231
RealClearPolitics*: 303 - 235
AVERAGE Prediction: Obama 303 - Romney 219
Note that some of the sites don't include states that they deemed statistically tied.
*The RCP number is from their map with no toss ups.
------------------Obama-------Romney-----------------
Electoral-vote.com: 294 - 220
Princeton Election Consortium: 303 - 235
Votamatic: 332 - 206
Pollster.com: 281 - 191
FiveThirtyEight: 307 - 231
RealClearPolitics*: 303 - 235
AVERAGE Prediction: Obama 303 - Romney 219
Note that some of the sites don't include states that they deemed statistically tied.
*The RCP number is from their map with no toss ups.