Poll analyses are back! Yes…after about four years of rest, I’ve decided to get back into the poll analysis business. If you are unfamiliar with my analyses, take a look at the FAQ. This is a shakedown run with a new HA template, so feel free to mention quirks or problems in the comment thread.
I’ve started with a match-up between Hillary Clinton and Jeb Bush. I know, I know, it should be a match-up against Donald Trump instead of Bush. The problem is, there are a total of 20 state head-to-head polls available that match-up Hillary and Donald. For Clinton–Bush match-ups at the state level, there are 212 polls. Apparently, pollsters haven’t quite gotten serious about Trump. I expect that most state head-to-head polls will start including him, however.
For the same reason—lack of state head-to-head polling—I cannot really provide analyses for Sanders, O’Malley, Webb, Biden, Warren, Warner, Cuomo, Schweitzer, or any other Democrat (for now). I will, however, do Republicans other than Jeb Bush, because the polling exists.
After 100,000 simulated elections, Hillary Clinton wins 56,003 times and Bush wins 43,997 times (including the 2,369 ties, which go to the Republican for 2016). Clinton received (on average) 273 to Bush’s 265 electoral votes. In an election held now, Clinton would be expected to win with a 56.0% probability and Bush with a 44.0% probability. That is a close result…damn near a tie.
I remind you that these analyses are driven by state head-to-head polls, not national polls. The state-wide polls are aggregated in a way that emulates the electoral college.
A peek at the long term trend is given after the jump….
The long term trends in this race can be seen from a series of elections simulated every seven days using polls from 31 Aug 2014 to 31 Aug 2015, and including polls from the preceding month (FAQ).
Clearly, until this month, Clinton has maintained a solid lead over Jeb. This is, perhaps, why the media portrays some element of panic among the Democratic establishment (like the largely unfounded rumor that Joe Biden is considering jumping into the race). Is this a reflection of “server-gate”? Or do Democrats spend more of their summer away from phones in national parks and forests? Share your thoughts in the comment thread. (Please remember…the comment threads for these analyses are not “open threads”, so please stay on topic.)
Here is the distribution of electoral votes [FAQ] from the simulations:
An animated sequence of maps and electoral vote distributions over time can be seen here
The most likely outcome of an election held today would be a Clinton victory with 274 electoral votes. Of the ten most probable outcomes, Clinton loses in four of them (269 or fewer electoral votes):
- 274 electoral votes with a 2.86% probability
- 273 electoral votes with a 2.38% probability
- 269 electoral votes with a 2.37% probability
- 284 electoral votes with a 2.21% probability
- 266 electoral votes with a 2.16% probability
- 261 electoral votes with a 2.12% probability
- 272 electoral votes with a 2.09% probability
- 285 electoral votes with a 2.06% probability
- 289 electoral votes with a 2.01% probability
- 256 electoral votes with a 2.00% probability
After 100,000 simulations:
- Clinton wins 56.0%, Bush wins 44.0%.
- Average (SE) EC votes for Clinton: 273.2 (22.1)
- Average (SE) EC votes for Bush: 264.8 (22.1)
- Median (95% CI) EC votes for Clinton: 273 (231, 319)
- Median (95% CI) EC votes for Bush: 265 (219, 307)
Each column of this table shows the electoral vote total aggregated by different criteria for the probability of winning a state (Safe=100%, Strong=90%+, Leans=60%+, Weak=50%+):
|Threshold||Safe||+ Strong||+ Leans||+ Weak|
This table summarizes results by state. Click on the poll count to see the individual polls included for the state.
|2||8||Votes||polls||Votes||Clinton||Bush||% wins||% wins|
* An older poll was used (i.e. no recent polls exist).
Details of the methods are given in the FAQ.
The most recent analysis in this match-up can be found from this page.