Presidential election forecasts :whoami

whoami

My name is Andrea Moro and I am an Associate professor of Economics at Vanderbilt University. More information about myself, including my professional cv, can be found on my personal page: http://www.andreamoro.net.

I have started this project as a hobby for the 2004 elections, when I was at the University of Minnesota. The project obtained a decent success, and many have asked to repeat it this year.

Please do not email unless necessary. I will do my best to find the time to clarifiy the methodology and procedure on the appropriate page.

If you are Italian, and do not yet know about it, please visit noisefromamerika.org, a group blog (written in Italian) I edit together with 5 Italian economists about italian politics and economics.

Thanks for visiting.

On the first page of the site, Obama's unconstrained likelihood of winning is given as 98.2%. His likelihood of winning if McCain wins Florida is also given as 98.2%. This would seem like it can only be true if McCain has a 100% chance of winning Florida, which should not be the case given your methodology. Is there a difference between the two figures which is being rounded away? Is there something else that I am missing?

Dear Andrea, You postulate "what if" scenarioss. Curiously, in these permutations you do not include Colorado and New Hampshire. Potentially, Co and NH have the same pivotal role as Ohio in 2004 and Florida in 2000. So, why not include these two states. with Obama building a lead in Co, NH could decide the election?

Andrea, just a suggestion. You should add to your frontpage probabilities what would be Obama's winning pct. in case he loses New Hampshire (and possibly also that of other combinations which include NH). Looking e.g. at RealClearPolitics one can easily check that in the "no toss up" count of electoral votes, which has been stuck at Obama 273 - McCain 265 for at least a month, shifting NH (= 4 EVs) from Obama to McCain would cause a draw at 269 (Nancy Pelosi would then decice!). Note that the polls give NH alternatively to one or the other, depending on the source.

Dear Andrea, Tou have posted Colorado 47-49% to Obama. However, OB holds a definite, possibly slightly statistically significant lead over McCain. No other pollster, simulation, web-site has this can you please explain why?

Dear Andrea, Looking at your electoral map, it appears to be less sensitive to state polls than many of your other competitors? Have you compared your poll samples with the other sites? Do you compare day-to-day projected Obama EV? Perhaps, this would be of interest for not only polling but also for methodological issues. After the election, it will be interesting to work out which forecaster was closest to the outcome and why?

The Bradley effect refers to the fact that in the past 6-7% fewer voters actually voted for a Black candidate than those who told the survey that they would do so. Is there any research to indicate that this factor will be weaker now than in the past?

"But if obama loses ...", Dear Andrea, I understand that the ca. 22% prob of an Obama victory in the event of defeat in several key swing states is generated from the same simulation. But, looking at the electoral map is there a one in five probability that OB will construct 270 votes out of the next most winnable swing states:ND, WV, MT, GA? Isn't that prob closer to 10%. If a scenario occurred affecting the swing states you postulated this could infer a shift in national not state voting. Assuming this, the prob. of OB getting to 270 votes via ND et al is 0%. Best wishes.

Dear AM, Your final forecast was "out" significantly. Indeed, much more than Prof.Sam Wang(notwithstanding Sam's very late, personal intervention!), 538, Pollster et al. I accept that your methodology is statistically sound but do you have any explanations? Yours, Nicholas J. Alcock.

Though I don't at all like politics and am already sick and tired of all these political issues (constantly stubmble over only political reviews even at rapidshare se ), but your explanations are easy to understand and I even read through your articles from time to time to get the slightest inkling of what is going on. Thank you, Andrea!

am already sick and tired of all these political issues (constantly stubmble over only political reviews even at Torrents search engine ), but your explanations are easy to understand and I even read through your articles from time to time to get the slightest inkling of what is going on.

of all these political issues (constantly stubmble over only political reviews even at Rapidshare Files ), but your explanations are easy to understand and I even read through your articles from time to time to get the slightest inkling of what is going on.

to the fact that in the past 6-7% fewer voters actually voted for a Black candidate than those who told the survey that they would do so. Is there any research to indicate that this factor will be weaker now than in the past? Download movies

Documents

Post a new comment