PDA

View Full Version : FiveThirtyEight continually downgrading B.J.'s chances of winning




Romantarchist
10-01-2010, 05:02 PM
A couple weeks ago, FiveThirtyEight declared that B.J. Lawson had a 31% chance of winning his race. Then a couple weeks passed by and they downgraded his chances to 6% (!!!). Now, they downgraded it even further to 4%!

Does anyone actually believe this? Other election predictors are raising B.J.'s chances of winning right now but FiveThirtyEight keeps lowering him further and further. This is especially outrageous in the face of these facts:

-the Republican Party is favored by American voters this year
-B.J. Lawson has libertarian positions that will have wider appeal in NC-4 than a neocon
-Republicans won NC-4 in 1994

Why is FiveThirtyEight doing this? Bribery from the Price campaign? Incompetence? It demands an explanation.

BenIsForRon
10-01-2010, 05:12 PM
How many people in the 4th district do you think look at FiveThirtyEight. .00001%? .000001%?

Epic
10-01-2010, 05:45 PM
Yeah I don't understand this. Today they just moved BJ's chances down to 4%, which makes no sense, cause there hasn't been any news in the race at all since BJ released an internal poll showing him slightly ahead.

Imperial
10-01-2010, 08:01 PM
I think the reason is that the pollster for BJ's internal is an extremely partisan one. It isn't used for their forecasts, and some other prominent pollsters derided the result.

Jeremy
10-01-2010, 08:20 PM
Inconsistency.

wormyguy
10-01-2010, 09:03 PM
FiveThirtyEight uses a computerized statistical model. Older polls, and internal polls, are not weighted as heavily as newer or independent polls. Given that, the model is probably predicting almost solely based on the district mean, with some degree of uncertainty due to Lawson's internal.

TCE
10-01-2010, 11:02 PM
Nate Silver is pretty trustworthy. I'm sure he knows what he's doing. That internal poll wasn't reality and it is likely B.J. was truly down 8-10 points in that poll. Sure, that's pretty amazing considering the Cook Rating of that district and the incumbency effect, but not nearly good enough to win. Check out some of his postings for methodology. He just posted 2 parts in a 3 part series on what factors he uses, very interesting. Statistically, unless that last poll can be verified, it is statistically highly unlikely B.J. can take this race.

boneyard bill
10-15-2010, 12:34 AM
I can think of two explanations:

1. Silver is a liberal Democrat (which I believe he is known to be), and he is trying to hurt B.J.'s fund-raising abilities in a race that is getting close. Or

2. The mere passage of time reduces B.J.'s chances of winning without a major movement in the polls. So if you discount B.J.'s internal poll and have no other evidence of progress, the chances of victory become less. You have three weeks to close the same gap as when you had six weeks left.