WaPo's "Election Lab 2014" Senate Ratings

[njwes]

Does anyone know offhand why the Washington Post's Election Lab 2014 feature has had such wildly different ratings than all the other prediction models this cycle? Are they using a model that's markedly different than the other half-dozen other "quantitative" forecasts?

Don't get me wrong, I really do appreciate that they're really going balls to the wall and claiming a 90+% chance of Republican takeover, and haven't been boringly wandering around between 40-60% for the last several months like those other milquetoast forecasts, but it seems like they could at least include a note of explanation somewhere.

[Torie]

Because  its metrics weight current poll numbers less, and underlying fundamentals more. In other words, the model "predicts" where the polls will end up rather than the current numbers, more than other forecasters do.

[njwes]

Because  its metrics weight current poll numbers less, and underlying fundamentals more. In other words, the model "predicts" where the polls will end up rather than the current numbers, more than other forecasters do.

Thank you, that makes sense. Though the forecast also seems to have the quirk of moving VERY sharply between extremes in the individual Senate races--e.g., Iowa and Colorado careening between parties at 80+%. North Carolina and Alaska have also seen huge and fast swings, though not as extreme. Not sure why that should happen in a model that emphasizes fundamentals.

That said, I guess you can see the effect of underlying factors in, say, Kansas, which I don't think has ever fallen below 85 or 90% Republican even in the last couple months.

[IceSpear]

Those percentages gave me a good laugh, thanks.

[Princess Nyan Cat]

Their methodology has some odd quirks, especially in House race ratings. I've been watching them closely, and it's not unusual for a seat to jump from 75% Republican to 85% Democrat (or vice versa) overnight.

Nevertheless, all the various predictions I follow are generally converging to the same results.

[Bigby]

Their methodology is weird, but it sure makes me happy. I think it's different due to how they calculate everything.

[Adam Griffin]

They have had a horrible right-wing bias throughout the entire cycle, but that is giving way to a more bi-partisan ridiculousness as election time draws nearer. They've been more willing than others throughout thick and thin in this cycle to classify a variety of races as one candidate having 90%+ chance of victory, in an abnormally fluid year. Certainly, many races would solidify in the run-up to the election, but 99% chance for Shaheen and Rounds? A 96% chance for Cassidy and Hagan? A 95% chance for Roberts? 87% for Ernst and 83% for Gardner? These projections don't seem to be based in reality. According to them, Nunn has a better chance of winning than Braley, Udall or even Orman.

Because  its metrics weight current poll numbers less, and underlying fundamentals more. In other words, the model "predicts" where the polls will end up rather than the current numbers, more than other forecasters do.

I'm wondering if they've just threw that out, because as outlined above, their rankings have shifted by very large amounts very rapidly. The only way that should happen given a consistent formula is if they were weighing polls very heavily, and not the opposite.

[Eraserhead]

Junk Senate ratings!

[○∙◄☻¥tπ[╪AV┼cVê└]

SD is still > 99% Republican? Toss it.

[dmmidmi]

This is the election equivalent of HOT SPORTS TAKES. Gut-based opinions that are meant to stir a reaction.

Washington Post is garbage, and so is this map.

[windjammer]

That was WaPO who had Michigan as a pick up not a long time ago

[KCDem]

Hahaha this model is a joke.

[njwes]

Here's the explanation, hadn't seen this article:

http://www.washingtonpost.com/blogs/the-fix/wp/2014/10/10/the-washington-posts-election-model-says-republicans-have-a-95-chance-of-winning-the-senate-um-what/

"At this point, we're almost 100 percent weighted toward the polls, except in races where the polling is extremely light."

That would explain the insane swings.