Midterm predictions: Sam Wang says Dems. have a 70% probability of holding the Senate

[telemann.livejournal.com]

Sam Wang of the Princeton Election Consortium, is not as well known as Nate Silver of Five Thirty Eight, who was the media darling in the 2008 and 2012 elections with his amazing predictions for the Presidential elections and midterms. Despite current predictions by the Washington Post, the New York Times, Larry Sabato, Sam Wang currently is predicting the Democrats will keep control of the Senate, albeit by a razor thin margin.

Senate Democrats are doing surprisingly well. Across the board, Democratic candidates in the nine states above are doing better in the polls-only estimate than the mainstream media models would predict. This is particularly true for Alaska, Arkansas, and North Carolina. In these three states, Democrats are outperforming the expectations of the data pundits (The Upshot’s Leo, Nate Silver, Harry Enten, John Sides, etc.).

Currently, the long-term forecast for Democratic control on Election Day is 65%, about 2-1 odds in favor.

I’ve been asked why the PEC Senate poll snapshot is more favorable to Democrats than forecasts you’ll find elsewhere: NYT’s The Upshot, Washington Post’s The Monkey Cage, ESPN’s FiveThirtyEight, and Daily Kos. All of these organizations show a higher probability of a Republican takeover than today’s PEC snapshot, which favors the Democrats with a 70% probability. Today I will show that in most cases, added assumptions (i.e. special sauce) have led the media organizations to different win probabilities – which I currently believe are wrong. I’ll then outline the subtle but important implications for a November prediction.

Dr. Wang explains his methods why everyone else is missing what he sees. Sam Wang's predictions were even more accurate than Nate Silver's predictions, getting every close Senate race right in 2012 when Silver missed 2 of them.

Dr. Wang also wrote a lengthy article explaining his modeling.

Comments

[hardblue.livejournal.com]

Wouldn't that shock Jeff!
He's pretty cocky on this.

[badlydrawnjeff.livejournal.com]

It would surprise me, but not shock me. Wang is quite good, but he's really the one out on a limb here.

[the-rukh.livejournal.com]

What do you believe he is missing in his calculations?

[badlydrawnjeff.livejournal.com]

He's made it quite clear that he doesn't believe in using fundamentals, but then notes that the fundamentals disappear as we get closer and the polling reflects it. So he, I believe, is making the same mistake many here have in just assuming the polling today will reflect the polling in the end.

A more important flaw is that he believes we have a lot of robust polling in this cycle, when polling hasn't been nearly as significant in terms of amounts of polls this cycle so far, especially in gubernatorial races. It might be enough for him to make his call for his purposes, but part of the reason averaging and such worked was because of the amount of polling done.

[the-rukh.livejournal.com]

What are fundamentals? What evidence do you have that polling is not "robust"?

[badlydrawnjeff.livejournal.com]

What are fundamentals?

Local quirks, PVI, national approval, Presidential approval, campaign ability, finances, etc.

What evidence do you have that polling is not "robust"?

I can't find the pieces I had read about it, but it's mostly attributed to the more local firms not putting out the money for polling like they used to. It's a value judgment in any regard - Wang feels he has enough polls to make a call, others do not see the polling as enough, and everyone agrees it's basically moot as we get near to the election anyway.

[the-rukh.livejournal.com]

Can you show the supporting math behind "local quirks" and the impact it has on a state to state basis?

Not to sound rude, but "fundamentals" sounds like an x-factor of the gaps. If someone comes up with a clearly supported reason for "Y" you just bring up any fuzzy thing so you don't have to agree.

Here's the real question though, have you taken in to account lunar phase during the time of polling? This is important.

[malasadas.livejournal.com]

A person making predictions based on chicken entrails may be correct -- may even be correct often.

But it wasn't because of the entrails that much is certain.

[badlydrawnjeff.livejournal.com]

Can you show the supporting math behind "local quirks" and the impact it has on a state to state basis?

Not in a way that would be of value here. This is the type of stuff that the pundits and prognosticators are doing, however, and it's why, say, Sabato has a different look than Cook, who has a different look than Silver.

Not to sound rude, but "fundamentals" sounds like an x-factor of the gaps. If someone comes up with a clearly supported reason for "Y" you just bring up any fuzzy thing so you don't have to agree.

That's fair if you're simply coming at it from that angle as opposed to understanding that election projections are about more than simply what the polls say on a random summer day. In reality, you should have a good reason to ignore the way that a lot of the local issues, finances, etc, should be discarded. Wang's theory is that Democrats are outperforming polls now, so they'll outperform later. That doesn't make a ton of sense overall, historically or otherwise, but his track record lends it more credence than anything else.

Here's the real question though, have you taken in to account lunar phase during the time of polling? This is important.

Well, it depends on whether werewolves wear pants when they transform on a full moon. Pantsless werewolves are disproportionately impacted by ID laws because they have nowhere to carry their IDs.

By the way, the piece I was thinking of regarding the polling? Turns out it was a FiveThirtyeight piece (http://fivethirtyeight.com/features/is-the-polling-industry-in-stasis-or-in-crisis/) (and here, too (http://fivethirtyeight.com/datalab/somebody-poll-a-senate-race/)).

[telemann.livejournal.com]

but his track record lends it more credence than anything else.


Oh absolutely.

Longtime readers of PEC will not be surprised to know that I think the media organizations are making a mistake (i.e. relying or weighing "fundamentals" more heavily). It is nearly Labor Day. By now, we have tons of polling data. Even the stalest poll is a more direct measurement of opinion than an indirect fundamentals-based measure. I demonstrated this point in 2012, when I used polls only to forecast the Presidency and all close Senate races. That year I made no errors in Senate seats, including Montana (Jon Tester) and North Dakota (Heidi Heitkamp), which FiveThirtyEight got wrong.

[yes-justice.livejournal.com]

But is he more shocky and cocky than Dr. Wang?

[hardblue.livejournal.com]

Well, I take it he is playing in the big leagues. I suspect Jeff is a pro at this, too, but it's not the same league.

[malasadas.livejournal.com]

I suspect Jeff is a pro at this



I was unaware that "unqualified yet insufferably smug hobbyist" and "pro" were synonyms.

[hardblue.livejournal.com]

Are you sure that he is strictly amateur? I got the impression he works for Republicans, and I've imagined this is one of the reasons he is busy these days, on account of the elections.

[malasadas.livejournal.com]

Yes, amateur and well beyond the confines of any field of competency.

[badlydrawnjeff.livejournal.com]

I do not work for Republicans, but I have been offered paid positions in the past. I'm not a "pro" by the strict definition, but I'm far from an "amateur smug hobbyist." Attempts at character assassination are par for the course here.

I'm busy these days because of my real life, not because of the elections. I haven't been able to spend nearly enough time on the campaign trail this year, sadly (but probably happily for the peanut gallery).

[oslo.livejournal.com]

I do not work for Republicans, but I have been offered paid positions in the past.

Paid positions to do what, exactly? And on what past experiences were these offers predicated?

If I recall correctly, your opportunities in this area didn't have anything to do with devising political strategy for national or local campaigns, based on poll data or "fundamentals." Weren't they opportunities made available to you based on your volunteering for campaigns?

[lafinjack.livejournal.com]

No, I wish you had all the time in the world to volunteer to them.

[farchivist.livejournal.com]

Are you sure that he is strictly amateur?

Yes, I am absolutely and positively sure he is strictly amateur.
His "experience" is limited to volunteering for a couple of local campaigns.
His job has nothing to do with politics. His lack of time is because of his kid and caring for his Alzheimer's-ridden mother.

[telemann.livejournal.com]

Ok, if we're going for Double entendres :P

[yes-justice.livejournal.com]

HA!

[usekh.livejournal.com]

Well now that Jeff has opined against him I have greater confidence this pollster is right.

[malasadas.livejournal.com]

http://www.hiyoooo.com/

[telemann.livejournal.com]

lol

[yes-justice.livejournal.com]

Steer into the slide