Discussion for article #228320
Nate sure does get into a lot of verbal slap fights. Politico, PPP, and now Wang. Its almost as if he wants media attention or somethingâŚ
I have a feeling that Nate fails the âguy you want to have a beer withâ test pretty handily.
30 Likes
Indeed. The more I see of Nat the less I like him.
12 Likes
Can these folks simply do their jobs without becoming primadonnas?
18 Likes
I clicked on this only to ask;
What would happen if we simply waited until the elections?
Why do we engage in the predictions at all?
Could the predictions themselves affect the results?
(Either way; "They say my guy is going to win, so I donât have to bother / They say my guy is going to lose, so why bother?)
Again, why not just wait and see?
6 Likes
I clicked on PEC for the first time and suffered stimulus overload from all the charts and data.
Wish I had a Cliffs Notes version.
1 Like
It is useful to do predictions for a candidate so they can see if their campaigning strategy is working or not. If they are still down in the predictions they know they need to change strategies. Sometimes that works well. Other times we have flailing candidates picking Sarah Palin as a running mate.
There are other political reasons to do predictions as well. If the Massachusetts election that resulted in Sen. Scott Brown had been predicted in advance, maybe the Senate wouldnât have waited for the election to push through the ACA, and we could have gotten something better than what we got.
Yes, sometimes the predictions do affect the results. If the prediction is very one sided, then the exact impact you described happens. If the prediction is close, the opposite happens where both sides are motivated to go vote.
If we are going to be hit with another tea party wave, it is helpful to know in advance so we can get done what we can get done, and so we can try to preemptively staunch the bleeding. That is why we donât âjust wait and seeâ.
11 Likes
This sounds like nothing more than jealousy on Nateâs part. From my perspective, Wongâs predictions are more accurate than Silvers.
4 Likes
I think itâs pretty clear that whatâs eating Nate is the fact that Sam is writing for the New Yorker and has much more impressive academic credentials. An obvious threat.
Iâve read enough of Wangâs blog to have a rudimentary grasp of his approach and the ways it differs from Silverâs. I certainly donât have enough knowledge to evaluate its accuracy on that basis alone, but I know that he has a very good record going all the way back to 2004. As the ability to make accurate predictions is the only important criterion here, the idea that his model is âflawed in profound waysâ is obviously stupid.
The more interesting story in all of this to me is the way that, as far as conventional âwisdomâ is concerned, 2014 is the inverse of 2012. In 2012 all the legitimate pollsters had Obama winning handily but the official story was that it was too close to call (leaving room for the Fox pundits to forecast a Romney landslide). This year though the conventional wisdom is that Republicans are overwhelming favorites to take back the Senate. But the Wang vs. Silver debate aside, the fact is that at this point it is pretty much too close to call. Most of the meta-analysts give the Republicans something like a 55% chance of winning. Given that the election is not going to be run 100 times in a row, that is saying very little. You donât have to be a statistical whiz to realize that is not very different from a coin toss.
21 Likes
My take on this is that these guys differ on how subjectivity should be factored into making predictions based on polling data. For example, take an extreme comparison like Joe Scarbourghâs âgut feelingâ after a night on the town predictions versus a strict averaging of all pure polling data on the same subject. What is missing from the former is any objectivity whatsoever, while the latter ignores the political reality that elections are won based on who actually get off their butts and vote as opposed to simply answering questions from a pollster. What these guys are feuding over is how that subjectivity is factored into their equations. Wang tends to be more liberal and Nate is more conservative and each can write a book length justification for their views. Why is Nate so crotchety on the subject? Because thatâs who he is. Based on his performance to date, he deserves respect and expects it.
7 Likes
Stop it, boys.
Youâre both pretty.
23 Likes
Silverâs book was deeply flawed and he knows itâŚ
2 Likes
That is essentially correct. Wangâs models are driven by polling data, while Silver tries to quantify and introduce some of the âgut feelingsâ. Personally, I do find a certain irony that Silver is defending his âspecial sauceâ by attacking a guy that is strictly data driven, considering his earlier arguments against rest of the pundits was that his predictions are data driven and theirs werenât. I do understand that Silver does go to lengths to quantify non polling data factors, but I lean more towards Wangâs approach. My thought is that all the âspecial sauceâ is already baked into the polling data, and considering the special sauce guysâ, Silver included, models are start to converge with the actual polling data, it seems to be the right approach.
14 Likes
I agree with you.
And add, everything will be finally decided by which side can motivate the most people to get off their butts and vote.
5 Likes
Silverâs 2012 call was easyâŚHis polling method will blow up one day the layman wonât understand and Silver will sufferâŚhe knows his stardom is numbered.
1 Like
I suspect those 55% predictions are going to be shifting downwards this week, too. Kansas is turning into a big fly in the ointment. The GOP is looking better in IA, but thatâs the only place they have shown any gain. Every other race seems either pretty flat, or a slight Dem bump so far.
As a slight add on to the part about "conventional wisdom, what I find interesting is looking at Silverâs summer predictions. It was all conventional wisdom, and Silver bought into it. Now that the conventional wisdom is turning out to be questionable, Silver is going on the warpath in a pre emptive defense of his âspecial sauceâ.
9 Likes
Wasnât easy in 2008 and he missed by two electoral votes. It also wasnât easy to go up against the huge âtalking headsâ wisdom that is based on party talking points expressed by folks with massive egos and lots to loseâŚKarl Rove comes to mind as well as a dozen more of the GOTPâs âassets in place.â
If Wang is right again. then heâs the man. Itâs clear he has fans here for obvious non-technical reasons.
1 Like
Nateâs tweets look like something a mean high school girl would doâŚor a Republican. Weird.
9 Likes
In addition to Blueberryâs response, for political scientists, one of the big problems is there is very little actual science because there is very little data and tests are extremely hard to do for obvious reasons.
So polling data over the length of the race is incredibly important in studying what works and doesnât on campaigns. How effective are direct mailings vs. TV spots? What sort of effect did the candidateâs appearance make in peopleâs interest, and how long did the stay interested?
All of that is important not only for the current campaign, but in determining future approaches and strategies.
5 Likes