How Nate Silver Missed Donald Trump

[Slander and/or Libel]

I would be interested in seeing how well the "endorsement points" correlated with polling in the runup to the primaries, historically. It strikes me that Silver may be overweighting endorsements when in the past, they've tended to track which candidates were doing well in the polls anyway.

[Slander and/or Libel]

@Figs: Here ya go.

That's interesting and goes part of the way toward what I was wondering, but it also doesn't necessarily give me an indication of whether endorsements as a variable is separable from polling support. He sketches a very, very rough idea of how endorsements might translate to votes, but I still don't have a good handle on exactly how much that effect might be captured by, or duplicating, other data.

I'm struck in particular by the Democratic charts for 2004. Kerry's endorsements spiked after he started succeeding in primaries, before which point they were lagging Dean and even Gephardt.

[Slander and/or Libel]

I don't disagree, and I think he was in a tough place. If he admitted flexibility, then it would cut against the perception that he was trying to cultivate that the numbers tell us way more about the race than anything else (and that his specific interpretation of the numbers was as close to correct as anything). But if he stayed rigid, he ran the risk of a hard-to-predict event biting him on the butt, which is what appears to have happened. I don't envy the position he's found himself in, and I don't think it's all his fault, though he probably could have handled it better.

[Slander and/or Libel]

I know it'd be next to impossible to quantify, but it'd be nice if Silver could come up with some rough way to assign a number to the value of the endorsement of a particular leader. Maybe in general, or on average, the 1 point for representatives, 5 points for senators, 10 points for governors thing works out. But what about a really prominent representative? Wouldn't that endorsement carry more weight than that of a backbench freshman senator? All this, of course, is to say nothing of the potential endorsements of former office holders, media figures, and other party bigwigs. Would Mitt Romney's endorsement mean something? My intuition says yes, but Silver's model gives no way to quantify that. Would Rush Limbaugh's endorsement mean something? Again, my intuition says yes, more than the endorsement of a first term Republican congressman, but there's no way to capture that here.

I think one of the biggest methodology problems, though, is that the model lumps all Republican primary voters into one basket when there seems pretty clearly to be an establishment/insurgent split. That means that maybe an endorsement from Mitt Romney would affect, say, Marco Rubio in a different way from how it would affect Trump or Cruz (not saying they'd get Romney's endorsement, but just for sake of argument). It's unclear to me how to quantify this split, but it's pretty clear to me that it's real, and that a model that doesn't attempt to take it into account is not really reflecting reality.