Wednesday, January 20, 2016

DATA ANALYSES OF THE GOP RACE WE NEVER FINISHED READING

Okay, David Wasserman of FiveThirtyEight, explain to me why everything I think I know about the Republican presidential race is wrong:
In a few weeks’ time, it’s possible that Donald Trump and Ted Cruz will steamroll their way through Iowa, New Hampshire and South Carolina and dominate the so-called “SEC Primary” -- the collection of 13 mostly Southern states that will vote on March 1 -- horrifying many GOP elected officials and depriving any other candidate of a night to celebrate.

Yet even if that happens, it’s still possible that Marco Rubio (or another more establishment-friendly candidate) could end up with the nomination, thanks to quirks of the GOP’s complex delegate math.
Really? Tell me more.
The GOP’s primary calendar is surprisingly front-loaded with states friendly to insurgents like Trump and Cruz. But because of Republican National Committee rules, all but one of these states will award their delegates on a proportional basis, intentionally making it difficult for any one candidate to build a durable or commanding lead.

Instead, Florida and Ohio, which tend to support more conventional Republicans, are likelier to shape the race’s destiny than Iowa or South Carolina. That’s because they will award a whopping 99 and 72 delegates, respectively, in huge winner-take-all primaries on March 15.
Florida, "which tend[s] to support more conventional Republicans"?

Have you seen the latest polling from Florida, David?



In this poll, from Florida Atlantic University, Donald Trump is leading by 32 points. He and runner-up Ted Cruz combine for 64% of the vote. According to the Real Clear Politics average of recent polls, Trump leads Florida by 19.3 points. (Click to enlarge.)



And Wasserman thinks that lead is going to shrink if Trump comes into Florida looking like a winner after victories in February and early-March contests?

There hasn't been much polling in Ohio this year, but Trump was leading there in the last poll, well ahead of local boy John Kasich, who was in third place.

Clap louder, data journalists! Pay no attention to the actual polls!

5 comments:

Victor said...

One cannot see, that which one paid not to see.

mlbxxxxxx said...

Maybe you need to wonder why "data journalists" don't seem to trust the data. 538's been a pretty reliable organization with a solid empirical bent. They do not seem to have a lot of faith in the numbers this cycle. Their inclination this year -- unlike, say, 2012 when their analysis was very poll driven -- to question the polls gives me pause.

That said, I've got my fingers crossed for Trump to sweep the primaries. He'd be the easiest to beat (though I think you disagree -- some polls, e.g., those that show him hated by more and more minorities, you don't seem to credit as much) and even if he won, he's clearly the least conservative of the bunch. You can't ask much more from a GOP president than that he not be a died-in-the-wool conservative.

Steve M. said...

No, I agree that he's the most beatable. I just think it's far from a certainty that he'd lose.

Unknown said...

We'll know a great deal after Iowa. If Trump can deliver turnout commensurate with his poll numbers, it's all over but the shouting. If not, we'll know that he's a paper tiger.

Yastreblyansky said...

Apropos of Florida, something tells me all the Cubans under 65 are going to be voting Democrat this year. It's amazing that Bush and Rubio both have no legs in the state, you'd think some Republicans would have some affection for at least one of them.