Super Post-Tuesday Mk. II Wrap-Up: Where I Went Wrong

Here we are past midnight, Wednesday, June 4, and we've one race yet to be decided. As of 1:35 AM in Mississippi, with 98% of precincts reporting, state senator Chris McDaniel leads incumbent Sen. Thad Cochran 49.6-48.8: a difference of 2,528 votes. Unfortunately, we probably won't know the exact results until tomorrow morning. Here's Emily Wagster Pettus of the Associated Press in Jackson:

Zero precincts reporting from Covington County. An election commissioner just told me she's home in bed & can't give us numbers. #MSSen
— Emily Wagster Pettus (@EWagsterPettus) June 4, 2014

At this point I think it's safe to say that Mississippi's going to a runoff.

When I predicted Mississippi, I made two claims: 1) that Sen. Cochran would come out ahead of McDaniel, and 2) that the election would proceed to a runoff. If what people are saying about the Mississippi results is right, then I'm right on the second claim and wrong on the first. With regard to the first claim, I did say that it was too close to call, and that I slightly favored Cochran--it's just that in Mississippi polls were not that great, with a lot of variation in how various pollsters and handicappers thought the composition of the electorate would turn out (especially in the median age of the voters). This is a fairly important variable: with Sen. Cochran running for a seventh term, he likely appealed to older voters. Turnout by age could've made the difference here, and we won't really know until exit polls get released (if they do at all).

The only other place I went wrong was in Iowa. Polls for about a month before the election consistently showed Ernst ahead, but with 30, 35% of the vote--the weighted average came out just above 35%, which is why I said that Ernst would avoid a convention but would not win a majority. (That's what happens when I get too detailed with estimates.) I reasoned that the undecided voters--which made up as many as 30% of the respondents in one poll--would break more or less proportionately, with maybe a little more going to Ernst, which would bring her into the low- to mid-40s. Here's what happened on election night:

Ernst picks up almost 21 points compared to polling, while candidates like Jacobs and Schaben come out to pretty much what the polls said. Aaron Blake of The Washington Post writes that "Ernst surged late in the race and got both establishment and tea party support, moving past the erstwhile frontrunner, businessman Mark Jacobs." This is the sort of thing polls don't pick up, like when "Dewey Defeats Truman" in 1948 (although at least in this case the polls got the overall result right). 

Maybe it isn't as much of a "surge" as Mr. Blake says, since Ernst had been gaining momentum since about February, largely drawing undecided voters while Jacobs sorta flatlined and Clovis bounced around an average of about 10%:

But hey, that's part of the learning experience here. Hope you'll read more on Super Tuesday Mk. III (June 10), which should work better than this time around.

CORRECTION: The post previously said that Super Tuesday Mk. III was on June 24. Turns out there's a bunch more primaries on June 10.