Reviewing Our Predictions for Election 2014

Last month, I posted my predictions for the 2014 midterm elections. With the final Senate seat being decided in a runoff election this past weekend, it seems like a good time to revisit those predictions to see how we did.

The general consensus was that we'd see a strong Republican showing. I went a little further out on a limb, projecting the GOP to outperform state-by-state polling data. Reality was even more of Republican win.

Pre-Prediction Prediction
While this doesn't necessarily count as an election prediction, I wrote "I know Mitch McConnell promised that a Republican majority in the Senate will kill Obamacare. I've heard that line before. The GOP promises big things before an election. After the election, there's no tough decision that can't be delayed indefinitely."

Senate Majority Elect Mitch McConnell didn't even wait for the new Congress to be sworn in before backing off of that promise, saying that our hopes for repeal come from the Supreme Court "as opposed to actually getting the president to sign a full repeal, which is not likely to happen..."

Shocker.

Prediction #1: Walker: 51%, Burke: 49%
Result: Walker 52%, Burke, 47%

Well, I was right in predicting the winner, but off a few percent as Walker pulled out nearly a six point victory. It wouldn't be the only time the GOP outperformed the polls on election night.

Prediction #2: "I'll also predict that the GOP pulls off the attorney general race and narrowly expands the majority in the Legislature."
Result: The Republicans picked up two seats in the Senate and three in the Assembly.

We nailed this one.

Prediction #3: "...the Republicans pick up an even 10 seats in the House of Representatives."
Result: The Republicans picked up a dozen seats in the House for their biggest advantage since the Truman Administration. There is the opportunity to claim a 13th seat pending a recall election in Arizona.

Not bad, not bad!

Prediction #4: "...the GOP wins Louisiana, Georgia, Colorado, Iowa, Alaska and Kansas for a 53 seat majority in the next Congress."
Result: The GOP picked up all those seats, plus North Carolina, to win a 54 seat majority.

Perhaps the most interesting part of my prediction was this:
Those eight seats are as follows (ranked in order of how likely I think the Republicans are to win):
  1. Louisiana
  2. Georgia
  3. Colorado
  4. Iowa
  5. Alaska
  6. Kansas
  7. North Carolina
  8. New Hampshire
....If my estimation is correct, we will be able to draw a horizontal line somewhere on this list, and everything about will go red; everything below will go blue. 
It turns out that this was almost exactly correct. One through seven went red (with a nail-biter putting North Carolina in the GOP column), and New Hampshire went blue in a closer than expected race. Missing from this race was Virginia--which I did not expect to be close. The GOP nearly pulled off an upset here.

Rationale #1:  "Well, that's not a very bold prediction, Karlen. You're just telling us what the polls are telling us!"

Yes, my analysis was heavily based on polling data, which I have long (correctly) argued is incredibly reliable on predicting elections. While my poll-based projections turned out to be quite accurate (as you'll see above), I think one of the most interesting aspects of this election is that the polls were way off in the magnitude of the GOP landslide.

For example, the polls and I predicted a two point Walker win. We look correct for getting the winner right, even though the final margin was nearly four points greater than the projection. If we miss by four points the other way, Mary Burke is the governor-elect, and we look like idiots. We might have gotten lucky.

So has polling broken down? Are response trends now too unpredictable that there's no way to use public opinion surveys to accurately gauge elections? That, my friends, is the million billion dollar question that we'll need at least another two years to answer.

Rationale #2: "One party usually takes home 70-80% of the tossup seats....That means that either the Democrats sandbag their losses, take a majority of the tossups and save their Senate majority or the GOP heads into 2015 with a stronger-than-expected majority."

If there's one thing I got right, it's the paragraph above.

Exit Prediction #1: Hillary will not be the next president of the United States.
Exit Prediction #2: Right now you're quietly cursing me for this subtle reminder that we've got two years of NONSTOP speculation, advertising and chatter about the 2016 presidential election.

No comments: