April 4, 2016
We’re less than 24 hours from the start of live voting in Wisconsin, and Hillary Clinton is within 3 points in the Real Clear Politics average. This doesn’t mean she’s going to win. Bernie Sanders is ahead in 5 of the 6 most recent surveys. The one prior to that (Emerson) favored her, but their recently released update has him up 8.
The one poll from the past week to favor Hillary is from Loras College. They surveyed Iowa regularly in the weeks and months before the caucus. As a rule, Clinton did better in their polls than almost anyone else. They were wrong. When Loras shows Hillary ahead by 6, that’s good news for Bernie.
Sanders is poised for an actual win, Clinton is on the brink of a de facto victory.
There are any number of guideposts for this contest. The simple one is who wins. That person gets their face on the TV with the big check mark next to their name. It’s fun. You give a victory speech and everything.
Based on the current numbers, Bernie is likely to get to do this. The next level is matching Barack Obama’s 2008 performance, when he grabbed 58% of the vote against Hillary in Wisconsin.
Until the recent Pacific Saturday caucuses, Sanders trailed Obama’s numbers in virtually every state that has voted. The big exceptions were in New England. Many of the remaining states were Clinton victories in ’08, so if he can’t start doing better than Obama, he’s got a problem, even beyond the already bleak delegate math.
If Bernie is making real progress in converting voters, Wisconsin would provide important evidence. We now have enough data to narrow this down to two possible outcomes. Either the polls are seriously off, the way they were in Michigan, or Bernie is going to fall short of Obama’s result.
Caucus polls have consistently understated Bernie’s final result. Often, the margin is off by a full 40 or 50 points. We haven’t seen this in primary polls outside of Michigan. New Hampshire, Vermont, Massachusetts, Illinois, Ohio, and Missouri were all in the ballpark.
Down South, if the polls were wrong, they understated Hillary’s total. Again, excepting Michigan, Bernie does sometimes do a couple/few points better than expected, but from a base of +3, he’s not getting to the 18 point margin Obama had unless they are just dead wrong.
To start noticeably clawing back delegates and dramatically reversing the narrative, we figured Bernie would need a 65/35 type of win. Based on the current numbers, that would require a bigger polling error than Michigan. We can safely eliminate this as a possibility.
His best news in the numbers is the improvement with Emerson. As of now, they are the only organization to have polled multiple times in the past couple of weeks. He went from minus 6 to plus 8 in their survey, with only a 10 day gap in between.
His worst news is the result with CBS/YouGov. They tend to favor him in their results. In Michigan, this meant they had him closer than the other incorrect surveys. In Illinois, they showed him ahead, when he wound up narrowly losing.
Here they have him at +2. That doesn’t mean he’s going to lose, but it’s a strong indicator he won’t win by that much. Interestingly, they have him lower than a couple concurrent polls (PPP, Fox Business), so it’s possible they’ve just stopped feeling the Bern, but the internet polling method is usually to his benefit.
Over on the Republican side, it’s a clear two person race. Ted Cruz leads each of the six most recent polls, with Donald Trump still lurking. John Kasich is a very distant third. He’s trailing Trump by 15 points in the RCP average.
If Kasich was going to contend, he needed a couple of favorable poll results so the media would cover it as a three person race and possible Kasich supporters wouldn’t feel they were tossing a vote towards Trump or Cruz by picking him.
With about 10 percent officially undecided and 20% with Kasich, it looks like almost a third of Wisconsin GOP voters aren’t particularly sold on either of the front runners.
When voters are asked about favorability, Kasich scores best. He isn’t showing any ability to translate this to polling support or votes though.
Based on the data, Cruz should win, but he has the chance to put up a big number. Trump still hasn’t cracked the 50% barrier in a contest. If Cruz does here it would make a statement.
With a polling average of 40 and a high number of 43, it would require a few Kasich defectors, the majority of undecided voters and a strong supporter turnout. Don’t bet on this, but it’s mathematically possible.