What Went Down In The Indiana Primary

The Democratic race remains fundamentally unchanged after tonight’s win by Sanders. Yes, his victory was somewhat surprising, given that all of the polls had Clinton winning and by an average of 7 percentage points. And yes, Sanders has promised to fight on in the primary until perhaps the convention. The problem for the Sanders campaign remains delegate math and demographics. Right now, Sanders looks like he’ll earn about five to 10 more delegates than Clinton in Indiana. That means Clinton will have an elected delegate lead by the end of the evening of around 280 to 285 delegates. In order to catch Clinton in the elected delegate count, Sanders would need to win over 65 percent of the remaining elected delegates. That’s actually higher than it was before Indiana voted. Perhaps as importantly, there’s not anything in the Indiana result that should make one think that Sanders has dramatically changed the result. According to a demographic model published last week by Nate, Sanders was expected to win the state of Indiana by 7 percentage points. That’s about the size of his lead right now. Indeed, you can look at the exit polls and see that Clinton is holding onto the demographic groups she usually wins. For instance, she is beating Sanders among black voters by 52 percentage points. That’s actually slightly better than she did among black voters in New York. I know that some people will think tonight’s polling error in favor of Sanders could be predictive of errors to come. As I pointed out in a different post, it’s actually par for the course so far in this primary. It’s nothing like the Michigan polling error we saw earlier in the campaign. Sanders will need far larger polling errors going forward to have a shot at the nomination. As I said at the top, Sanders will continue to fight on, and he will win votes. He looks like the favorite in the West Virginia primary, for example, which is coming up next. He’ll also do well in the remaining states in the middle of the country. Still, it looks like Clinton and Trump are going to be the nominees of their party.

We’re always on the lookout for potential new ways to forecast election results, even as we remain skeptical about whether they work. Here’s an update on how two measures of online interest in the candidates did in the Republican race today: Google search volume continues to be a decent proxy for how Republican primaries and caucuses will play out. Trump dominated in searches all day — until Cruz announced he was suspending his campaign and achieved a Pyrrhic victory. Meanwhile, TweepsMap projected a Trump landslide based on its analysis of the geographic location of the candidates’ Twitter followers and their engagement in the last 24 hours before the primaries. On the GOP side, Trump had the most Indiana followers, and was the subject of the most positive hashtags in the state. This has less of a track record than the Google Trends approach, but at least today, in this race, it pointed the right way.

Now that we can call Trump the presumptive nominee of the Republican Party, I suppose, Democrats seem to be bouncing back and forth between glee and panic. Which is the more evidence-based emotion? Back when we thought political science mattered, there were two pretty consistent schools of thought that suggested candidates and campaigns were less important than news reports often suppose. Election forecasts based on what political scientists like to call the “fundamentals” — the state of the economy, how long the incumbent party has been in office, and how popular the incumbent is — are usually pretty accurate. Second, polarization has emerged as a powerful force in American politics. With an electorate that’s pretty set in its party preferences, it’s not too hard to figure out what the vote might look like. All of this suggests that the GOP nominating a wild card like Trump doesn’t matter that much. But candidates do sometimes underperform based on what the “fundamentals” suggest, and campaigns may well play a role in helping voters figure out the cues from the political environment. In particular, voter mobilization seems to have been a big part of the story for Obama’s last two campaigns. Mobilization is perhaps the biggest question mark for a Trump general-election candidacy. He’s made some friends within the elite party tent, but not that many. Will the party coalesce around him and mobilize on his behalf? Or does the candidate really matter when it comes to inspiring these efforts?