Donald Trump won a sweeping victory in the Indiana Republican primary. Here's how. (Peter Stevenson/The Washington Post)

With truly remarkable timing, CNN and its polling partner ORC released a new national survey previewing the general election before the sun even rose Wednesday after Donald Trump essentially clinched the nomination. How did they do this? Magic.

What the poll shows is that the almost-certain general-election contest between Trump and Hillary Clinton (barring some dramatic change of fortunes) is Hillary Clinton's to lose. What it also shows, perhaps more alarmingly to Team Trump, is that Clinton's current lead has expanded since an initial surge of interest in Trump last fall.

Overall, the pattern looks something like a U. Last June, shortly after Trump announced his candidacy, Clinton led him by 24 points. By September, the two were tied in CNN-ORC polling, thanks to a combination of Clinton's rough summer and Republicans starting to embrace Trump. But as time passed, the pendulum swung back to Clinton.


As is often the case in our polarized politics, that swing is in part thanks to movement among independents. Republicans were initially skeptical of Trump, but he persuaded them; they now prefer him by a wide margin (though Democrats prefer Clinton by a wider one). Independents, however, moved quickly toward Trump and then more slowly away from him.

One of the arguments Bernie Sanders has used in defense of his ongoing campaign is that he can win independents in November. CNN suggests that Clinton can, too. (For what it's worth, Sanders leads Trump by a wider, 16-point margin among independents -- the same margin as his overall lead.)


Since the beginning of the year, another big gap has opened up based on gender. Women have preferred Clinton for a while, but in CNN's three most recent polls, they've preferred her by more than 25 points. That's about where they were last June, but back then, men were there, too.


Speaking of huge gaps! White voters prefer Donald Trump, as they have since last June. But nonwhite voters vastly prefer Hillary Clinton -- at this point by a 67-point margin. (Sanders leads with this group by a smaller margin.)


The split by education is interesting. In June, those who had or hadn't attended college were about split between the two candidates. (In October, the demographic split became those who did or didn't have a college degree.) Over time, a consistent split emerged: Less-educated voters were more likely to favor Trump.


Normally, education and income correlate, but here, higher-income voters still prefer the Republican even while more educated ones don't. This reveals some of the challenge with the broad buckets into which voters are gathered for these polls; splitting out white and nonwhite subsets of these groups would probably explain most of the differences.


We've noted frequently that general-election polling conducted well before the election isn't particularly accurate. Once you know who's actually running, as we now apparently do, things start to get better. But the trend itself is significant, even if the top-line margins aren't.

After launching his campaign and convincing Republicans he was for real, Trump managed to run competitively with Hillary Clinton. As the campaign progressed, voters turned against him. His challenge now isn't to figure out how to steer the ship as it leaves port; it's to execute a wide turn against a current.


Other demographic matchups. Before December, the age split was older or younger than 50; after, it was 55.