NPR has a nice piece here about how likely Hillary Clinton is to beat Donald Trump. The best part is that, starting with 2012 Obama-Romney turnout by various demographic groups likely to be key in this election, you and I can, by making hypothetical changes this year, with sliders, see what the likely electoral vote margin will be. (NOTE: NPR appears to assume that third-party voting will be unchanged, a dangerous assumption.)
First, if nothing changes, Clinton will nonetheless win bigger over Trump than Obama did either time.
Now, my personal changes.
I moved white women 2 percent more for Clinton plus 1 percent more turnout. Cut black turnout 3 percent. Left all else unchanged.
Clinton wins 363/175. She wins by that margin, if women turnout and break her way like that, even if black turnout falls 5 percent.
Let's make things a lot tougher. Women only break 1 percent more her way, on the same turnout. White men break another 2 percent GOP, and turnout is 1 percent higher. Blacks break the same on percentages, with 6 percent lower turnout than 2012; Hispanic turnout falls 2 percent, same break, as 2012.
Trump ekes out a 282-256 win. And, given that Trump is reportedly below 50 percent among college-educated whites, even that would seemingly be big uphill sledding.
If I keep those parameters, but cut the white male break to 1 percent, she still wins, 294-244. On all parameters, she wins, by that margin, unless total turnout falls dramatically.
Add in that Ted Cruz clearly refused to endorse Trump, in prime time, at the Republican National Convention. And won't this year.
So, she doesn't need my vote. If things are that bad, she's in Dewey Defeats Truman range.