Via Newsbusters, I don’t know what he’s basing this on. True, that NBC/SurveyMonkey poll this week did show Trump a few points better than Cruz against Hillary and Morning Consult’s 50-state poll had Trump winning four more electoral votes against her than Cruz would. Both of those polls, though, fly in the face of most other national surveys. On March 8th, RCP’s average of head-to-head match-ups between Clinton and Trump had him trailing her by four points; a month later, he trails by 10.4. Cruz has slipped a bit too over the same period, but he’s still within striking distance: He led Clinton head to head on March 8th by 1.5 points and now trails by 2.8. In the last major national poll testing him against Hillary (from Marist/McClatchy), he managed to tie her at 47 apiece. Trump trailed by nine.

If you want more current data, you’re in luck. WaPo/ABC are out with a new national poll today confirming that lots of Americans dislike Cruz but many more dislike Trump. And the ones who dislike Trump tend to dislike him strongly.

Thirty-one percent of Americans have a favorable view of Trump while 67 percent are unfavorable — nearly identical to an early March Post-ABC poll which found he would be the most disliked major-party nominee since at least 1984. Over half the public (53 percent) continues to see Trump in a “strongly unfavorable” light, ticking down from 56 percent last month…

Trump also continues to receive strongly negative ratings among several key voting blocs that are at least partly up for grabs this year. Two-thirds of political independents have an unfavorable view of Trump, as do 74 percent of Americans under age 40; 75 percent of women, and 81 percent of Hispanics. Majorities in each group see Trump in a “strongly unfavorable” light, exceeding intense negative ratings of Cruz or Kasich by at least 20 points.

Should those ratings fail to improve, Trump’s potential path to victory rides on a surge in support and turnout among whites, particularly those without college degrees. Yet Trump’s image among both groups is underwater. Whites see him negatively by a 59 to 39 percent margin, while non-college whites tilt negative by a narrower 52 to 45 percent.

Cruz’s numbers are also poor at 36/53 favorability, although he’s 20 points lower than Trump is among voters who view them “strongly unfavorably.” Per ABC, Trump’s overall rating of 67 percent unfavorable is the worst they’ve measured for any presidential candidate since 1984, with one exception. That exception: David Duke, who was barely worse than Trump at 69 percent. This is the guy whom Scarborough thinks is a better shot than Cruz to save some Republican Senate seats.

The logic here is this, I take it: Cruz can’t play in purple states because he’s a doctrinaire conservative whereas Trump is a “radical centrist” with some crossover appeal thanks to his strength among working-class independents and Democrats. (Hypothetical strength, I mean. As noted above, the actual numbers don’t bear this out.) It’s understandable that Scarborough, a centrist himself, would subscribe to the “right-wingers can’t win general elections” theory but using Trump and Cruz as an illustration of that is dubious in two different ways. First, there’s no actual evidence (yet) that Trump would play better in purple states than the more dogmatic Cruz. A Quinnipiac poll of Pennsylvania taken last week had Trump trailing Hillary by three but Cruz tied with her at 43. An Ohio poll taken in mid-March had Trump trailing by six — and Cruz ahead by two. The last poll of Virginia had Cruz trailing by eight and Trump trailing by nine. And in Wisconsin, where Cruz thumped Trump in the primary, he’s within three of Clinton in the latest poll while Trump trails by 10. I think it’s true that a generic populist Republican with protectionist leanings would do much better in swing states than Cruz would, but we don’t have a generic candidate in that niche. We have the opposite, a guy who’s roughly as popular as a famous Klansman was when he ran.

Which brings us to the other flaw in Scarborough’s theory. Trump is so deeply unpopular that he might make red states vulnerable to Hillary, which would offset his hypothetical gains in purple states. This is a guy, remember, who’s neck and neck with her according to some polls in Utah and Mississippi. If that polling is borne out this fall, he’d have to decide whether to (a) spend resources defending those states, which will limit the resources available to him in purple states, or (b) ignore those states, hoping and trusting that they’ll stay loyal while Democrats pour money into them. And that assumes that Trump’s weakness in places like Mississippi wouldn’t show up in more competitive states like Ohio, which it almost certainly would. For someone like Kelly Ayotte, the key difference between Cruz and Trump, I think, is sheer predictability. With Cruz you know what you’re getting, you can tailor your message accordingly, and you can trust that he’s not going to say or do something off the wall that dominates the news for a week and which you’re made to answer for. With Trump, every day brings new “WTF” possibilities. In fact, wasn’t it, er, Joe Scarborough who pronounced Trump “disqualified” from office a few months ago out of disgust when Trump wouldn’t give Jake Tapper a straight answer about Duke and the KKK on one of the Sunday shows? That’s the risk you run with him as nominee. You never know when he’ll do something to evoke that same response in swing voters.

Bottom line: The choice for Ayotte is between someone at the top of the ticket who’s superbly organized, politically familiar, and sure to have party regulars turning out for him en masse and someone whose campaign is famously disorganized, whose political positions in the general are a black box right now, and whose support is unknown beyond the minority of Trumpist Republicans who are fiercely loyal to him. With whom would you rather take your chances?