No poll this early can rightly be called “significant” but this seems … significant-ish. True, there’ll be no head-to-head contests in the primaries (until the very end), in which case all this really tells us is preferences within a certain niche. But that niche is crucial: These three guys have emerged early as the favorites among the center-right and donor class, the groups that typically decide the nomination. They’re also widely viewed as the three most conventionally “electable” candidates. If Rubio’s leading both of his competitors by double digits, that’s an inkling of how undecideds might break as the field narrows.

We’ve already reached the stage in the campaign where fans of different candidates are whining at me on Twitter that I’m biased against their guy, so here you go, Rubio fans. Pure data sunshine from me to you.

The sample size of Republicans is relatively small, so that Rubio/Walker match-up may be much closer in reality than those numbers suggest. Same goes for Rubio/Bush, but in that case, Rubio’s lead is so large that even a “much closer” race probably still means a double-digit advantage. That’s bad, bad news for Jeb, a guy whose name recognition is supposed to give him an edge over a comparative unknown like his old protege. The “Bush” brand, which is still well regarded within the GOP, isn’t helping him here. And don’t think the GOP’s big donors, who are backing Jeb now but don’t want to ride a loser all the way to the finish line, won’t notice.

Even worse for Jeb, Rubio’s out to a conspicuous lead among Latinos both head to head against Bush and against a larger GOP field. The sample size of Latino voters here is very small at just 80 people, but these results are encouraging enough for Team Marco that a bigger poll of Latinos is in order ASAP. When given a choice between Rubio and Bush, Latino voters prefer the former 46/14. (In a Rubio/Walker race, they split 48/17.) When Rubio is lumped in a field with Ted Cruz, Rand Paul, Mike Huckabee, Carly Fiorina, and Ben Carson, he takes 29 percent of the Latino vote versus just 12 percent for Fiorina, the next highest vote-getter. If Latinos nationally are starting to pay special attention to Rubio, Bush’s donor-class hype as the one man who can heal the rift with a group alienated by immigration policy will be up in smoke. Once that happens, the only thing to recommend Bush over Rubio if you’re a billionaire GOPer looking for a horse to back is his family connections. Rubio’s a better retail pol in almost every way and, unlike Jeb on Iraq, he won’t spend a week during the campaign stepping on rakes rhetorically.

One more reason for Team Marco to cheer: Predictably, his big liability isn’t a key concern among Republican voters. Here’s how GOPers answer when asked what the most important issue of the election is:

Immigration’s a respectable third, but still in single digits. When you give Republicans a laundry list of issues and ask them which ones they consider important (but not necessarily most important), 84 percent say so of immigration — fewer than who say the same of the economy, terrorism, health care, Social Security, Medicare, and taxes. And of course, as GOP voters slowly come to realize that Rubio’s really no further left on amnesty than anyone else in the field, whatever Scott Walker might be telling people this week, his Gang of Eight track record will matter even less. If terrorism remains the top issue, Rubio’s well positioned as a conspicuous super-hawk; his foreign policy may be nothing more than old wine in new bottles but Republicans really like that vintage. if the economy emerges as the top issue, Rubio’s well positioned because of his family-friendly tax reform plan. He’s got big, big money backing him now too in Larry Ellison, Norman Braman, and potentially Sheldon Adelson. There’s really no reason, apart from Bushworld’s influence within the donor class, why he can’t win the nomination. And like I say, the better his poll numbers look compared to Jeb’s, the less that influence will matter to people who want to win above all else.