For the purposes of this analysis, I think we should view Trump as an incumbent when measuring his performance in the primary. Republican voters know exactly what Trump is like in the White House, and they can judge for themselves how he did on the issues they care about (he generally does well on this axis) and his character (things are more muddled here).

If we think of him as an incumbent, he is doing quite poorly. Remember how Pat Buchanan embarrassed H.W. Bush by getting 38 percent of the vote in New Hampshire in 1992 and is blamed for weakening him for the general? Haley has been doing better than Buchanan across the early states. There are plenty of Republican voters who don't think of Trump as ideal. We've probably all heard their arguments. Something along the line of, "his character isn't ideal, but he fights for me."

Now, what does this all mean for the contest against Biden in the fall? Not much. The vast majority of those people will vote for Trump in the fall. And if we are sticking with the 1992 comparison, Bill Clinton was a popular candidate (net +11 favorability in September of '92). Joe Biden is not. His favorability rating is currently at net -15 points. So, Trump is doing poorly, but so is Biden. When it's a race to the bottom all around, doing poorly in a GOP primary doesn't tell us all that much.

—Galen Druke, 538

I see your point, Mary — Trump has gotten “only” around 50-60 percent in the first three contested states — but I think in the grand scheme of things, that’s still a strong performance. Iowa winners much more often finish in the 20s or 30s, and New Hampshire is just weird because of its demographics and high proportion of independent voters. And at the end of the day, he’s the first non-incumbent Republican to win all three of Iowa, New Hampshire and South Carolina, and there’s a good chance he’s going to win all 56 contests.

—Nathaniel Rakich, 538

For a non-incumbent (if former president), Trump has made short work of the primary calendar, basically annihilating any real competitors in record time without ever having to deign to step foot on a debate stage. The fact that he didn’t completely cruise to the nomination captures some of the hesitancy in the party that Julia highlighted, but at the end of the day, all Trump needs to do now is be more popular than Biden, who isn’t exactly the Prom King these days.

—Kaleigh Rogers, 538

I'm honestly not sure there is a clear way to go about establishing what Trump "should" be winning, Mary. On the one hand, polls had him at 61 percent, so 60 percent of the vote (as of now) is pretty good — especially if you consider he'll win the nomination easily by mid-March if this is the case.

Another benchmark would be the percent of the GOP electorate that calls themselves "part of the MAGA movement." According to preliminary ABC exit polls, that's about 44 percent of SC voters (50 percent said no and 6 percent were unsure). So by this benchmark, Trump is actually overperforming.

Too bad there aren't a whole lot of modern examples of former presidents + failed presidential candidates running in subsequent primaries, or else we could try to establish historical baselines. Trump's performance right now is somewhere between George H.W. Bush vs. Pat Buchanan and Jimmy Carter vs. Ted Kennedy. Both of those incumbents lost in the general.

—G. Elliott Morris, 538

I have become a little obsessed tonight about what we should be expecting Trump to hit in this primary a priori. That is, given Trump is assumed to be the eventual party nominee and almost universally liked in the GOP, should he be winning more than 60 percent in South Carolina?

I already gave my case for answering "no" to that question: Strictly speaking Trump is dominating the delegate count and running ahead of his 2016 vote share in most counties with complete counts this primary cycle. And if you consider that Haley gets a home-state advantage in South Carolina tonight, Trump's adjusted vote share is close to 65 or 70 percent; our delegate benchmarks think Trump should have won 68 percent of the vote based on the demographics of the state alone. That's not the highest number, but it's not the lowest right? Would 65 percent be "good" for Trump? 75 percent? 80?

One counterargument to this centers around how the media has covered historical performances by incumbent presidential candidates. Journalist Jill Lawrence points out that in 1992, Patrick Buchanan challenged incumbent President George H.W. Bush for the GOP nomination and won 40 percent in the New Hampshire primary, holding Bush to 58 percent of the vote. That's an almost identical split to the results from tonight. The New York Times journalist Robin Toner wrote up the results with the headline "BUSH JARRED IN FIRST PRIMARY" and said the result "amounted to a roar of anger" from Republican primary voters.

If Trump was a true incumbent, I imagine the news media would use a similar headline to describe tonight's results in South Carolina. Perhaps our expectations for him are too low, or we're too focused on the broader state of play? Haley said in her concession speech tonight that she will stay in the race indefinitely, so I guess we'll get more data on Super Tuesday — only 10 days from now. The primary lives on!

—G. Elliott Morris, 538