The Republican Party racked up record numbers of voters in the 2016 primaries and caucuses, a phenomenon for which Donald Trump took credit. He claimed that his populist-outsider campaign, sparked by large rallies and tons of media coverage, had brought a massive wave of new voters to the GOP. Some of us wondered whether the large jump in turnout came from new Republicans, or rather Republicans that normally didn’t vote in primaries. A new analysis of the vote from Politico claims it’s almost entirely the latter:

While Trump’s insurgent candidacy has spurred record-setting Republican primary turnout in state after state, the early statistics show that the vast majority of those voters aren’t actually new to voting or to the Republican Party, but rather they are reliable past voters in general elections. They are only casting ballots in a Republican primary for the first time.

It is a distinction with profound consequences for the fall campaign.

If Trump isn’t bringing the promised wave of new voters into the GOP, it’s far less likely the Manhattan businessman can transform a 2016 Electoral College map that begins tilted against the Republican Party. And whether Trump’s voters are truly new is a question of urgent interest both to GOP operatives and Hillary Clinton and her allies, who have dispatched their top analytics experts to find the answer.

“All he seems to have done is bring new people into the primary process, not bring new people into the general-election process … It’s exciting that these new people that are engaged in the primary but those people are people that are already going to vote Republican in the [fall],” said Alex Lundry, who served as director of data science for Mitt Romney in 2012, when presented Politico’s findings. “It confirms what my suspicion has been all along.”

Sone of this changes from state to state, too. In Ohio, there does appear to have been a genuine expansion of the GOP’s footprint in the primary. However, it’s not clear that it was Trump that drove it. John Kasich made significant inroads into both independents and Democrats in his last gubernatorial election, and he ended up beating Trump by eleven points in the state primary. Less than 10% of the GOP’s turnout in Ohio had not voted in any of the three previous presidential election, and party switchers only ticked up a little more than in 2012 (5.9% from 4.4%). Could some of that stick around for Trump? It’s possible, and the GOP only lost Ohio by 160,000 votes in 2012.

In Florida, however, the news does not look good. Without the Sunshine State, there is almost no path for Republicans to the White House, and the amount of new voters who participated in GOP the primary but had not voted in 2012 or 2014 was just 6%. Party switchers were all but a wash in the closed primary:

Smith, the data-crunching political scientist, said that about 8,500 voters changed their registration in the last two weeks from independent to Republican and then voted either early or absentee in the 2016 primary. But that figure during the same period was almost identical for independents who became Democrats: about 7,600. The statistics show that about 2,000 more Democrats reregistered as Republicans during that period than vice versa.

These analyses are backed up, at least for the moment, by the data from head-to-head polling at the state level, at least thus far. In each of the key states that Republicans must win — either on the traditional swing-state path outlined in Going Red or the Rust Belt path that Trump’s supporters project an Electoral College victory — there is no evidence that Trump will outperform Mitt Romney’s 2012 results. If Trump really was bringing millions of new voters to the GOP, that should be evident in at least some of that polling, but from Florida to Virginia to Pennsylvania to Wisconsin, there hasn’t been a hint of it.

That doesn’t mean that Hillary Clinton has the election wrapped up. If this holds up, it means that Trump and the GOP have to find ways to truly bring in new voters to the Republican Party in the fall, rather than just get general-election Republican voters engaged earlier in primaries. That will take a different kind of campaign, one that Trump seems reluctant to wage.

On the other hand, it also poses a conundrum for conservatives who claim that Trump doesn’t represent the GOP and that he owes his nomination mainly to non-Republicans. If these analyses hold up, then they may also need to come to grips with the reality that movement conservatism has lost its hold on the GOP and has a lot more work to do, too.