Donald Trump is constantly boasting about the record-setting numbers of GOP primary votes he has received, claiming that this shows he can be competitive with Hillary Clinton in November.

While Trump is correct in pointing to the record millions of voters who have cast ballots in the GOP primary, an analysis by Politico shows that most of those voters are not "new."  They may be new to voting in a primary, but most of them would vote in November anyway.

"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.”

For this analysis, Politico obtained voting statistics from GOP officials and independent analysts in the handful of states that have so far released such information. To varying extents, the findings rebut both of Trump’s central claims: that he has brought in waves of new voters and that he has attracted flocks of Democrats. Among the highlights:

In Iowa, the Republican caucus turnout smashed its past record by 50 percent this year, jumping from 121,000 to nearly 187,000. But, according to figures provided by the state party, 95 percent of the 2016 caucusgoers had previously voted in at least one of the past four presidential elections—and almost 80 percent had voted in at least three of the past four.

The new caucusgoers, in other words, are likely to vote in November anyway.

In South Carolina, which also saw record turnout, data from the state GOP show that first-time voters amounted to 8.4 percent of the GOP electorate. But triple that amount—roughly 25 percent—were only first-time voters in a Republican primary. Even with historically high turnout, the data from the state party show that the Trump-led ballot brought almost exactly the same number of former Democratic primary voters into this year’s GOP primary as a Trump-free ballot did four years ago.

And in Florida, one of the nation’s most critical battleground states, Republican primary turnout jumped by 40 percent from 2012 to 2016. But only 6 percent of those who voted in the 2016 Republican primary did not vote in either of the 2012 or 2014 general elections and were registered to vote then. That amounts to a lot of people—about 142,000—but it’s a fractional share of a populous and fast-growing state that has added almost 1 million voters to the rolls since the beginning of 2012.

While this analysis shows not many "new" voters in the primaries, it still doesn't prove that Trump can't attract Democratic votes in the general election.  Also, the sample size is suspect in that the states that appear to have been analyzed are largely GOP states anyway.  An analysis of New Hampshire, for instance, would be valuable.  New Hampshire is an open primary state, and if Trump were able to attract working-class Democratic voters there, might he not be able to do the same thing in Michigan, Pennsylvania, Ohio, and other rust belt states?

But there is still cause for worry for Republicans.  It means that Trump's general election message must be tailored to maximize his appeal among white, working-class Democrats in swing states.  Can he do that without losing his core constituency in the GOP? 

On this, the election may turn.