During his triumphant return to the stage at the Conservative Political Action Conference, President Trump — a onetime curiosity who now dictates the direction the wind blows — made an unexpected and erroneous aside. Well, he made lots of those, but this one in particular was interesting.

“You know, we haven’t been as a group given credit for this, but look at how much bigger our party has gotten during this cycle,” he said. “During the early days when we had 17 people running, the primaries, millions and millions of people were joining. I won’t say it was because of me, but it was. And we have an amazing, strong, powerful party.”

Trump’s speeches often have a skipping-record quality, where he’ll keep rehashing applause lines even well after they make any sense. He regularly bashes Hillary Clinton, for example, as though he still has her looming as an opponent on the horizon. This claim about the Republican Party is a skip, too; he has been saying it for a while.

It started before the primaries. The 2016 election saw a legitimately big surge in turnout among Republican voters, as indicated in data from the U.S. Election Atlas.

But millions more Republican primary voters is not the same as millions more Republicans. Politico looked at who was turning out in the primaries and found that the new voters were heavily people who didn’t usually vote in primaries but did in the general. In other words, Trump was ginning up interest in primary voting, not in the Republican Party.

What’s more, Trump was likely inspiring people to come out and vote both for and against him. He ended up getting less than half of the votes cast in the primary — and more people cast votes for someone other than the eventual nominee than in any other cycle in history.

Again, Trump’s been saying this thing about there being millions more voters for about a year. Meaning that it predates a bit of data that severely undercuts the claim: turnout in the general election.

The 2016 election saw more votes cast than any other in American history. That’s in large part a function of population growth, of course; there were about as many third-party votes last year as there were votes for the Republican candidate a century earlier.

There was not, however, a massive spike in new votes for the Republican candidate. There was a slight uptick, while the votes for the Democratic candidate stayed about the same, but as a percentage of growth, it was fairly modest. Contrasted with the change in Republican votes cast between 1976 and 1980 or 1996 and 2000 or 2000 and 2004, it was a blip.

A look at data from large states that report party identification shows that the change in Republican registration was scattershot.

In each of the five states we looked at — blue states California and New York and red states Florida, Pennsylvania and North Carolina — Republican registration was up in 2016 relative to 2014. But only in the red states was it up relative to 2012. This is percent change in registered voters — meaning that even as the population in California increased, the number of Republicans dropped. Sure, there were gains in Florida, but a net reduction in Republican registered voters in the largest state in the country is not exactly evidence of a giant surge in enthusiasm.

There’s another way to look at this — a particularly interesting one.

Gallup regularly polls Americans and asks them how they identify by party. Did you notice all those yellow bars on the graph above? That reflects the long-term trend in Gallup’s findings: People are more and more likely to identify as independents. The percentage of respondents who identify as Republican has increased slightly from late 2013, when the government shutdown hurt the Republican Party. But mostly the story has been people moving from the parties to calling themselves independents.

I hope you noticed the remarkable thing that happened at the far right of those lines. The number of people who identify as Democrats plunged from Election Day until January of this year (the most recent data available). Who gained? The independents. People are less likely to call themselves Democrats — but it isn’t Trump’s Republican Party that gains. In fact, the percent of people who identify as Republicans — 28 percent — is precisely what it was on Election Day in 2014.

One of the main skips in Trump’s speaking-record is how great his electoral victory was. It wasn’t; he lost the popular vote and had one of the lower electoral vote totals in modern history. But he seizes on whatever metrics he can to demonstrate that he’s got the will of a powerful movement propelling him forward. On this day, to that conservative crowd, Trump was hoping to suggest that his candidacy had made the Republican Party a much more powerful force.

In reality, though, it was skeptical Republican voters coming home to the party on Election Day that gave him the narrow margins he needed in the Midwest to win the electoral vote. In other words, the power flow worked the other way.