It will come as a great relief to Hillary Clinton’s supporters that Bernie Sanders is telegraphing an uneventful end to a campaign he once promised to take all the way to the Democratic Convention. After meeting with President Obama on Thursday, he dialed back that promise substantially, and announced he would meet with Clinton in the coming days “to see how we can work together to defeat Donald Trump.” Presumably he will not be asking her to be his running mate—and presumably, challenging her nomination at the convention doesn’t count as working together to defeat Trump.

So in an arithmetic if not exactly explicit sense, he’s promising to suspend his campaign and endorse her sometime before the end of July.

On the surface, Democrats are eager to turn the page on the primary because they want to run against Trump with a unified party. But at a deeper level, their impatience reflects their anxiety about Clinton’s difficulty appealing to young voters.

Even in her home state of New York, which gave her a sizable and crucial primary victory in April, Sanders won among voters under 30 by a 30-point margin. Somewhere around a quarter of Sanders’s voters say they will never vote for Clinton. If the #NeverHillary contingent of Sanders’s base resembles his base more generally, then these #NeverHillary voters are disproportionately young. Combine that with the fact that young people are famously unreliable voters, and it follows logically that Democrats are concerned that getting out the youth vote in November is going to be very difficult—that unifying the party will be harder for Clinton than it was for Barack Obama eight years ago.

The youth vote is a perennial source of both hope and dread for Democrats. The drop-off in these votes contributed mightily to the GOP’s huge landslides in the 2010 and 2014 midterms. In 2004, Democrats rejected the candidate of the youth vote (Howard Dean) in favor of the establishment candidate (John Kerry) and went on to lose the election. In 2008, they nominated the youth-favored candidate and crushed the GOP.

It makes perfect sense that Democrats are alarmed Clinton doesn’t inspire young people, and worried that she can’t inspire this crucial part of the “Obama Coalition.” But that fear is based on a statistical error; it’s a consequential conclusion drawn from an extremely low sample size. In truth, the differences between the presidential races in 2004, 2008, and 2016 (to say nothing of the midterm elections in between) are so enormous as to render most comparisons between them irrelevant. Clinton can win handily without 2008-level youth turnout. Assuming the Clinton campaign is aware of that, it stands to reason it will not be conceding much to Sanders in exchange for his endorsement.

The youth vote did indeed climb between 2004 and 2008, but the increase wasn’t the difference between victory and defeat for Barack Obama. Indeed, Census data show that youth turnout collapsed in 2012, in absolute terms and as a percentage of the overall electorate, to below 2004 levels. And of course, Obama won anyhow.

In 2004, Democrats were trying to unseat a reasonably popular president. In 2008, that same president was catastrophically unpopular, and the election was a referendum on whether to give his party a third straight term. This year’s election will likewise be a referendum on a president’s two terms, but unlike George W. Bush in June 2008, Obama is actually well-liked. Obama’s presumptive heir is also overwhelmingly popular among Democrats, if not the electorate as a whole. But her opponent, Donald Trump, is one of the most deeply divisive people to ever run as a major party nominee.

To beat him, Clinton can’t write off disaffected Sanders voters altogether. But she doesn’t need to convince the overwhelming majority of them to fall in behind her, either.

A recent Pennsylvania PPP poll illustrates her challenge neatly:

Pennsylvania is a great microcosm of the issue Clinton faces in winning over Sanders fans. Among people who support Sanders in a head-to-head match-up with Trump, only 72% support Clinton in the general. 10% would go to Trump, 6% to Stein, 4% to Johnson, and 9% are undecided. If Clinton could win over even just half of those Sanders-supporting holdouts, her lead over Trump would go from a tenuous 41-40 to a comfortable 47-40.

Just half? It’s hard to fathom that Clinton won’t convert at least that many Sanders holdouts simply by dint of not being Donald Trump. And if that’s the case, she doesn’t have to do much at all to unify the party, or tailor her message to millennials, enough to win the election. She’ll need to concede something to Sanders to insure against him going rogue, and in terms of “issues”—which Sanders vowed on Thursday to take to the convention—that may amount to more than Obama conceded to her in 2008. (The likelihood of Sanders being offered a job like Secretary of State are not just slim, but none.)

At the end of the day, Sanders doesn’t have much that Clinton needs to win. She doesn’t need his infrastructure. She doesn’t need his political network. And, as it turns out, she probably doesn’t even need that many of his discontented voters either.