On the day he formally announced his candidacy for president, Bernie Sanders put the country on notice: “Don’t underestimate me,” he said in an interview with ABC News. He was speaking to those who doubted the chances of a 74-year-old democratic socialist from Vermont—as unkempt as he was unknown—winning the highest seat in government.

Despite this warning, even his supporters initially saw his campaign as more of a statement than a serious bid. Almost no one believed he had even an outside shot at contending with the financially supercharged Hillary Clinton.

Yet here we are. By February 10 his campaign had received an unprecedented 3.7 million contributions, by far the most of any candidate. Now, coming off a decisive victory in the New Hampshire primary, a candidate whose name was recognized by less than five percent of those polled at the beginning of the race is neck-and-neck with Clinton. Enthusiasm for Sanders’ campaign and optimism about his ability to win the 2016 election have never been higher.

An outsider and an underdog, his simple populist message and decades-spanning consistency on the economic and social issues he is so genuinely passionate about is hard not to like. His unpolished manner and appearance seem to authenticate his portrayal as an alternative to so-called establishment candidates—Clinton, Jeb Bush, and Marco Rubio in this election—who seem focus group-tested and board-approved.

Unsurprisingly, the support of young voters has been crucial for Sanders’ success: he received 84 percent of the vote for the 18 to 29 demographic in Iowa and 83 percent in New Hampshire. Of the 301 sophomores, juniors, and seniors the Observer polled at Tufts, 53 percent plan to vote for Sanders, while 43 percent plan to vote for Clinton in the primary.

Young people will play a critical role in deciding the Democratic nominee, and their support seems to be staunchly with Sanders. It’s hard to blame them—he’s inspiring and authentic. But no matter how desperately some might hope Sanders wins the presidency, the decision to vote for him in the primary should not be an obvious one.

In the primary, you’re not simply electing who you like, but who you think stands a chance to win the general election. It is critical that this belief be based on an assessment of the available evidence. The question of electability should figure into the party’s choice of nominee in a significant way for a simple reason: Marco Rubio, Ted Cruz, or Donald Trump winning the election would be nothing short of a tragedy.

This is not to say that Sanders is necessarily less likely than Clinton to win the general election. There are, in fact, reasons for Sanders enthusiasts to be optimistic. In the New Hampshire primary, Sanders won 60 percent of votes among self-identified moderate and conservative democrats, the same percentage of votes he won overall. This suggests that moderates on a national level might be willing to vote for him in a general election as well. He has proven uniquely capable of mobilizing young and first-time voters, which may hold true in a general election, and liberals typically fare better in elections with a greater turnout and more young voters. The momentum of his campaign has increased unabated for months, and perhaps it’ll continue to do so through November.

Certain forecasts predict Sanders to succeed in a general election. Western Illinois University’s famous mock election, which has correctly predicted the past several elections and bills itself as the “largest and most elaborate mock presidential simulation in the nation,” predicted a Sanders landslide victory, with Martin O’Malley as his running mate. Alan Abramowitz, a political scientist at Emory University, told the Washington Post he believes Sanders is a “risky candidate.” According to Abramowitz, his nomination would not necessarily result in a loss, at least not a blowout on the magnitude of Lyndon Johnson’s 1964 landslide victory over Barry Goldwater or Richard Nixon’s 1972 demolition of George McGovern (Goldwater and McGovern, like Sanders, were perceived as outside the political mainstream).

This is due to the rise of what he calls “negative partisanship”: “In both parties, voters are much more ideologically cohesive and dislike the opposing party much more than in the past,” he said. “On the Democratic side, there are far fewer conservative voters who would prefer a Republican to even a very liberal Democrat like Sanders.” Abramowitz argues that because the GOP won’t nominate a moderate—all of their frontrunners are extreme, “high risk” candidates—they would be no more likely to attract moderate voters than Sanders.

On balance, however, there are many more reasons for caution than optimism, and Abramowitz’s assessment is at odds with the consensus among the most qualified analysts. Earlier this month, Vox asked six of the top political scientists in the country how Sanders would fare in a general election, and they agreed nominating Sanders would make the Democrats’ bid to retain the White House an uphill battle.

Seth Masket, a political science professor at the University of Denver, estimated nominating Sanders would cost the Democratic Party 2 to 3 percentage points in a general election over a more conventional candidate like Clinton. Bruce Miroff, a researcher at the University of Albany, said the disparity would be more in “the vicinity of 6 to 10 percentage points.”

James Glaser, the Dean of the School of Arts and Science at Tufts and a political science professor who researches electoral politics and political behavior, believes Clinton has a better chance than Sanders in a general election because her positions are more moderate. “Depending on who the Republican is, Sanders is less likely than Clinton to hold onto moderates and independents in the general election,” he wrote in an email. “And while he has an edge in enthusiasm among his core voters, her voters are more reliable.” Sanders’ appeal may be limited to these core supporters, enthusiastic as they are.

As the question of electability has entered the race, Sanders has repeatedly cited his favorable polling numbers in general election matchups with various Republican candidates, but a look at past elections shows such polls are essentially meaningless so many months before the election, in part because the issues the public cares about can shift hugely.

“The impressions people have of the eventual nominees months from now will be so different from today,” Seth McKee, a political science professor at Texas Tech University, told Vox. “That’s a nice thing to point to, but what does a head-to-head poll mean in early February? … It’s worthless. It’s absolutely worthless.”

In 2008, Barack Obama and John McCain were tied in the polls at the beginning of the primaries, and “at that point in the campaign, more people cared about foreign policy than domestic issues,” Harry Enten of FiveThirtyEight said in an article. “That changed dramatically as the global financial system collapsed, and Obama went on to win by over 7 percentage points.”

Under the scrutiny of a general election race, Sanders’ obvious lack of experience with or knowledge of foreign policy could be his undoing. A meaningful discussion of national security has been conspicuously absent from his campaign, and Sanders’ committee assignments in his career in the House and Senate have included Senate Budget and House Financial Services, with Senate Veterans’ Affairs the closest to foreign policy. In his opening remarks at the Democratic debate the day after the November 13 Paris attacks, Sanders devoted just a few seconds to the crisis before returning to economic injustices, his bread and butter, for which he was widely criticized.

“When [national security] rears its ugly head and when a crisis comes out and the American people are talking about it, candidates create risks for themselves if they don’t communicate on those topics,” Joel Rubin, a former deputy secretary of State for House affairs told The Hill. If events again thrust foreign policy issues into the spotlight, Sanders’ campaign would likely suffer. Regardless, his inexperience is a clear shortcoming his opponents will exploit in a general election race.

For months, the GOP has largely ignored Sanders to focus their attacks on Clinton, but were he to secure the nomination, the resources expended against him would be immense. That a self-identified socialist with minimal foreign policy experience could withstand such a campaign is uncertain: a June Gallup poll found that while 24 percent of Americans would discount a presidential candidate for being gay or lesbian and 38 percent would refuse to vote for a Muslim, a full 50 percent would never support a socialist candidate, the least permissible demographic of all. Limited as polls are, this is telling. Maybe Sanders will overcome the public’s aversion to the prospect of a socialist in office, but it’s a risky bet.

Another reason Sanders could struggle in the general election is his lack of support from voters of color, who comprise a crucial, often pivotal swath of the Democrat voting coalition. National polls show that Sanders trails Clinton by almost 40 percentage points among non-white voters. The Sanders campaign is aware of this and has invested in ground organizing efforts in states with large black populations like South Carolina, made a show of meeting with Black Lives Matter activists and black public figures, and has made a series of campaign stops at historically black colleges. But closing a gap that large will be difficult. Sanders’s success in Iowa and New Hampshire, both of which are more than 90 percent white, may not translate to Nevada and South Carolina, which are only 65 and 43 percent white respectively. The momentum his campaign has enjoyed for so long could suddenly flag.

Electability aside, there are other reasons to think twice before voting for Sanders. His sweeping, drastic policy proposals—a pure single-payer healthcare system, dramatic tax reforms, and free tuition at public colleges, among others—would only have a chance of finding success if the Democrats can take control of both the Senate and the House. This would require taking five and 30 seats respectively at the November elections. Gaining 30 seats in the House is basically out of the question, and should the Democrats fail to do so, Sanders will be hamstrung by a Republican-dominated Congress. Sanders’s “political revolution” would depend on an unprecedented turnout at the Congressional elections; that Republican voter turnout was better in both Iowa and New Hampshire is not encouraging.

If Democrats fail to gain a majority in both the House and Senate, which will almost certainly happen, a Clinton presidency may prove more effective. “I think Clinton is more willing to compromise than Sanders,” wrote Glaser. “And this will lead her to be more likely to have legislative success, depending on the composition and attitude of the Congress. Political revolutions require a large swath of the electorate behind you.” Sanders is unlikely to have that sort of backing.

Of course, many Sanders supporters know this, and understand that should he be elected, his policy proposals will not be realized in full. The tacit truth of his platform is that is more an expression of his utopic vision for the country’s development than it is a realistic plan. But idealism is not action, and Congress will torch anything but the most diluted forms of his proposals—and maybe not even those. For better or worse, the legislative process is defined by compromise. Progress is incremental. Ours is not a system built for revolution, and a revolutionary may not be what we need.

“I don’t think it’s cynical to think about things like legislative success or electability as you cast your vote,” wrote Glaser. “The question is which candidate will maximize the likelihood that your values and your positions will be well represented in government and these concerns are part of that equation.”

No one can tell you who to vote for, but it’s worth considering how you should go about making the decision, and ignoring the question of electability is irresponsible. It’s clear nominating Sanders is a risk, and voters have an obligation to factor that risk into their decision, whatever it may be.