In the aftermath of the presidential election, economist and political science professor Maria Guadalupe of INSEAD wondered, like so many people, whether Clinton would have lost if she were man and whether Trump could have won had he been a woman. Hypothesizing that in a gender-flipped race, Clinton would have come out the winner, she devised an experiment: To restage the debates with the roles flipped, so that Trump is a woman who not only said everything Trump did, but also acted the same way he did, with Hillary becoming a man who spoke and acted just like her. Guadalupe turned to NYU professor Joe Salvatore to direct while she created the script for Her Opponent, a debate between Brenda King (a female version of Trump) and Jonathan Gordon (a male version of Clinton), performed on January 28.

Guadalupe selected moments from all three debates and wove them into one 35-minute play, then Salvatore worked with the actors using Anna Deavere Smith’s ethnodramatic technique in which the actor memorizes the exact inflections and exact gestures/movements of a real-life subject. A 25-minute post-show discussion followed in which Guadalupe and Salvatore were amazed to find that even as a woman, Trump still came out the favorite.

In the discussion and in an online survey completed later, audiences found King/Trump to be concise, authoritative, and commanding. Alternately, they found Gordon/Clinton’s incessant smiling to be totally off-putting. When King attacked, Gordon didn’t fight back; she just nodded and smiled. In the body of a man, this response was disconcerting at best and at worst, one audience member found him “extremely punchable.”

Cut to one-month later and an article about the show on NYU’s news page goes viral. Salvatore is drowned in press inquiries and requests to see the footage they captured when they filmed the performance. Much of the coverage is from right-wing sites and individual bloggers, where they rejoice that two liberals were proved so wrong by their own endeavor. This proves, they argue, that Trump’s win was not the result of sexism but rather of the strength of his message and the inherent unlikability of Clinton.

That Guadalupe and Salvatore’s hypothesis was wrong does not, by itself, prove that Trump deserved to win. Rather, it shows that we are so programmed to see femininity as weak and masculinity as strong that even when masculine behaviors are embodied in a woman, she comes across as authoritative and confident. Feminine behaviors on the other hand, make even men read as subordinate and even a little laughable.

In fact, even though in real life, Clinton does not come across as all that stereotypically feminine, her behaviors are so inherently feminine that audiences assumed that the actor playing Gordon had been directed to act feminine or even to play gay. (Commenters on right-wing websites that picked up the story were thrilled to call him a faggot.) In reality he had not been directed that way; that impression arose purely from him exactly imitating Clinton.

Whereas people who have met Clinton up close find her easy to connect to and personable, her debate performances and speeches are so heavily coached that she comes across a bit stale. Like most women, she has probably been told all of her life to smile more, and even planned when to smile during the debates. Unfortunately, smiling and nodding in response to being attacked may be feminine, but it didn’t make her any more likable than she would have been had she fought back. In fact, the audience response to King – a woman who speaks and behaves with all the bravado, aggression, and sweeping masculinity of Trump – may imply that women have far more freedom to behave that way than they think without being thought of as a bitch.

Guadalupe and Salvatore are remounting the experiment for an off-Broadway run beginning March 22 at the Jerry Orbach Theater and are working on a film that replicates, shot by shot, the debate broadcast. They hope that the discussions after the next round of performances can help them further unpack why Gordon’s femininity was quite so off-putting and how even people who reject Trump’s policies could have found themselves inclined to vote for King, whose policies are exactly the same, rather than Gordon, whom they had in many cases already voted for in the form of Clinton.

Guadalupe told me that her theory is that their experiment says nothing about the appeal of the two candidates’ platforms, because when audiences watch a play instead of an actual debate with real world consequences, they are free to react almost entirely emotionally rather than intellectually. She believes that this decontextualization reveals not the power of Trump’s ideas but rather the power of the way he conveys them. The film, she hopes, can be used as a teaching tool, whether in political science or gender studies classes, to show how deeply reception of content is shaped by context.

Though Guadalupe and Salvatore’s hypothesis that Clinton would have would won were she a man was wrong, their experiment still proved that it is sexist assumptions – in this case, the one that women can be liked only when they act super feminine – that brought her down. Had Clinton had the liberty Trump had to act in a manner as bold as her ideas, or had Trump been forced by coaching to smile the whole time and nod even as his opponent beat the crap out of him, the election might have turned out differently. Trump didn’t win based on the power of his message, nor did Clinton loose on the issues. Trump won because he came across as the most direct, strong, and yes, masculine, candidate.