Discrimination and the Female Playwright

While researching for a town hall meeting held last fall at New Dramatists to discuss the low numbers of female written plays reaching production, I noticed that, by every estimate, work by women made up only approximately 17% of the total number of new plays produced in this country; yet, in an apparent paradox, 31% of the plays on the Theater Communication Group’s list of the “Top Ten Most Produced Plays in American Theatre” were written by women.

Seven months after that lunch, on May 27, 2009, Emily presented the findings of her study, “Opening the Curtain on Playwright Gender: An Integrated Economic Analysis of Discrimination in American Theater” at the 59E59 Theatre. The results were enormously surprising and complex. Unfortunately, most of the subsequent media coverage was sensationalist in nature — “Women Beware Women!” — and offered highly misleading and reductionist reports. But more on that later...

In her presentation, Ms. Sands explained that her “Integrated Economic Analysis” consisted of three separate studies, each employing a highly trusted form of research methodology. (We’ll refer to these studies as the Audit, the Doollee, and the Broadway). The aggregate results of these studies did indeed indicate that gender discrimination exists in the American theatre and also suggested compelling reasons why. Here’s a brief rundown of how each study was conducted and what the findings were:

The Audit Study

Discrimination is notoriously difficult to prove. Was a person not hired because of their race or did they come off badly in an interview? An audit study overcomes this ambiguity by isolating a single characteristic, such as race or gender, and controls for everything else. A famous study by Marianne Bertrand and Sendhil Mullainathan has a title that says it all: “Are Emily and Greg more Employable than Lakisha and Jamal, A Field Experiment in Labor Market Discrimination.” Spoiler alert: yes. Based exclusively on resumes that were identical in every way but name, fictitious Emily and Greg were granted many more interviews for employment than fictitious Lakisha and Jamal. The number of opportunities “Emily” and “Greg” enjoyed could be directly attributed to societal bias and called discrimination because the study effectively eliminated every other possible explanation.

Using this as a model, Ms. Sands asked us to gather four previously unseen ten-page script samples written by women, two with male protagonists and two with female protagonists. She then made equal copies of the scripts, assigned a male pen name to one half and a female pen name to the other, and sent them out to the 252 theatres (culled from The Dramatics Sourcebook and The Dramatists Guild Resource Directory) that had agreed to participate in what she presented as a “Princeton-sponsored study about the process of script evaluation.” Eighty-two of these recipients completed the survey. The majority of the respondents were Artistic Directors (56%) and most of the remaining 44% were Literacy Managers. The gender breakdown was approximately 50/50. Each respondent received all four scripts, identical except for the male and female pen names on the title pages. Some theatres received four scripts with all male names, some with all female, most with a combination. The collective responses to “Script A” with a male pen name were then compared to the collective responses to “Script A” with a female pen name, and so on for all four plays. The results were very surprising.

But first:

Cecilia Rouse predicted that we probably wouldn’t find bias through an audit study since the cat usually falls out of the bag during the process (a few letters from respondents confirmed this) and subjects will hide bias if they suspect it’s being looked for. But, she added, not finding something means very little in economics. For example, imagine we send you into a dark room to find hidden oranges. You have five minutes. Go. You emerge empty handed. We ask you how many oranges were hidden in the room. The correct answer would be “I don’t know.” You can’t say “zero” just because you didn’t find any. After all, there could be oranges hidden deep under the floorboards. Or maybe you’re a lousy orange hunter. Or both could be true. If, on the other hand, you emerge from the dark room with three oranges, you can confidently and accurately state, “There were at least three oranges hidden in that room.” Your “findings” are those three oranges. They do not constitute proof that there were only three oranges hidden in the room.

Why are we comparing gender discrimination to oranges? Because the results of Ms. Sands’s study yielded similar scenarios. While it didn’t find evidence that men discriminate against female-written scripts, that is not proof that they don’t. And though it did find proof that women discriminate against female-written scripts, that is not proof they’re the only ones who do.

The crucial point that much of the media and bloggers missed in their reportage of these particular findings, is that the discrimination displayed by the female respondents of the study was not the straightforward, taste-based variety that indicates personal bias; rather it was an unusual form of statistical discrimination that, for lack of a better word (and believe us, we looked), we’ll call prophetic discrimination. We define prophetic discrimination to be a projection, based on personal experience or observation, that others will discriminate against a person’s own group, causing them, in turn, to discriminate against their own group as well, because of the assumed likelihood that hiring (or advocating for) a person in that group would be economically damaging. For example, an African-American who believed that although Obama was the superior candidate, he could not win the U.S. presidential election because of racial bias, might have voted for a more “electable” white candidate in the primary. Voting in such a way could be called discrimination, because it was based in the candidate’s race, but that would be a laughable and gross oversimplification.

Similarly, the survey’s female respondents did not judge work to be of lower artistic merit when they believed it to be written by women; the lower marks were given on questions that related to how they believed others would receive the scripts (i.e., the prospects of it winning awards and critical acclaim, of it being widely produced elsewhere, and if the materials would resonate with the theatre’s audience). Moreover, Ms. Sands found that the scripts rated the lowest by the female respondents were the ones that had both female pen names and female protagonists. In particular, the question “How likeable are the play’s characters?” yielded the most statistically robust results. Female characters that were “likeable” when purportedly written by men, evoked the opposite response when they were purportedly written by women. The word “likeable,” of course, is problematic because it’s unclear whether it pertains to a person’s own opinion or the perceived opinion of others. In the context of a study however, in which the majority of the questions were about the perceived opinions of others, it seems a fair assumption that the “likeability” question was viewed that way as well. When respondents were asked to rate a script based purely on their own personal tastes (i.e. “How eager would you be to produce X’s script?”), absolutely no discrimination was found. Unfortunately though, female respondents did believe that scripts with a female pen name were far less likely to “fit with their theatre’s mission statement” than the identical work when it wore a male pen name. In essence, their responses indicated that although they would like to produce work by men and women equally, they believed their hands were tied. (It should be noted here that the audit study’s findings were so statistically significant, that if the same study were repeated under perfect circumstances, one could expect to get the same results at least 95 out of 100 times.)

The Doollee Study

In this part of Ms. Sands’s thesis, she tested claims made by Artistic Directors that they simply receive more male-written scripts for consideration (claims that female playwrights have been reluctant to believe) by looking at Doollee.com, a database of over 80,000 plays and 20,000 playwrights. Included in the records for each play is information about such things as whether or not it has been produced and if the playwright is professionally represented. It’s an admittedly flawed resource, so Sands obtained a second database of plays and playwrights from the Dramatists Guild. She found both data sets to show basically the same thing and therefore merged the findings.

Artistic Directors are right. Fewer women than men are writing professionally. (This was also corroborated by what we learned “off the record” from some of the top graduate playwriting programs: approximately 45 percent of applicants are female and 55 percent are male.) In addition, women playwrights were found to have written fewer scripts than their male colleagues.

It turns out actresses are right too when they make their claims that there are fewer roles for them. Although Doollee doesn’t specify whether a play has a female or male lead, it does count the number of male and female roles. And it shows that of plays written by men, 81% have a majority of male roles and 19% have a majority of female roles. You’d expect female written plays to be the mirror opposite, but only 33% of women’s plays had a majority of female roles, 67% had a majority of male.

Remember, Ms. Sands found in the audit study that the female respondents believed the scripts most likely to face discrimination were those by women with female protagonists. Female writers obviously agree and are protecting themselves from the worst discrimination by limiting the amount of estrogen in their work. In contrast, the Doollee study showed that a play written by a man with mostly female characters has a much better chance of being produced than a play by a woman with mostly female characters. Therefore, the population with the most innate impetus to write for actresses is being economically discouraged from doing so. The root of the actresses’ problem is the same as the root of the female playwrights’ problem, the discrimination against plays by women about women. (Please pass this information on to powerful actresses.)

The kicker is that, despite such bias, plays with female protagonists do not appear to be less economically viable or less likely to win critical acclaim than those with male protagonists. You don’t need to be a Princeton economist to figure this out. Look at the past ten years worth of TCG’s Top Ten Most Produced Plays in the American Theatre and count the number of male vs. female protagonists for the two most widely produced plays each year. Out of the 27 plays (the math doesn’t add up due to some ties), fourteen have female protagonists, seven have male protagonists and the rest are ensemble works. Seven out of the past ten Pulitzer Prize-winning plays had female protagonists. There’s ample evidence that both award committees and the ticket buying public (by every estimate at least 60% female) find female characters extremely likeable, whether written by men or women.

The Broadway Study

In the third portion of Ms. Sands’s thesis, she followed the money. She is an economist, after all. She went straight to Broadway and left behind such subjective notions as “likeability” and “artistic exceptionalism” and focused on more objective criteria such as weekly revenues, ticket prices, and length of a show’s run. First she compiled a list of all the new plays or musicals produced on Broadway between 1999 and 2009 and then identified the gender of the playwright or bookwriter. She found that women exclusively wrote only 11% of the shows. And that the shows written by women earned an average of 18% higher revenue and sold 16% more tickets, weekly. And yes, she controlled for the type of play (i.e., straight, musical, or one-person show) due of the large variations in production costs across play types. She controlled for outliers, i.e., the one big hit that could tip the results unfairly. She controlled for all sorts of things and her methodology has been vetted by all of her numerous advisors, who happen to be some of the most prominent economists at our universities and in our government. Sands used numbers that came from BroadwayLeague.com and looked at them every which way, but nothing changed the fact that the female-written plays on Broadway were more profitable than their male-written counterparts. And even more remarkable is this paradox: even though the female-written plays made more money and sold more tickets, they ran for the same amount of times as the less-profitable male-written works.

This finding has caused the most head scratching and skepticism. If those female-written shows were so profitable, why wouldn’t the producers keep them running longer than their less profitable male counterparts? We offer one possible explanation: Imagine there is a show with a bankable star whose contract is expiring. The producers must predict whether or not the play will continue to sell tickets without that star. Based on the evidence, we can guess that when a play has a male writer, producers are more likely to recast the role and keep the show running. But when a female has written a show, they may be more likely to assume that the high ticket sales were due to the departing star and not the writer’s work. The producers’ prophecy that there will be a revenue drop when the star leaves the show becomes a self-fulfilling one when they close it down.

The higher relative success of work by women is not in any way, shape, or form proof that women are better writers than men. Rather it suggests that the bar is set markedly higher for female writers to be produced. A young male writer who shows promise receives a production, and then a few more, and then he writes a hit. He develops his craft along the way. A young female writer who shows promise, however, is not as likely to be produced until she writes a hit. She must come to the table with her craft highly developed. It’s a Catch-22 that catches all female playwrights. And it explains the statistics that started this whole thing. Female playwrights are twice as likely to land on TCG’s list of the Top Ten Most Produced Plays in the American Theatre as their percentage of productions suggest they should, because the American theatre doesn’t take as many “risks” on women writers. This is corroborated by a quick accounting of last year’s season in New York: out of all the “unknown” writers (i.e., playwrights receiving their first major production), only around 10% were female. The bar is clearly set much higher for women, and there aren’t very many reasons to hope that future conditions will improve.

There are basically two ways to interpret all of this. The first is that women don’t want to become playwrights as much as men and/or they just aren’t as good. The second is that female playwrights are what economists call “discouraged workers” – people who drop out of the labor market because of an inability to find suitable employment. One can certainly recognize and acknowledge this byproduct of bias when it’s displayed in other professions and against other groups, so why not with female playwrights? Women ARE as good at playwriting as men, but not as many are writing because as difficult as it is for men, it’s far more difficult for women to eke out a living in the theatre.

If you subscribe, as we do, to the second interpretation, then what is to be done? We’re glad you asked. First, a concerted effort must be made by Artistic Directors to find and develop female writers with the same eagerness and enthusiasm they do with male writers. That means they are going to have to eliminate the de facto quota of twenty percent or so of production slots that, in practice, have been set aside for women and writers of color to compete for. They also need to put out the call to schools, agents, and their own literary departments, and then actually read all the scripts by women that come in. And since women inside the theatres have essentially reported that they perceive bias in their institution’s play selection process, Artistic Directors must make sure they create environments in which these perceptions can be freely addressed and satisfactorily handled. And the women need to acknowledge the possibility that they engage in prophetic discrimination, subconsciously or otherwise. Basically, if you read and like a script by a woman but think it’s not a good “fit” with your theatre, pretend it was written by someone named John and read it again. Lastly, both Artistic Directors and producers should take advantage of the free market research Ms. Sands has provided and acknowledge that her results make perfect sense. After all, everyone knows that audiences are predominantly female. So, start looking for more female-written plays with female leads. They are the least-produced and most successful plays around. We’ll even tell you where you can find them — in the hearts, minds and hard drives of female playwrights.

Discrimination and the Female Playwright was originally published in The Dramatist, September/October 2009. Reprinted by permission.