How Has the MFA Changed the Contemporary Novel?

We wrote a program to analyze hundreds of works by authors with and without creative-writing degrees. The results were disappointing.

BrAt82 / charles taylor / Shutterstock / Kara Gordon / The Atlantic

This year, about 20,000 people applied to study creative writing at MFA programs in the U.S. It’s a funny fact to consider, given that the idea creativity could be taught used to be widely mocked—the literary scholar John Aldridge once said the programs produced “clonal fabrications of writers.” For a time, MFA programs were oddities on college campuses: In 1975, only 52 existed. Much of this has changed in the last two decades. Today, there are more than 350 creative writing programs in the U.S. alone, and that number doubles if you include undergraduate degree programs.

The rise of the MFA has changed how both writers and people in general talk about creativity. The debate has shifted from whether creativity could be taught to how well it can be taught and whether it should be taught. The stakes are real: Creative writing has become a big business—it’s estimated that it currently contributes more than $200 million a year in revenue to universities in the U.S.

Today’s debate falls along predictable fault-lines: One side eyes the teaching of writing suspiciously, and concludes that MFA programs may produce some good fiction, but they don’t produce enough “great literature.” The other side defends the institution by saying, if nothing else, that programs give aspiring writers the time to “dedicate oneself” to the craft of writing. But there’s an underlying assumption that the MFA does something. There’s a widespread belief that if you get an MFA, at the very least, it will change your writing in some discernible way.

But what if there’s no change to speak of? Is it really possible to tell the difference between novels that have been through the meat-grinder of the MFA and those that haven’t? What if this is just something that’s been imagined into existence, by both detractors and supporters alike, to satisfy a collective need to believe that institutions can improve anything, even creativity? Or conversely, that institutions ruin everything, especially creativity? Whether you valorize the Romantic ideal of the lonely, humble artist or the neo-liberal belief that education can solve any problem, the MFA has become a kind of Rorschach test for how writers and critics feel about creativity, where it comes from, and how best to nurture it.

Until now, no one has used much evidence beyond the anecdotal to test whether or not the MFA has actually influenced the contemporary novel. What if this debate, furious as it is, is just a distraction from more important questions surrounding creative writing, like problems of diversity within publishing or financial exploitation on the part of universities?

We’re two professors of language and literature who regularly use computation to test common assumptions about culture. So we decided to examine to what extent writing from MFA graduates differs from writing by non-graduates. We collected a sample of 200 novels written by graduates of MFA programs from over 20 leading programs (including Columbia, University of Texas at Austin, Iowa, and others) that have been published in the last 15 years. (This sample includes authors like Rick Moody, Alix Ohlin, and Ben Lerner.)  For the sake of comparison, we also collected a similarly sized group of novels published over the same time period by authors who haven’t earned an MFA degree (including writers like Donna Tartt, Miranda July, and Akhil Sharma). To make these two groups as comparable as possible, we only gathered novels by non-MFA writers that were reviewed in The New York Times, which we took as a mark of literary excellence. Using a variety of tools from the field of computational text analysis, we studied how similar authors were across a range of literary aspects, including diction, style, theme, setting, and even how writers use characters.

Needless to say, novels consist of much more than just these features. What makes a single novel a great novel, what makes, say, Junot Diaz sound like Junot Diaz, is of course mostly immeasurable. But these features remain the fundamental building blocks of any novel, so if MFA writing were in aggregate to have some essential difference from books written by authors without MFAs, it should be perceptible at the very least at this genetic level of prose. There has to be something that makes them different, and those differences, according to the vigor and tenacity of critics’ claims, ought to be recognizable. As Mark McGurl, the author of the sweeping history of the MFA, The Program Era, writes, creative writing programs “obviously” teach writers how to become a specific “creative type.” Or as Chad Harbach has argued more recently in his popular essay “MFA vs. NYC,” “the university now rivals, if it hasn’t surpassed, New York as the economic center of the literary fiction world.” If there are indeed “two literary cultures” in Harbach’s words, we should be able to detect it.

We began by looking at writers’ diction: whether the words used by MFA writers are noticeably different than those of their non-MFA counterparts. Using a process known as machine learning, we first taught a computer to recognize the words that are unique to each of our groups and then asked it to guess whether a novel (that it hasn’t seen before) was written by someone with an MFA. When we did this, the computer was successful only about 67 percent of the time at guessing correctly. You don’t need a degree in statistics to know this isn’t very good—you can be right 50 percent of the time just by accident. To put this number in context, with the same procedure we can predict bestselling novels about 82 percent of the time or whether a novel is a mystery or romance 85 percent and 95 percent of the time, respectively.

Nevertheless, there are some words that are different, but given that we’re talking about over 200,000 unique words, this is hardly surprising. For example, MFA novels tend to focus more on lawns, lakes, counters, stomachs, and wrists. They prefer names like Ruth, Pete, Bobby, Charlotte, and Pearl (while non-MFA novels seem to like Anna, Tom, John, and Bill). But on the whole, these distinctions look pretty meaningless; the words that appear more often in MFA novels don’t seem to be related to each other in a significant way. To test whether this was the case, we used a method called topic modeling that examines themes instead of individual words. And while MFA novels do tend to slightly favor certain themes like “family” or “home,” overall there’s no predictable way these topics appear with any regularity in novels written by creative writing graduates more than other people who write novels. To sum up: So far, no real difference between MFA and non-MFA works.

How about style? Surely, we thought, there should be some stylistic differences between these novels. The way writers put their words in order, that special MFA voice, should be detectable at some level. As one brochure has it, the goal of the adjunct faculty of an MFA program is to “work closely with their students to help them develop their own voices, styles, and form.” Presumably upon graduation those voices should be discernibly different than what’s already out there on the market. However, taking syntax as a measure of style—if we see style as the way writers sequence their words, the way they put their sentences together—we saw little difference between the two groups. MFA novels tend to use pairs of adjectives or adverbs less often, or avoid the more straightforward structure of a noun followed by a verb in the present tense. But other than that, there’s nothing detectably unique about the so-called “MFA style.”

So far, nothing. No real distinctions at the level of language, themes, or even syntax. When we went further to test whether the way writers constructed their characters was any different, once again nothing significant showed up. It was extremely difficult to separate the MFA and non-MFA writing groups in any meaningful way. If these results seem unbelievable, we shared this feeling as we carried out our tests. Our starting point was that there must be some mark of distinction. Why else were critics like Elif Batuman saying things like “the creative writing program has exercised the single most determining influence on postwar American literary production”? Why else were people paying for these schools?

Contrary to the critics, many top MFA programs explicitly state that they’re “doctrine free” and allow students to develop their writing “on their own terms.” They do not, they claim, actively try to make their students sound any particular way. As the University of Texas program says, “The best thing we do for fiction writers at the Michener Center for Writers is leave them alone.” But then why go? If a program isn’t going to train you or change you in any significant way—and the data suggest that by and large most don’t—then the costs of that investment start to seem deeply questionable. According to the latest research, only 7 percent of MFA graduates are fully funded, which means 93 percent are investing some portion of their own money to sound like everyone else.

Some might say that’s precisely the point. The MFA isn’t about developing a unique style at all, but about learning how to sound like already published writers. It’s about gaining entrance to the club. Look closely at the promotional materials of creative-writing programs and you’ll almost invariably see a host of proper names—these are the people with whom you can expect to rub shoulders, if not directly, then by association through the former graduates that have passed through the program or the mentors of your mentors whose influence will surely rub off on you. It’s about having the opportunity to insert yourself, however virtually, into that literary social network.

But this absence of distinction also has its hidden costs. Things begin to cut more deeply, for example, when we look at issues of gender or race. A major claim of the MFA is that it not only helps an aspiring writer find his or her voice, but it particularly helps minority writers discover some authentic self through the process of writing. As McGurl has shown, “find your voice” was a mantra at Iowa in the 1960s, and starting in the 1970s, it took on particular significance for writers of color. Programs like Iowa trumpet their success in training writers like Margaret Walker and Sandra Cisneros.

But when we refined our tests to look at how race factors into the results, we found the opposite to be true. We took each separate body of work—books by MFA writers and books by non-MFA writers—and compared all of the writers in each individual corpus along the metrics of diction, style, and theme we describe above. For both corpora, we expected white and non-white writers to group together in clusters, and we anticipated that non-white writers would especially group together in the MFA corpus (authors like Tayari Jones, Chieh Chieng, and Daniel Alarcon). But we found no such thing. Again, based on diction, theme, and syntax, these two groups, in both MFA and non-MFA writing, are impossible to distinguish.

The MFA promises to make the distinction of race come alive, take on literary heft, through learning how to write and the work of writing. But we have no evidence that MFA authors are any better at this than their less educated non-MFA peers. If there’s a quality that distinguishes a writer as Asian American or black, we could not find it. Junot Diaz has argued that MFA programs are “too white” and reproduce the “dominant culture’s blind spots and assumptions about race and racism.” It’s a claim that fits in with our algorithm’s inability to tell apart works by non-white writers and white writers.

But this erasure of voice gets an even more negative spin when we look at gender. A second major claim of the MFA is that getting an education in writing is an enlightening experience, and a key part of this enlightenment we can assume is learning how to challenge society’s gender norms. Many MFA programs, like the universities they are a part of, say they actively promote a culture of challenging “patriarchy” and “heteronormativity.” Cornell’s MFA program, for instance, celebrates the gender diversity of its faculty, which is “evenly split” between men and women. We’d expect MFA writing to actively resist gender stereotypes, especially given that MFA graduates skew overwhelmingly female (about 66 percent of MFA grads are women, which is about 10 percentage points higher than for the master’s degree more generally).

Once again, the data tell a different story. The percentage of male protagonists in novels written by MFA grads is well over half, at 61 percent, while that figure is 65 percent for non-MFA novels. Further, if a novel has a female lead, the chances that it has two strong female characters is only 32 percent for both MFA and non-MFA novels. Last, the percentage of novels that have a majority of male characters in the non-MFA group is 99 percent, whereas it is 96 percent for MFA novels. These are terrible numbers by any standard. They suggest that the contemporary American novel is disproportionately preoccupied with the experiences of men. And they suggest that the MFA novel is only barely better than its non-MFA counterparts. It’s possible that MFA writers have found more subtle ways to create strong female characters that go beyond simple numerical representation. But the raw numbers are damning: MFA writers are no better at representing women, and both groups are downright bad at it.

These results are hard to square with the increasingly amplified discourse that surrounds the MFA, whether for and against. While something may happen in MFA programs, perhaps that thing is more behavioral than artistic. When we look at the data, the MFA seems to be helping people sound like everyone else. To put a positive spin on it, we could say the degrees help writers fit into the literary landscape. Like the universities to which these programs belong, the MFA may offer a way of gaining entrance to an elite club. You learn the rules of the road, at least as defined by the publishing industry and literary reviews. At its worst, it doesn’t do anything at all.

The intensity with which readers and critics feel and think about the MFA, we might assume, has become disconnected from its moderate-to-minimal effects on the literary landscape in America. So it seems to us that the MFA doesn’t merit many of the hyperbolic claims about its impact on literature. $200 million per year, after all, is a high price to pay for very little measurable impact.