Can a Computer Program Discern the Value of an MFA?

(screenshot via YouTube)

To MFA or not to MFA? The debate about the value of an Master of Fine Arts degree rages on in the art and literary worlds. Some see fine arts grad programs as an aspiring artist’s surest ticket into an elite club of sorts, as well as a crucial time to hone one’s craft; others see them as factory-like places for artists to navel-gaze as they prepare for lifetimes of debt. While there’s plenty of discourse, there’s been little quantitative research done to support either side of the argument — until now. 

A computer program has weighed in on the MFA debate. Over at The Atlantic, Richard Jean So and Andrew Piper, professors of language and literature who regularly use computation to study cultural trends, devised a program that analyzed and compared hundreds of novels by writers both with and without MFA degrees. Their aim was to gather data on how the rise of the creative writing MFA degree has affected the contemporary novel. In short? The researchers found no significant differences between MFA novels and non-MFA novels.

Since computers have no souls (yet) and can’t measure, say, the wondrousness of Oscar Wao’s life or the twee-ness of a Miranda July story, you might take this analysis of fiction with a grain of salt. Still, the researchers’ methods were rigorous. They studied a sample of 200 novels published in the last 15 years, written by graduates of more than 20 leading MFA programs. Writers included Rick Moody and Ben Lerner; programs included Columbia University and the University of Iowa. The researchers compared these novels to about 200 New York Times-reviewed novels by authors without MFA degrees, including July and Donna Tartt. Using a variety of tools within the field of computational text analysis, they studied diction, style, theme, setting, and the use of character in these novels. Summed up, the results revealed, as the authors put it:

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.

The distinctions the authors did find were fairly innocuous and highly absurd. In terms of specific words used, for example, “MFA novels tend to focus more on lawns, lakes, counters, stomachs, and wrists,” the authors write. “They prefer names like Ruth, Pete, Bobby, Charlotte, and Pearl (while non-MFA novels seem to like Anna, Tom, John, and Bill).”

Using syntax as a measure of style, the researchers found only that “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.” (Learning to avoid adverbs: worth $100,000 in student loans?)

Piper and So also studied novelists’ use of character in terms of race and gender. Their most significant and troubling finding may be that despite the fact that about 66% of MFA grads are women, 96% of the MFA novels the researchers sampled have a majority of male characters, a sign that the rise of the MFA isn’t doing much to diversify the range of human experience represented in contemporary fiction.

It’s hard to say what these findings mean, if anything, for the debate about the value of an MFA in the visual arts. What differences would a similar program find if it could study contemporary paintings by MFA-degreed versus non-MFA-degreed artists (besides, perhaps, more tears shed on canvases after brutal critiques)? Sounds like a job for the art world’s new computer vision technologists. 

Read Piper and So’s full account of their findings over at The Atlantic.