Engineer crunches literary data to seek truth about fiction

Katy Milkman found inspiration for her senior thesis by exploring the roots of American fiction, but thinking like an engineer.

Katy, who majored in operations research and financial engineering and earned a certificate in the Program in American Studies, took a course last spring on American short stories.

Elaine Showalter, now an emeritus professor of English, emphasized to the students that fiction writers often use themselves as models for their characters.

Searching for a thesis topic that would combine her interests in engineering and literature, Katy was intrigued by Professor Showalter’s theory.
“As someone who is used to thinking about things analytically and using statistics, the question arose in my mind whether this was something that could be quantified: Could you test the hypothesis that short fiction is autobiographical?” Katy said.

Based on Professor Showalter’s advice, Katy decided to pursue the question by researching stories published in The New Yorker, which she described in her thesis as “arguably the world’s compass for defining the best modern short fiction.”

She broadened the scope of her research to include questions about whether fiction published in The New Yorker is affected by editorial changes at the magazine, as well as which segments of society are represented in the stories.

With the help of faculty members in engineering and the humanities, Katy devised a statistical study of the 442 short stories published from October 1992, when Tina Brown’s much-hyped tenure as executive editor began, through September 2001. She decided to halt the study then because she was unsure how the Sept. 11 terrorist attacks would alter the magazine’s story selections.

Data collection
Katy collected demographic information about each of the stories’ protagonists and secondary characters—culling data on 34 variables, including age, race, gender, and country of origin—and collected the same information about each of the authors. The strong correlations between the characteristics of authors and their protagonists proved her main hypothesis.

In addition, she found some interesting trends within the demographic categories. For instance, male writers wrote about men more often than female authors focused on women, and white and Asian authors focused on characters of their own ethnic backgrounds more frequently than black and Hispanic writers did.

Katy suggested in her thesis that perhaps women or members of minority groups might feel pressure to write about male or white characters because of “real or imagined preferences of New Yorker editors and readers.”

“I always have wanted to do things that use both sides of the brain, and I don’t think I could have chosen a better topic,” said Katy, a native of Washington, D.C., who plans to pursue a Ph.D. in business. “It’s been perfect, in that it contains challenging math and challenging literary analysis.”

Summer reading
Katy began reading the stories and compiling data last summer, and she visited The New Yorker’s offices after returning to campus. In face-to-face or phone interviews, she spoke with several key figures at the magazine, including executive editor David Remnick ’81, who took over the magazine’s helm in July 1998, as well as current fiction editor Deborah Treisman, former fiction editor Bill Buford, and longtime writer and editor Roger Angell.

Two advisers
Reflecting the interdisciplinary nature of her thesis, she turned to René Carmona, the Paul Wythes ’55 Professor of Engineering and Finance, and William Gleason, an associate professor of English, as her advisers. Katy also communicated via e-mail with Joyce Carol Oates and Edmund White of the creative writing faculty to gain additional insights from authors.

Because he teaches in the Program in American Studies, Professor Gleason said he sees many theses that cross disciplines. But, he noted, “Katy’s thesis has certainly been one of the most innovative in this regard, combining fields that are not thought to be particularly adjacent.”

“When I got drafts back from my advisers, it was funny to see how completely different the comments were,” Katy said. “One was talking about hypothesis testing, and the other was talking about how to better justify my decision to focus on short fiction. Both of them were really enthusiastic about this thesis.”

Bonus information
In addition to her primary hypothesis, Katy gathered statistics to answer two secondary questions. She determined that the switch in the magazine’s fiction editor from Charles McGrath to Mr. Buford in 1995 led to significant changes in story variables such as narrative voice, length, gender of authors, and characters’ age, country of origin, and religion.

She also found that the stories’ settings were highly correlated with the geographic distribution of the magazine’s subscriber base but not with that of the U.S. population, suggesting that its story selection “may be more market-driven than its editors are aware [of] or will admit.”

Professor Carmona, her co-adviser, said that Katy’s choice of such a nontraditional subject for an engineering thesis could have created obstacles to achieving statistically significant results. But he said Katy skillfully completed her statistical tests, and “the various sources of ‘noise’ in the data she collected did not conspire to hide the facts. Her gamble paid off.”