After her freshman year at Portland's Reed College, Elizabeth Bruch got a job as a one-person census bureau for a community-based group in the San Francisco Tenderloin, for many years a red-light district populated by drunks and prostitutes. A tide of immigration had brought families into the area, but fearful of the area's dangers-not the least of which were the cars speeding through what they thought was a bad neighborhood-the families were barely visible. "There were tons of kids," Elizabeth says, "but they were all locked inside. It was a neighborhood about to burst open."

Going from building to building to count the kids, Elizabeth got to know everybody, even "the crazy guy on the corner [who] would stop screaming, smile, and wave" when she went by. With the help of the data her census provided, advocates were able to get funds to build an elementary school in the neighborhood.

Elizabeth returned to Reed with the knowledge that "statistics can have an impact on people's lives," and she was on her way to knowing what she wanted to do with her life: "study social problems and use that information to try to implement reasonable solutions," she says.

The social problem Elizabeth has decided to examine in her sociology dissertation is residential segregation, and the tool she is using is computer modeling. Her results test the theory of economist Thomas Schelling, whose model showed that if people simply don't want to be in the minority-that is, they'll stay in a neighborhood as long as their group holds even a slight majority-the eventual result is high segregation.

In her master's thesis in sociology, Elizabeth used data from the Los Angeles Family and Neighborhood Survey (LA FANS), which showed that people's residential choices were actually more nuanced. Her model indicated that decisions based on race/ethnicity alone would not lead to the highly segregated neighborhoods typical of Los Angeles. For her doctoral research, she's adding income and class to the input factors, reasoning that people's residential choices are in part determined by their financial means.

Her results are certain to be interesting, but so is the method she's using to get there. In Schelling's model, the artificial people lived on a computer grid and made either/or choices based on the racial mix of nearby squares. Elizabeth "wanted my artificial people to live on maps on actual city blocks," she says, and she created them to reflect a 5% representative sample of the area's actual residents.

Her adviser, Professor of Sociology Richard D. Mare, says: "Simulation without empirical validation can be an idle exercise, whereas data analysis without a conception of underlying social processes often produces only superficial conclusions. By combining these strategies, Elizabeth has the potential to produce social science at its very best."

Elizabeth brings to her work in sociology a growing expertise in statistics. Having always been interested in quantitative methodology, she decided to take some statistics classes when she had finished her coursework in sociology. "I had a good time," she says, "but more important, I realized how important it was as a sociologist not just to know statistics but to be able to communicate with statisticians."

At one level, Elizabeth is talking about understanding the language statisticians use: The statistics courses designed for graduate students are far more rigorous than those sociology majors typically take. However, there's also a social aspect to her statement. Because Elizabeth had met some statisticians, she knew where to go to when she couldn't find a way to check the fit of a computer model. "I dropped my pile of articles on [Professor] Jan de Leeuw's desk," she says, "and three weeks later, he e-mailed saying he thought he'd figured out how to do the necessary diagnostics." Together, they wrote a paper on their findings.

And knowing statisticians paid off even more grandly in her connection with Professor Richard Berk, whose statistics class she had taken. He sent news that the National Science Foundation "had a pot of money for work related to my dissertation on complex models of how cities change over time." As a result, Professors Berk and Mare, with Elizabeth, have a $280,000 grant for their research, and Elizabeth has a National Science Foundation Dissertation Improvement Grant.

The project asks "whether it is possible to simulate well the patterns of residential segregation as they have evolved in Los Angeles from 1980 to 2000," says Professor Berk. "Is it possible to develop a useful computer model of how people in Los Angeles choose where to live?" As "a first-rate social scientist and a first-rate applied statistician," Professor Berk says, Elizabeth is a valued partner.

She is also a first-rate builder of social networks, and this was true long before she started hanging out with statisticians. Her undergraduate adviser, John Pock, was also mentor to three members UCLA's Sociology Department, including Professor Mare. When Elizabeth hit a snag with her undergraduate thesis at Reed, Professor Pock suggested that she ask UCLA's Bill Mason for help. He invited her to spend some time in Los Angeles, and Elizabeth found it "a really exciting place, where "they took graduate students very seriously."

As she was making her choices about graduate school, UCLA's demography group was just getting started and the first interviews on the LA FANS were being done. Even before she was officially a graduate student, she became "immediately entrenched in this project" as a research assistant, she says. "I had the sense that I could be part of something new and exciting."

Probably next June, Elizabeth will receive a master's degree in statistics along with her PhD in sociology, her master's thesis being a methods chapter in her dissertation. She looks forward to an academic career, and she's already outlining her first course: teaching social dynamics using toy models students can play with to study everything from residential segregation to marriage patterns.

The same social attitude that took her to the Tenderloin nearly a decade ago remains, with a twist: "If you really want to help people," she says, "you have to change the structure of the society they're living in."

Published in Fall 2004, Graduate Quarterly