Electoral expert will be CBS pundit

In Stephen Ansolabehere’s sunlit, minimalist Cambridge Street office, there’s a wide, wall-high shelf of books — not a remarkable circumstance for a Harvard professor.

But alongside weighty tomes on statistics, history, and political science is a toy made of Popsicle sticks. If you pull the bottom rails back and forth, an angel trades punches with a devil.

That may sum up what interests Ansolabehere the most: elections, the fightlike contests that are so often cast as battles between good and evil.

Ansolabehere’s academic career spans more than two decades, but every year has been marked by a fascination for the electoral give-and-take. “If you can vote for it,” he likes to say, “I’m interested in it.”

With a 1989 Ph.D. in political science from Harvard, Ansolabehere is among a small group of U.S. social scientists who study elections, using novel survey methodologies and intricate statistical formulations to tease out evaluations of American voting systems.

For one, they mine data from a robust Internet project called the Cooperative Congressional Election Survey, a collaboration involving 40 U.S. universities that in 2006 polled 24,000 voters.

Wry and slight, the 46-year-old Harvard professor of government is a nationally recognized expert on opinion polls, elections, and elections technology. This Tuesday (Nov. 4), he’ll be in Manhattan as an analyst with CBS, helping to parse election returns during the McCain-Obama matchup.

That’s a long ride from his teenage years in Reno, Nev., where Ansolabehere first joined the rough-and-tumble of politics in high school, taking the helm of a campaign or two. (It surprised him even then, he said, “how much effort people put into this, given how low the stakes were.”)

Home also provided a political charge. Ansolabehere’s father, an auto mechanic and shop teacher, was active in labor politics and the Democratic Party. With five boys at home — and sequential households in California, Nevada, Illinois, and Michigan — his mother played a political role too, said Ansolabehere: “She was a referee.”

At the University of Minnesota, the budding political operative spent his extra time knocking on doors and getting out the vote in a variety of mayoral and state contests. “It was almost recreation,” he said.

As a sophomore majoring in political science and economics, Ansolabehere started a five-semester track as a research assistant to Frank Sorauf, an expert in campaign finance.

His doctoral studies at Harvard, with Morris P. Fiorina, were interrupted by a one-year fellowship at the Stanford Graduate School of Business. In company with political scientist David W. Brady and others, said Ansolabehere, “We just wrote papers.”

In the classroom at Harvard, he mastered probability theory and the other statistical tools needed to evaluate polls and election results. But it was the research and writing experiences outside class, said Ansolabehere, that best prepared him for the world of political scholarship.

With his Harvard dissertation still to write, Ansolabehere took a job at the University of California, Los Angeles, where he taught for four years and completed his first book, “The Media Game: American Politics in the Television Age” (1993), co-authored with Roy Behr and Shanto Iyengar.

After a year at the Hoover Institution at Stanford, Ansolabehere in 1995 took a job at the Massachusetts Institute of Technology (MIT), where he taught until early this year.

MIT was the springboard for his last two books. “Going Negative: How Campaign Advertising Shrinks and Polarizes the Electorate” (1996), co-authored with Iyengar, earned the coveted Goldsmith Book Prize.

“The End of Inequality” (2007), written with James M. Snyder Jr., argues that redistricting since 1962 has changed the face of U.S. politics — shrinking the authority of rural districts and bringing urban centers to the fore.

From 2000 to 2002, Ansolabehere churned out two dozen papers as a Carnegie Scholar — among them, studies of voting machines, race, Congressional roll-call voting, and campaign finance. (This last a subject he now says will fill out a future book.)

In December 2000, Ansolabehere helped found the Caltech/MIT Voting Technology Project, a research collaboration that evaluates the reliability of U.S. voting systems. Its inspiration was the contentious Bush-Gore contest just a month before.

In those days of “enormous frustration,” the Caltech/MIT project gave a place for the frustrated to go, said Ansolabehere, who co-directed the group until 2004. (The project influenced the 2002 Help America Vote Act, which called for replacing punch card systems and establishing minimum federal election standards.)

By January 2001, Ansolabehere was in Florida to witness the gathering political, legal, and technical scrum over voting technology — “a great moment, when social science and science could step in,” he said. “If we can send a man to the moon, we can fix the voting machine.”

Election officials were eager to go all-electronic, but Ansolabehere advised them to take it slow. Security is not the main issue with computer-based hardware, he said. System maintenance is — and how well that is done is a function of size.

There are about 5,000 local election offices in the United States, but only 500 or so are highly bureaucratic, well funded, and well staffed — the framework, Ansolabehere said, that is essential for maintaining large-scale computer systems.

Internet-based voting systems are not secure enough for widespread use, he said, and at the other end of the technology spectrum, paper ballots have their own problems. They are easy to retrieve, in case of controversy, but hard to count accurately. (Imagine doing hand-counting in Los Angeles, said Ansolabehere, or Miami-Dade.)

Until a revolution in computer design comes along, said Ansolabehere, the best voting technology may be optically scanned paper ballots. They’re retrievable and easy to count accurately.

Voting systems today are generally more accurate and fair than they were even eight years ago, he said, when at least four states used inadequate voting technologies: Florida, Iowa, New Mexico, and Wisconsin. (Nationwide, about 2 million disputed ballots were thrown out in the 2000 presidential election.)

Voting technology in place now will make a disputed McCain-Obama contest unlikely, said Ansolabehere. But he has set aside his November calendar in case he has to step in as an academic referee.

In the meantime, there’s always life outside the book-lined office. Ansolabehere lives in Newton, Mass., with his wife Laurie Gould M.B.A. ’96, who develops low-income housing, and their two daughters, ages 10 and 15. On the average day, he swims 3,000 yards — a task that leaves him only enough mental time for counting pool lengths (60).

Ansolabehere also paints, draws, and dabbles in lithography. “I love doing art,” he said, imagining life outside of regression analysis, warring angels, and the whirl of elections. “You feel your brain switching gears.”