February 9, 2002

TESTING `SIX DEGREES' THEORY IN E-MAIL WORLD

You know all about ``six degrees of separation.'' Everybody does; it's the charming notion that any two human beings are linked by a chain of acquaintances no more than six hops long.

But is it true? No one really knows, though the idea has enjoyed popular currency since 1967, when psychologist Stanley Milgram wrote about it in the magazine Psychology Today. Belief in the idea is rather more robust than evidence for it. Now researchers at Columbia University are trying to test the theory in the e-mail world, and they'd like a few hundred thousand people to help them by starting a chain. You can sign up at http://smallworld.sociology.columbia.edu just by sending in your e-mail address. I did.

For their original ``targets,'' Duncan Watts, Peter Sheridan Dodds and Roby Muhamad chose roughly 20 people, mostly from among their acquaintances, though otherwise quite dissimilar. More than 1,000 people have volunteered to be targets, and in addition to Americans they now include a veterinarian in the Norwegian army, an unemployed man in Indonesia and a policeman in Western Australia.

If you register to start a chain, the program assigns a you a target at random, and tells you that person's name, city of residence, current employment and date and place of highest degree earned. Your assignment is to choose one person from among your acquaintances whom you think has the best chance of getting closer to the target than you are, and sending that person an e-mail message with a request to try to pass it on.

You can't just look up the target and send an e-mail directly -- that's cheating. But asking around among people you do know to see if they might be able to help and are willing to participate is fine, Dodds says.

``We think people are good at searching through their own social networks because they intuitively understand how connections are formed,'' he told me (by e-mail, of course). ``People are entitled to look after their messages.''

I registered, and by return beep I was assigned a target -- a reporter at Bloomberg News Service, the financial wire.

Aw, gee. That's way too easy. She's even had dozens of bylines in this newspaper, which subscribes to Bloomberg News.

The very next day I checked the stories we got from Bloomberg. She was a reporter on one of tham, and among her editors was someone I know because we served together on the board of directors for the Society of Porfessional Journalists a few years ago. Two hops, two degrees of separation.

There would be other possibilities. She writes about telecommunications, and because there's been a lot of telecom news out of Denver the past couple of years, we've undoubtedly interviewed some of the same people. But knowing that a chain exists doesn't mean a message would get through it. Executives of telecommunications companies are too busy trying to keep their companies afloat -- or swimming away from the wreckage.

``People not carrying on chains is probably our biggest problem,'' Dodds said. ``Our `attrition rate' has always been high, but it's gradually improving.''

High attrition was Milgram's problem too, which is why no one really knows whether his sweeping claims were justified. Judith Kleinfeld, professor of psychology at the University of Alaska at Fairbanks, has studied Milgram's papers, which are archived at Yale University. Her study is available on the smallworld Web site.

Milgram's first study involved 60 people in Wichita, Kan., recruited to start chains attempting to reach the wife of a divinity student in Cambridge, Mass. Only three chains were completed (and two of those came to the target the same way, through the wives of the deans of the Harvard Divinity School and the Episcopal Theological Seminary). One chain, highlighted in the magazine article, had only two intermediate links, but the average was eight.

Nor were the starters chosen at random; they were recruited through a newspaper ad that, Kleinfeld notes, ``appeals to patriotism and pride in social skills.'' The ad asked, ``Could you as a typical American, contact another citizen, regardless of his walk of life?''

In a second study with starters in Nebraska trying to reach a stockbroker in Boston, 44 of 160 chains were completed.

Summarizing a number of unsuccessful attempts to prove Milgram's thesis, Kleinfeld says that the evidence suggests the world is not so much small as lumpy, with groups that are well-connected internally but largely separated from each other. ``Social capital, the ability to make connections, is rare and more apt to be held by an elite.''

Is that still true in the age of e-mail? You can help find out, and if you do sign up, please let me know what happens.

(789 words)

(correction to improperly hyphenated e-mail address included with next column)