With social software appearing everywhere, the news seems to be more and more curious about social scientists who are working in the digital domain.

SFGate ran an article today entitled Probing tech’s heart: Social scientists seek technology’s human side, focused on Marc Smith (MS – Netscan) and Josh Tyler (HP – email networks).

Probing tech’s heart
Social scientists seek technology’s human side

Carrie Kirby, Chronicle Staff Writer
Tuesday, August 19, 2003

Just as Freudian shrinks asked about your childhood, information age psychiatrists are asking people to “Tell me about your e-mail.”

Patterns of online interactivity — from how many posts you make on a discussion group to how long you take to answer an e-mail — say a lot about the people typing the messages, researchers are finding.

Joshua Tyler, a 25-year-old researcher at Hewlett-Packard’s famed Palo Alto labs, is one of an army of scientists delving into the tech user’s psyche at major technology firms.

If you thought high tech was only about crunching code, you’re wrong. Squishy, left-brain science is slowly gaining its place alongside hard-core technology, as competitive tech firms try to get an edge on what their users are thinking . . . and therefore buying.

HP’s Tyler has spent hours sitting in workers’ cubicles, asking them why they replied to certain e-mails immediately and waited days to answer others.

“We found there’s a lot of subtle information conveyed in the timing of how people respond to e-mails. . . . People will deliberately delay responding to e-mail just to say, ‘Don’t expect fast answers from me in the future, because I’m a busy guy,’ ” said Tyler, who began his study of e-mail rhythms while pursuing his master’s degree at Stanford University and continued it after he was hired at HP Labs.

In his spare time, Tyler contrived another study of interoffice communication — he bought a toy blimp and created a computer navigation system that allows the dirigible to ferry Post-It notes from one cubicle to another. Off-the-shelf Web cams note the “helio courier’s” course and allow the steering computer to make adjustments.

“I would love to study how people use the blimp,” he said. “What kind of message would you choose to send by blimp?”

Although Tyler is a computer scientist, not a social scientist, he took a few courses in field research at Stanford, where he specialized in human- computer interaction.

In HP’s Information Dynamics Laboratory (www.hpl.hp.com/shl/index.html), he works on a team led by a physicist that also includes theoretical economists. It’s not uncommon at corporate labs for sociologists, psychologists and other social scientists to work alongside programmers, physicists and technology designers.

At Microsoft Research in Redmond, Wash., sociologist Marc Smith said it is sometimes lonely being one of the few social scientists in a business full of technologists. But he has been embraced by the tech crowd, he said.

“The technologists have been quite good at providing a level of respect,” he said. “They realize that the ultimate application of their work has got to be social computing.”

Take the wildfire spread of a simple little Web site called Friendster, where users write nice little endorsements of their buddies — and potentially find dates.

“People like computers because there are people in them,” Smith said.

In Redmond, Smith has been studying the online space Usenet, a vast network of discussion groups that predates the Web. By charting different types of behavior in Usenet groups, he’s able to steer users to the kind of group they want — not just a group that discusses the right topic but a group with the right goals and pace.

“Right now, Usenet is like trying to pick a restaurant blindfolded with your nose stopped up, and you can’t even hear,” said Smith, a fast-talking, bespectacled man who would look more at home in front of a college classroom than at a high-tech corporation. “You’re trying to feel your way around. Once you get into one, and they put some food in your mouth, what’s the chance you’re going to leave this restaurant? Low.”

In real life, indicators such as the number of people eating in a restaurant, the decor and the smells tip off the consumer, he said. In an effort to create such atmospheric cues online, Smith and his group have created charts that represent with big colorful bubbles how chatty, argumentative or helpful a given group is.

They do this without reading any of the words in the messages. It’s all based on the pattern of activity. People who post multiple replies on every discussion thread tend to be the arguers, the nitpickers. Those who post just one reply — especially if that reply ends the thread — tend to be the expert problem solvers.

The software Smith’s team created, known as Netscan, is available online at netscan.research.microsoft.com, and about 1,000 people a day use it to help them choose discussion groups that fit their needs.

Someone who wants to know how to configure a printer would probably choose a group with a track record of quick answers, while someone looking for entertainment might choose a group whose history is riddled with flame wars, or online arguments.

“It’s like tea leaves in the bottom of the cup. We allow people to stare into that and form your own judgments,” Smith said.

Like most researchers at corporate labs, Smith would like to see his work incorporated in a product someday. But it’s often difficult for any laboratory researchers to catch the eye of the suits who make the marketing decisions — and even more so for researchers who study social interaction instead of microprocessors.

There’s a busy freeway running between Microsoft’s main Redmond campus and its research lab, and the joke around the coffee machine there is that it’s tough to cross that freeway, Frogger-like, lugging servers full of data on your back.

Tyler, of HP Labs, has gotten suggestions from consultants at his company that one of his projects might be useful in the business world. The project tracked e-mail among hundreds of workers at HP Labs for four months, and charted the clusters of workers who frequently e-mailed one another.

The results look like a map of the night sky, with most of the constellations glommed together in the middle and a few scattered around the edges. Each constellation represents a group of people in heavy e-mail contact.

The study would be useful, Tyler said, “if you want to understand informal leadership structure, informal work groups. People have also said it might be used to track people involved in terrorist networks.

“I don’t know if it would really work for that,” he said, noting that there would certainly be privacy issues involved in such an application.

Tyler has fewer long-term hopes for his blimp messenger project.

“The engineering challenges are significant,” he said, looking over at the plastic helio courier, which had lost much of its helium and was now slumped in the corner of his cube. “One thing I discovered is that (toy) blimps can’t carry very much.”

E-mail Carrie Kirby at ckirby@sfchronicle.com.