The social side of Sandy

Marvin Nauman/FEMA photo

Marvin Nauman/FEMA photo

When disaster strikes, we rely on our social networks for support. During hurricane Sandy, neighbors helped neighbors by sharing electrical power with those who’d lost it or removing tree limbs from each other’s rooftops. In many cases, the help we get during emergencies comes from whomever happens to be nearby, but more and more our cell phones—if they still work—connect us to more distant resources.

Northeastern professor David Lazer hopes to understand how people behave during disasters through an app he and his team developed for mobile phones, which archive much of that behavior in call and text logs.

The Sandy app is available through the website Volunteer Science, which Lazer and Drew Margolin, a post-doc in his lab, recently launched. It allows users like you and me to participate in research in the form games and apps like the one for Sandy. They hope to eventually see a huge flow of people through the site, playing these sorts of games for fun. After all, the question of how people behaved during Sandy is just one of a limitless number that researchers could ask using this approach.

But back to the Sandy app. Once downloaded, it will ask volunteers to answer questions about their situation during the storm, what resources they needed or provided to others, and how they connected with those they did. It will then reach into the archives, choose ten random calls, and ask questions like “How is this person related to you?” and “Did you get what you needed from this person?”

Post-disaster interviewing is standard protocol for emergency organizations like the Red Cross, but these rely entirely on people’s memories, which are often skewed during times of stress. The new app is “not meant to substitute for other methods, it’s just adding a powerful complement,” said Lazer.

“We think of it as the 21st century improvement over the interview,” said Margolin, who is leading the project. The method will add valuable data that interviewing alone cannot. For instance, people who had a flood of inbound calls during the storm may have been support-hubs without realizing it.

Previous research has looked at static call detail records, provided by the phone company, which includes things like the cell tower that was accessed and the number that was called. Those studies have revealed distinctive patterns of behavior. For example, the number of calls between people 20 years apart in age increases during emergency situations, suggesting more calls between parents and children.

“But the problem is we can’t ask those people questions about their context,” said Lazer. They can’t ask whether the older person someone called was in fact their parent. “In some ways it’s like one hand clapping. Here we’re trying to supply the context and make a little bit more noise.”

While the app will provide invaluable new information for emergency relief efforts, it will also provide an improvement over current network science research methods. “Trying to build a bridge between the observational data, the massive passive data that’s being created about all of us every day and deeper sociological constructs,” said Lazer. That’s the goal for the whole Volunteer Science platform, he said.

Readers can participate in the Sandy study and encourage New York and New Jersey area friends to do the same by visiting www.volunteerscience.com. The app is currently only available for Android phones.