Questioning the ‘Big Data’-Driven Decision

There was a time when we had to guess at a lot of things: Who’s buying what? Who’s traveling where? Which streets are busiest and at what times? They were educated guesses, but they were conjecture just the same.

Enter big data. The funneling of our lives into the digital realm has allowed us to collect vast swaths of data points that can be analyzed by computers to reveal patterns and trends—and even predict future behavior.

Dr. Regina Tuma

“Predictive analytics is another name for it,” says Regina Tuma, a doctoral faculty member in Fielding’s Media Psychology program. “For everything we do online, there’s a data point attached to it. They’re keeping track of everything. That’s the price we pay for going on the Internet.”

The information is used for lots of different purposes.

“In marketing, it’s being used to determine user tastes in order to niche market to the right consumer,” Dr. Tuma says. “We’re seeing it in the schools to gather data points on kids, from test scores to even what they eat and how that impacts learning. It’s used in law enforcement to look at neighborhoods and try to predict crime based on data points. It has emerged really as a way of making decisions.”

It makes sense: Better to make decisions based on fact than on hunches … right? But as a social scientist, something about it troubled Dr. Tuma.

“I was reading about how big data algorithms would determine your insurance rates and ability to qualify for mortgages, and I started thinking: How objective is this—and what are the tradeoffs? With politicians making decisions based on this, what does that do to democracy? And where is the discussion in all of this?” she says. “When we do research in psychology, we sample, we are objective, we describe how we select our sample, we detail our methods. None of this is
revealed with big data because it’s all proprietary.”

What’s more, the data is collected on such a huge scale and taken from so many samples that it’s often assumed to be infallible.

“My greatest concern about big data is it’s becoming a conversation stopper,” says Dr. Tuma, who jokes that nothing shuts up a room like the phrase “data-driven decision.” “But big data points are gathered to algorithms created by human beings, so our own biases come into the creation of these algorithms—and into how the data are interpreted, as well.”

To research the issue further, Dr. Tuma recently created a new course at Fielding: The Psychology of Big Data. “It’s a broader critical reflection,” she says, “asking, What sort of knowledge do we really get from big data? And can we trust the pictures it creates about ourselves and our world?”

Dr. Tuma has a PhD in psychology but has always been passionate about media and culture. She spent years “teaching in a media program and soft-pedalling the psychology, or teaching in a psych program and soft-pedalling the media,” she jokes. “I knew that Fielding, with its Media Psychology program, would make me whole.”

She was teaching the Psychology of Social Media when the big data course was born.

“We’re trying to layer the psychology into the big data movement,” she says. “We critique big data from a methodological perspective. For example, who has access to technology in society? Is the sample representative? Are we leaving out people of lower economic standing who are not using computers? Are we interpreting correlation as causation?

“What happens in a society that is under constant surveillance and how does that change our psychology? Do people react and try to subvertand outsmart the algorithm? That then begs a question about the truth value of big-data patterns.”

Students who understand the psychology of interpretation involved in reading big data outputs, and who’ve researched the socio-cultural aspects of an algorithm—namely, who goes in and who gets left out—are likely to be in high demand in the tech workforce, from government to private companies, she says.

“Our students really have an opportunity to go out there and make a difference,” Dr. Tuma says. “The goal is not to be anti-big data but to be able to interrogate it—and perhaps that’s where the knowledge really begins.”