New system allows researchers to conduct studies while participants keep privacy

In today’s data-driven world, information is an important commodity. Data is collected about what people search for on Google, buy on Amazon and post on social media. It fuels the decisions that businesses make and is essential to important scientific research.

But as data breaches increase, people are becoming more concerned about who exactly has access to their data. And these concerns are putting researchers in a pickle.

Timothy Brick, assistant professor of human development and family studies, said researchers are generally encouraged to share the data they gather in studies so others can verify the study was done correctly or perform research of their own. But participants are increasingly concerned about whose hands their data will fall into.

So Brick is working on a system called Maintained Individual Data, Distributed Likelihood Evaluation (MIDDLE) that changes the way data is collected and reported. The system will enable researchers to conduct studies and participants to keep control over their data.

In a traditional study, participants volunteer their time, complete the necessary activities and give any required information. Depending on the study, this could range from getting an MRI to keeping a log of emotions at various points in the day. Studies that pay participants are compensating them for their time, but Brick said they’re usually not paid for something far more valuable -- their data.

“We’re not factoring in the value of the data they’re giving us. Essentially, that data is a donation, and people don’t always realize the value of that data,” Brick said. “We want to give people control and ownership of their information while still being able to use it in studies.”

Traditionally, when participants get their MRI or record their emotions for the day, that information goes straight to the researchers. But with MIDDLE, the information goes to a private data store, never to be seen by anyone but the participant. Instead, researchers will periodically send out different “models,” and the participants will choose which one best fits their data.

For example, say there is a study that examines people’s moods and exercise habits during the week. Every day, participants record how they feel using the MIDDLE app on their phones while their Fitbits help keep track of physical activity. At the end of the week, the study sends out a statistical model representing a hypothesis about how each participant’s moods changed throughout the week. The model might convey feeling three times better on the days a participant exercised, for instance. The app then tells the study how accurate the model is based on individual responses, and the study continues to refine its estimates and send updated models until it determines the best model for everyone.

Researchers don’t see when or how often participants exercised (or how good they felt after) but the study still benefits.

“Essentially, the idea behind the project is that if the data is about you, you keep it,” said Brick. “You keep the data and let the researchers know which model it fits best. That way we learn about what works best across the population, but we can’t find out very much about any individual person in particular.”

In addition to helping to maintain participants' privacy, MIDDLE has a slew of other benefits, including its convenient delivery system.

Although the system is still in development, Brick describes it as eventually being like “an app store for experiments.” As an app on your smartphone, MIDDLE would let you search, sign up for and participate in studies that interest you.

MIDDLE also gives flexibility to both the researcher and participants. If a participant wants to quit the study, he can do so at any point simply by clicking a button. Likewise, if a researcher finds he has enough information in half the time he expected, he can also stop the study early.

Brick said the system also will open the door for health or education studies previously made impossible or more difficult by the Health Insurance Portability and Accountability Act (HIPAA) and the Family Educational Rights and Privacy Act (FERPA).

“Information is fiercely protected by these policies, and rightly so,” said Brick. “MIDDLE could be a big help and offer an opportunity to study these industries without violating people’s privacy. The improvements to be gained could be huge.”

But while privacy is important, so is accuracy. Brick said that while some are concerned with trusting participants to accurately report their data via the models, it’s important to remember that that is not a new issue.

“Yes, for this to work, you have to trust people. But we do that anyway,” Brick said. “It’s not a new problem or concern. When you start a study or research project, you’re already trusting people to give accurate data about themselves.”

Brick said that while the system is going through some final tweaks, he expects it will be able to be tested by researchers soon. In addition to its other benefits, Brick said MIDDLE is a chance to do some good in the researching world.

“As researchers and professors, we need to do things that benefit people, and do them ethically,” said Brick. “MIDDLE really gives us that opportunity. It gives people ownership over their data — it keeps the value with you, but also lets science learn.”

For more stories about IT at Penn State, visit news.it.psu.edu.

Media Contacts: 
Last Updated July 28, 2017