Can you predict a student's GPA? This app did.

Mobile app calculates grades using GPS, Wi-fi, sound, movement data

Researchers at Dartmouth College developed an Android app that tracked students' behaviors and predicted their cumulative GPAs within 17 hundredths of a point of their cumulative transcript GPA. 

Computer science professor Andrew Campbell led the SmartGPA study involving 30 Dartmouth students tracked across the 10-week quarter. The app uses automatically collected data on sleep, study, party, exercise, focus, and class attendance habits as determined by information from phones' GPS or Wi-Fi, motion detectors, and microphones. Researchers also incorporated students' periodic self-reports into their findings. 

For example, the app marks students as sleeping if it is nighttime, they are in their dorm rooms, not interacting with their phones, and the phones are charging. It records people as partying when they are in a Greek house they do not live in and ambient noise levels are high.

Previously, Campbell led a study on an app that used similar technology to predict mental health and academic performance.

In the most recent study, researchers found higher academic performers:

  • Felt an increase in stress leading up midterms followed by a decrease for the rest of the quarter;
  • Held shorter conversations in the evenings near the end of the term;
  • Studied more as the term progressed;
  • Had higher levels of happiness at the end of quarter; and
  • Were more conscious of their behaviors overall.

Among  some surprising findings:

  • Students who studied in louder locations had better grades—perhaps due to study groups; and
  • Better students had the ability to prioritize tasks, rather than just limit total party time.

"College life is complex," Campbell says. "There is no general agreement on why students with similar academic capability at the same institution do better or worse than one another."

Campbell says he hopes to make the "persuasive technology" available in the app store to help students stay aware of their behaviors, much like a fitness tracker for the mind. The data is sensitive, he says, but also useful and empowering.

Next, he plans to test the app on a more diverse, larger group of students at the University of Texas at Austin (Kamenetz, "nprEd," NPR, 6/2; Dartmouth College release, 5/25).

Thoughts on the story? Tweet us at @eab_daily and let us know.

  • Manage Your Events
  • Saved webpages and searches
  • Manage your subscriptions
  • Update personal information
  • Invite a colleague