“It’s very counter-intuitive to say at this stage, but the fact is, no one really knows what the heck people are seeing on their screens,” said Byron Reeves, a professor of communications at Stanford University in California, and an author on the paper. “To understand what’s happening, we need to know what exactly that is.”
Researchers have linked daily time spent on specific platforms, like Facebook, to measures of well-being and mental health. But to build a more compelling understanding of the effects of digital experience, they’ll need far more, the new paper argues. Scientists need to look over people’s shoulders, digitally speaking, and record everything, on every device, that an individual sees, does, and types. The researchers call this ultra-fine-grained record a “screenome”, adapting the concept from “genome”, the full blueprint of one’s genetic inheritance. Each person’s daily screenome is similarly unique, a sequential, disjointed series of screens.
“The point is, your thread is yours, mine is mine, and we use it to regulate our emotions, to balance facts with fun, in our own idiosyncratic way,” said Reeves, whose colleagues on the paper included researchers from Penn State University, Boston University, Apple and Toyota Research Institute. “We are not beholden to media companies to organise or direct what we do.”
In arguing to develop such an approach, the researchers presented the digital threads of several dozen people, recorded with consent: screenshots taken every few minutes for periods ranging from a day to several days. Those records showed that people switched from one screen activity to another continually, every 20 seconds on average, and rarely spent more than 20 minutes uninterrupted on any one activity, even a full-length movie.
One participant’s digital thread revealed when during the day her screen use was most- and least-concentrated, and where she was during those periods. Another subject’s record made clear why he stopped reading a news story about a couple being dragged off a United Airlines flight and switched to another site: in order to confirm his own reservations on United for an upcoming trip.
Perhaps most intriguing, the paper presented colour-coded graphs of the digital threads of 30 university students, monitored over four days. The graphs revealed wide differences in what people used their screens for, as well as in their patterns of switching from one kind of activity, like email, to another, such as entertainment or news. Some people sprinkle brief periods of work between huge chunks of streaming movies and YouTube, for instance; others appear to be bouncing between email, work, and news sites compulsively.
These patterns can vary day to day, of course, for any of us. The deeper question for researchers, and one which they have not had an easy way to study, is how these shifting patterns shape daily experience. The most commonly cited downside of excessive screen time is low mood, or depression. In a recent study, researchers led by Johannes Eichstaedt of the University of Pennsylvania examined (with permission) the Facebook activity of 114 people diagnosed with depression. Using machine-learning algorithms, the team analysed the content of the users’ posts from the months and years before receiving the diagnosis, and compared these to the posts of similar people who did not go on to develop depression.
The analysis found differences in how frequently certain kinds of words appeared. For instance, people who later received a depression diagnosis talked about themselves on Facebook measurably more often than people who did not develop the mood problem. The analysis, while small by big-data standards, was the first to link to diagnoses in medical records, and it solidified previous correlations between online language content and low moods.
“This is a well-documented process, that suffering generally contracts focus on the self, whereas mental well-being extends focus beyond the self,” Eichstaedt said.
The researchers found that, by analysing Facebook language in this way, they could predict whether a person was on their way to being diagnosed with depression about 70 per cent of the time. “That’s about the rate you get with clinical questionnaires, and we haven’t been able to do better so far,” he said.
The New York Times