When asked, students said they spent an average of 149 minutes *per day* on Facebook, but monitoring software found that this was a gross overestimation.

Reynol Junco

People are notoriously unreliable sources of information about their own habits. One study, performed in 2003, found that 50 percent -- 50 percent!! -- of self-reported non-smokers being treated for head and neck cancer "were in fact smoking actively," their lie revealed when scientists looked at their breath and blood for chemical clues.

But nevertheless, researchers often have to rely on self-reporting -- whether the question at hand is health habits or social-media use. And when it comes to the latter, people are apparently no more reliable than they are on questions of the former. A new study reports (pdf) found that college students estimated that they spent 149 minutes *per day* on Facebook, but when monitoring software was installed on their computers, the data revealed a much smaller number: just 26 minutes.

Reynol Junco, a scholar with Harvard's Berkman Center, says that the students in the study did have a pretty good sense of their relative Facebook use: Students who were heavier users estimated higher; those who were lighter users suggested as much. But ALL were wildly off when it came to the absolute estimate. (Though the number of students participating in the study was quite small -- just 45 -- Junco says the effect was so huge that the result is nevertheless reliable.)

Why were the students so off in their self-reports? Junco has four theories.

"It could be," Junco said to me, "that the self-report questions aren't specific enough." The question students answered was "how much time do you spend on Facebook each day?" which seems pretty straightforward -- except that that phrase "spend on" is actually a bit ambiguous. In focus groups after the study was complete, when Junco asked for their own theories about the results, "a lot of them -- and this was very surprising -- a lot of them said that they would take the question to mean how much time to do they spend *thinking* about Facebook," Junco explained. Spending time on something doesn't just mean the act of being on Facebook, but the act of expending mental energy on it. That ambiguity may have led students to overestimate their time on the site.

Another explanation -- one that Junco particularly likes, though he needs more research to support it -- is that students might have internalized societal expectations about their social-media use. "Society tells youths that they use technology a lot. They hear it from everybody. They hear it from the popular media; they hear it from adults; they hear it from their teachers," he said. "That would lead them, not very consciously but possibly subconsciously, to give very inflated estimates."

This might also be why so many people are surprised by the study's results: No one would expect people to very accurately estimate their Facebook time, but many would expect the discrepancy to run in the other direction. In other words, the assumption would be the college students lose themselves into Facebook, and spend way *more* time browsing it than they would self report. And this might be because people, like the students themselves, think that students use Facebook all the time. For this reason, it would be extremely interesting to see whether the effect was similar for other demographic groups.

The third explanation is the most practical, and is almost certainly a contributing factor -- and that is that students may be using Facebook from other devices, including their mobile phones and public computers that the software did not monitor. But would that explain the entire discrepancy? Junco doesn't think it would come even close, but he does write on his blog that "this is definitely an important facet to study in future research."

Finally, and most plainly, is just that students are bad at estimating how much time passes while they are online. Think about another thing we measure: distance. People have a pretty good sense of how far they have traveled based on how much time they've spent in the car. But with technology, we might have access to that kind of heuristic that would allow us to more accurately estimate how much time has passed.