Can you blame Facebook for your bad grades? Maybe.

It’s common to use Facebook as a scapegoat for poor academic performance. That’s because a few small studies have shown that grades are lower among students who spend more time on the social media site. The assumption has always been that more time spent on Facebook translates to less time spent studying, which leads to lower grades.

But a newer, bigger U.S. study has found that Facebook time and study time are only weakly related. It takes many extra hours of posting and chatting before grades start to slip. What’s more, although the new study found negative relationships between grades and certain types of Facebook activities, other types of activities appear to be a associated with higher grades.

Study participants were given a long list of potential Facebook activities, including tagging photos, updating their own statuses, creeping on other people’s pages, chatting, linking and more. Students rated how often they do each activity. They were also asked how much time they spend studying and their grades were determined through registrar’s offices.

The results show that sharing links and checking out what friends are up to (also known as creeping) are related to higher GPAs, while obsessively posting one’s own status, using the chat function and logging-on more often are related to lower GPAs.

And while more time spent on Facebook in general is correlated with less time spent preparing for class, the relationship is weak. The study’s model found that a student who spent four hours and 40 minutes more than the average student on Facebook in a week would only spend an hour and forty minutes less than the average student studying that week.

“Facebook use in and of itself is not detrimental to academic outcome,” researcher Reynol Junco of Lock Haven University of Pennsylvania told Mashable. “It depends how it’s used.”

To understand why there’s so much interest in how Facebook may effect grades, consider how much time students in the study report spending on the site—on average 106 minutes daily.

The authors suggest their work is more reliable than past studies, as the sample size is large (n=1,839) and they controlled for intelligence by using high school grades, in addition to gender, ethnicity and socioeconomic status. The study was published in Computers in Human Behavior.