Facebook Manipulates News Feeds of 689,003 Users for Psychological Experiment

Facebook conducted an experiment wherein they tweaked the posts displayed on their users' News Feed in an attempt to understand how our emotions were affected by posts made by other people.

The experiment ran for a week in January 2012, and it included at least 3 million posts that mentioned 122 million words. These words could be characterized as positive or negative. The researchers designed a special algorithm to categorize if a word evoked positive or negative emotions. They also tweaked the number of times that Facebook users were exposed to positive and negative words. That would explain why a person might have felt down while browsing his Facebook feed at that time.

Computers were used to analyze the data, not the scientists themselves. For those who were concerned about not being asked for consent, Facebook pointed out that all users agreed to participate in studies such as this one by agreeing with their privacy policy.

By altering what users see in their News Feed and analyzing over 700,000 posts of Facebook users, the researchers concluded that emotional stress could affect people even if they were not physically together.

California's Jamie Guillory and Jeffrey Hancock from the Cornell University concluded that users who were exposed to less positive messages would post less positive messages themselves. It also worked this way with users who saw too many negative posts.

The study proved that textual and verbal cues were enough to communicate emotions that transgressed physical borders.

"When positive expressions were reduced, people produced fewer positive posts and more negative posts; when negative expressions were reduced, the opposite pattern occurred. These results indicate that emotions expressed by others on Facebook influence our own emotions, constituting experimental evidence for massive-scale contagion via social networks," the researchers wrote.

Further details of the study were published in the June 29 issue of Proceedings of the National Academy of Sciences.

Real Time Analytics