In a report published at the Proceedings of the National Academy of Sciences (PNAS), Facebook data scientists conducted an experiment to manipulate the emotions of nearly 700,000 users to see if positive or negative emotions are as contagious on social networks as they are in the real world. By tweaking Facebook’s powerful News Feed algorithm, some users (we should probably just call them “lab rats” at this point) were shown fewer posts with positive words. Others saw fewer posts with negative words. “When positive expressions were reduced,” the paper states, “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 results shouldn’t surprise anybody. What’s more surprising, and unsettling, is the power Facebook wields in shifting its users’ emotional states, and its willingness to use that power on unknowing participants. First off, when is it okay to conduct a social behavior experiment on people without telling them? Technically, and as the paper states, users provided the consent for this research when they agreed to Facebook’s Data Use Policy prior to signing up, so what Facebook did isn’t illegal. But it’s certainly unethical.
Furthermore, manipulating user emotions in a digital space comes with uniquely disturbing consequences. In the real world, if you feel like the people around you bring too much negativity into your life, the solution is easy: Find a new crowd. But on Facebook, short of canceling your account, this is impossible to do if the company suddenly decides, whether as part of a research study or at the behest of certain advertising or engagement interests, to start sending more negative content your way. The whole point of the News Feed algorithm, to hear Facebook tell it, is to give users an experience tailored to their wants and interests. Clearly, that objective falls by the wayside anytime Facebook wants to turn its user base into a science experiment.
And then there’s the tone deaf gall of the whole thing: This research wasn’t uncovered by an investigative reporter, Facebook submitted the research to PNAS themselves. To make matters worse, there are questions about whether the methodology used was even sound. To determine “positive” and “negative” sentiments, the researchers used a technique called “Linguistic Inquiry and Word Count” or LIWC. But even the creators of LIWC admit that assessing its validity when applied to “natural language” (like a Facebook update) is “tricky.” LIWC’s reliability has largely been tested by analyzing essays, where there is more repetition than in natural language.
Our IP Address: