We know that most individuals are exposed to fake news via social media. Scholars have suggested that the algorithms dictating what kinds of material we see on these platforms (elevating material we like, depressing material we don't like) may help explain this exposure.
I thought this was a fascinating suggestion and decided to test it. I had MTurk workers read 10 short stories that were either all fake or half real and half fake. Each of these stories were coded to be agreeable to Democrats, Republicans or a mix of both.
Treatment conditions attempted to simulate a filter bubble by showing some partisans a series of news articles that they should agree with (i.e., “filter bubbles”). I used the “mixed” or balanced conditions as a control. In these bubbles, I placed several pieces of fake news.
The results showed that those who read a series of articles agreeable to their political world view found the fake news stories more believable compared to those who read a more balanced mix of news items.
This suggests that the social media algorithms feeding us news items we like may actually be making the fake news problem worse. In other words, seeing lots of news items that make us comfortable also make us less critical of fake news when does slip into our news feed.
Results from a follow-up study suggest that being exposed to lots of politically-agreeable information encourages more heuristic information processing compared to participants who received a balanced diet of news and exhibited signs of systemic information processing.
However, there is an important caveat here: only Democratic and Democratic-leaning participants seem to respond to this filter bubble "break up" effect.
Compared to their Democratic peers, Republicans assigned to filter bubble treatment conditions believed fake news stories at approximately the same rate as their fellow Republicans receiving a more "balanced" diet of news items.
This suggests that if Facebook were to tweak the algorithm and show us more news from the other side's perspective, it may depress the reach and effectiveness of fake news. However, practically speaking, it may not have much of an effect on the overall fake news problem. Why?
Because we know from other research that the vast majority of those being exposed to fake news are Republicans. Filter bubble or no filter bubble, Republicans and those who lean Republican appear more inclined to believe these fake news stories compared to Democrats in the study.
You can follow @samuel_rhodes.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: