A study published in Computers in Human Behavior suggests that brief exposure to online misinformation can unknowingly alter a person’s behavior. The experiment found that reading a fake news article slightly altered participants’ unconscious behavior, as evidenced by a change in their performances on a test called the Finger Tapping Test.
Social media plays a central role in the exchange of information among the public, and its influence is only growing. People increasingly use online platforms to read and discuss news, and algorithms help tailor this environment to a user’s interests and behavior. Study author Zach Bastick says that these filtered environments risk creating a “distorted reality” whereby users are exposed to content that reinforces their views at the expense of alternate viewpoints.
Online platforms also facilitate the spreading of inaccurate information. Bastick was particularly interested in studying misinformation that is intentionally spread for malicious purposes such as political or financial gain. Some scholars worry that such misinformation — termed disinformation — can lead to changes at the individual level that add up to impact society and potentially undermine democracy.
Still, the extent that disinformation can unconsciously affect behavior is unknown, prompting Bastick to conduct his own experimental study.
The experiment involved 233 students between the ages of 17 and 21 who were attending a university in France. Bastick wanted to see whether reading a fabricated fake news article would alter the students’ behavior without them knowing it. As a measure of unconscious behavior, Bastick had the students complete a cognitive and motor function test called the Finger Tapping Test (FTT) — a test that asks a subject to repeatedly tap a computer key with one finger as fast as they can.
First, students completed the FTT as a measure of their maximum tapping speed (MTS). Next, the students were randomly assigned to one of three groups. One group read a positive false news item that reported that a fast MTS was a trait of successful people, associated with happiness and positive relationships. A second group read a negative false news item that said that a fast MTS was associated with deviance and brutality. A third group read a control text that did not mention tapping speed. After reading the article, students completed the FTT a second time.
Remarkably, the students demonstrated changes in their tapping speed that were associated with the false news text they read. Specifically, the group that read the positive false information showed an increase in their maximum tapping speed by about 5%. This was after taking into account any practice effects (as determined by practice effects observed in the control group). The group that read the negative false information showed a slight increase in MTS of about 1.5%, but this was not statistically different from the control group.
Moreover, the subjects appeared unaware that the news items were affecting their behavior. Participants’ self-estimations for changes in tapping speed were not found to correlate with their actual changes in tapping speed.
Notably, the average exposure time to the news items was under 5 minutes, suggesting that a very limited degree of exposure to fake news can surreptitiously influence behavior. Bastick says his study most likely underestimates the impact of false information. As the author points out, disinformation on social media is likely to be far more influential, given the potential for repeated exposure and endorsement by friends and other users.
While the behavior changes witnessed were relatively small, Bastick says that the huge audience of online platforms and the potential for viral content can lead to large-scale outcomes. “For example,” Bastick says, “had every voting-eligible citizen been exposed to a disinformation campaign at least as effective as the 5.15% increase that was perceived in the positive group of this experiment (after controlling for practice effects), this would have been sufficient to flip the margin of the popular vote in the last two US presidential elections (2016: 2.09%, 2012: 3.86%).”
Bastick says that his study highlights the need to further investigate how disinformation can impact behavior, find ways to prevent such manipulation, and build stronger democratic processes. “These findings raise deep concerns for the future of society and politics,” Bastick warns. “Disinformation risks skewing individuals’ worldviews and deleteriously informing their behavior. Deliberately produced and targeted disinformation aimed at behavior modification amplifies these risks, by introducing incentives and optimization.”
The study, “Would you notice if fake news changed your behavior? An experiment on the unconscious effects of disinformation”, was authored by Zach Bastick.