It turns out that social media users may act more herd-like than you may think. The American Association for the Advancement of Science performed a study on the effects of prior social votes on current user behavior. The study shows that preexisting votes have a profound effect on the content’s overall social performance. The test analyzed comments on a website that allowed users to vote and respond to others. In the tests, some of the comments were artificially inflated, while others were left untouched.
If a user read a comment that had a previous positive score, they were 32% more likely to provide their own positive vote. Overall, those updates with an initial positive vote ended up with scores 25% higher than a control group. This research shows that the false positives helped to inflate scores and shape overall opinions of users.
On the overall variance, Sinan K. Aral, a professor of IT and marketing at MIT and a researcher on the project, stated “That is a significant change .. we saw how these very small signals of social influence snowballed into behaviors like herding.”
Additionally the survey showed that the overall relationship between the voter and the commenter played an important role in the outcome. Those who were friendly with commenters would try to boost up any negative scores. The research showed that those who were adversaries would not go out of their way to knock down others with a negative vote and that in the end the initial negative votes didn’t play as large a role as expected as their results were indistinguishable from those in the control group.
The website that was used in the test allowed users to vote on and comment on posted comments. The name of the site was not disclosed.
For more information see the full report or the NY Times.
Comments