A fascinating study on social approval, popularity, social media unraveled something that may prove key in getting clicks and attention. Scientists at MIT, NYU, and Hebrew University in Jerusalem discovered that, when a person initially "likes" or "upvotes" a link posted on another person's page, especially if they are a friend, people are more likely to "like" that article as well. However, when a user "dislikes" or "downvotes" an article, there is a likely chance that people will attempt to counteract it in response, especially if the link is their friend's. The study may prove there is some meaning to the term "popular contest," and the notion that hype feeds on itself.
The study, called "Social Influence Bias: A Randomized Experiment," will be published tomorrow in the new issue of Science magazine. In it, scientists conducted what is likely the first controlled experiment set entirely on the Internet, something previously thought impossible. They partnered with an unnamed website (obviously refusing to disclose their identity on the grounds that their site's integrity was compromised through this experiment) in which users submit links. These links were then allowed to be upvoted and downvoted by other users.
In the experiment, scientists left an arbitrary upvote or downvote (with a control group of comments left alone) following a comment, with the number of upvoted comments outnumbering downvoted comments, reflecting the site's more positive tendencies. Then, they monitored site traffic to see what behaviors would result.
Scientists found that, when a comment was upvoted, the first person reading that comment was 32% more likely to upvote as well than with downvoted comments. Furthermore, the overall feedback scores on a comment was 25% more positive than those comments left alone. What was more interesting however, was that downvoted comments eventually had scores that were indistinguishable from comments that were left alone. Given the site's ability to also rate users as part of a social network, the likely situation occurred in that friends were likely to "correct" feedback scores, while rivals stayed away.
In essence, scientists confirm that the notion of a slight, positive nudge can greatly impact the overall impact and influence of a link or comment. Such predicaments led popular news site Reddit to add a new feature that hides the overall feedback score of new links, allowing them to be tried and tested based on personal perception rather than popularity. Of course, there are variables in play, but the science is sound, and the question becomes then: How do we get that ball rolling?