Study finds online commentards easily duped, manipulated
Just garden-variety chumps, really
Internet forums often use reader moderation to determine which comments are the best, but new research suggests that tallying up and down votes for online comments is a poor measure of those comments' actual quality.
Oh, you may think you know who's brilliant and who's a troll in our forums, dear Reg reader – but according to a paper published in the journal Science on Friday, the so-called wisdom of crowds can often be misleading.
"When you rate things online, you are often exposed to others' ratings (either aggregated or listed individually)," Sean Taylor, one of the paper's authors, wrote in a blog post describing the research. "It turns out that this does impact rating decisions and creates path dependence in ratings."
Specifically, forum comments that receive positive votes are disproportionately more likely to be up-voted again, while comments that receive negative votes usually have those votes negated by positive ones shortly thereafter.
In other words, when people see that a comment has been up-voted, they tend to go along with the moderation in a "herd-like" fashion. When a comment has been down-voted, on the other hand, they tend to want to "correct" the moderation, producing an asymmetrically skewed snapshot of opinion.
Taylor and coauthors Lev Muchnik and Sinan Aral conducted their research by working with an unnamed "popular website that aggregates news stories" – not Reddit, they say, but something like it.
For a period of five months, the researchers randomly moderated every comment posted to the site – more than 100,000 in all – giving each an up vote, a down vote, or no vote at all (the latter comments being the control group).
They found that comments that received a random up vote were 32 per cent more likely to draw additional up votes than were comments in the control group, and those comments' final moderation scores tended to be 25 per cent higher than those of other comments.
Comments that received a random down vote fared differently. Other users didn't seem to jump on these negative ratings and drive them down further, as they did with well-rated comments, but often tended to up-vote the comments to cancel out the negative rating.
"Our experiment does not reveal the psychology behind people's decisions, but an intuitive explanation is that people are more skeptical of negative social influence," Aral told Science. "They're more willing to go along with positive opinions from other people."
The implication is that such positive voting can potentially be used to manipulate public opinion through chat boards, polls, and other online forums. But on the bright side, if there is an easy way to control just which way online herd behavior runs, this research wasn't able to find it.
"We conclude that while our manipulations do draw attention to comments and inspire more voting, they don't do it any systematic way that we can identify," Taylor said.
What do you think, dear reader? Is moderation of online forum comments to be believed? Or can comment moderation be used to sway public opinion like a virtual Pied Piper, herding online readers like a pack of gullible lemmings to the proverbial cliff? Naturally, here at El Reg we encourage you to make up your own mind. ®
Sponsored: Benefits from the lessons learned in HPC