Let me describe a situation you may have observed before. Some game reviewer –maybe a professional critic, maybe just another user– talks about some new game that you are interested in. Overall the review is positive. Let’s say the reviewer gives it 8 out of 10. But there’s one paragraph that’s critical of some element of the game, or the reviewer has some bone to pick with some mechanic or narrative element. And so their final word is something like “Super Buff Warrior 3 delivers a ton of really fun action, a rewarding core loop, and a surprisingly deep customization system. But the weapon upgrade system is pointless and honestly the sidekick character is annoying.”

As a result of this one slice of the larger review, some fans of the game convinced that the reviewer hated the game. They angrily post responses and comments about the review, accusing the author of having an axe to grind with the genre or not knowing how to have fun or completely ignoring all the great stuff about the game. Other less vocal readers may not share their outrage, but they may consider the review mixed and the game only okay. This despite the fact that the author overall liked the game and spent the majority of the review explaining why.

Ever seen that? I have.

One of the reasons this happens is that humans are biased to pay more attention to negative information and weigh it disproportionately when form ing opinions or judgments about something.1 (Another big reason is fanboy/girl-ism, but that’s a topic for another time.) In the absence of a good reason not to, we think about negative information more and tend to remember it more easily.2 If we meet a new person and learn one good thing about him and one bad thing of equal severity, we will form a slightly negative overall opinion of him because we weigh the negative quality more. 3

This effect has also been studied in the context of reader comments about news stories in social networking sites like Facebook. In one study4 researchers showed subjects a Facebook post about an article on legalization of marijuana. They also provided the full article the post was reefering –excuse me, referring to. The researchers then systematically varied the quality of comments so that some were positive about the linked article and others were negative. Within those positive and negative responses, they also created some that they called “subjective” and some they called “argumentative.” Subjective arguments simply stated an opinion without providing an argument for it (e.g., “Legalizing drugs is wrong”). Argumentative comments actually tried to make an argument one way or the other (e.g., “Prohibition creates illegal and harmful black markets; legalizing would undermine criminal operations and give buyers safe options”). They then randomized the presence or absence of these kinds of comments in the Facebook post and measured people’s general attitude towards the linked story.

The found that the presence of negative argumentative comments (that is, critical of the article and providing reasons why) led to more negative attitudes towards the article and towards the idea of legalizing pot. The same was not true of negative subjective comment. Nor was the inverse true of any kind of positive comment –nothing made readers in those conditions more favorable to the article than they already were. The negativity bias, the researchers, hypothesize, might be to blame.

It strikes me that browsing through user reviews on Amazon, Steam, or other virtual storefronts is not unlike reading comments on Facebook or other social networking sites. People post thoughts about and reactions to the game in question, and other people read those thoughts and form opinions. While many of the most highly rated reviews are nuanced and contain both positive and negative thoughts, it’s the negative ones –even in the overall positive reviews– that get the most attention from other people and result in the most discussion. Negative information is just so much more attention getting.5

It’s awful, terrible, and horrible, don’t you think?