Typically running an experiment with 216 variables is ridiculous, but not when you’re a scientist with 67 million subjects and 27 million more every day. League of Legends, an online video game super giant, is notorious for its toxic environment. The game’s company, Riot has been very transparent with their research and even assist other researchers by offering their large fan base as subjects for other tests such as teamwork. A sample size of the million’s magnitude would normally be impossible for a scientist to obtain, but with Riot’s assistance a new avenue for widespread science to be performed as opened. Neuroscience PhD graduate and gamer Jeffrey Lin was hired by Riot, to find how to curb the foul mouthing and anger.

A large scale scan of chat logs from games found that 1% of players were consistently toxic, but the other 99% were simply frustrated people who were lashing out.
So the team tried to prevent the toxicity before it happened through priming, presenting images or texts to influence behavior toward a desired outcome. This is where the 216 variables come in: 24 in-game tips presented in three colors at different points in the game.

From Nature “Can a video game company tame toxic behaviour?”

They found negative messages warning “Teammates perform worse if you harass them after a mistake” when posted before a match in red reduced offensive language by 11%, but the same message in white text was only made a 1.3% decrease. Positive reinforcement shown before a match such as “Players who cooperate with their teammates win 31% more games” showed a less significant change with blue text decreasing offensive language by 6.2%.

After sifting through the 216 different combinations, Lin found that a cautionary tip combined with the color red that people often associate with warnings and a positive tip colored blue, a whimsical and creative color, shown before a match effectively deter negative behavior.

Despite the priming, a toxic environment persists in League of Legends. The problem may stem from the internet’s anonymous, non-consequential culture, or even the game’s inherent sexist character portrayal.

Nevertheless, Lin’s efforts are the first steps to mass refining a community and with such a large sample size, he’ll likely find out how faster than any psychologist or scientist could.