The article already states that they found a very clear correlation. Besides if someone is turns out to be a complete @$$hole whenever they aren't afraid of repercussions do you really want them working at your company.
This means absolutely nothing if you don't compare against the equivalent rate of "employee toxicity" from "non-toxic players". 25% isn't that high; I would bet that in general 1 out of 4 LoL players have been toxic in chat at one time or another. You can't just state 25% for one side of the equation without also mentioning what results you found from the other side.
I suspect that 25% of LoL players in general have been "toxic". You can't just focus only on the "bad employees" and call that correlation.
Oh well that makes it ok then. If I find a clear correlation between say black skin and employee underperformance, I guess I can just go right ahead and sack individuals based on some statistical association discovered in a population they happen to belong to.
What the fuck is wrong with you people? I mean the entire lesson of not being racist or not being sexist or whatever is that it isn't ok to judge individuals by the population they happen to belong to. I mean guess what, black people are more likely to be criminals. Women are less likely to be coders. White people are less likely to play basketball well. These are statistical facts, though it seems awkward to say them. But that's ok, because the entire point is that you assess each individual you meet in life on their own merits, not some statistical generalization.
And somewhere along they way, you've missed this point. You don't get to assume that an individual is a shitty employee because they are a shitty LoL player just because you have some correlation to go on.