This is basically an A/B test, admittedly on a moderate sample size but the point Julia is making is that the amount of abuse differed depending on the gender o

This is basically an A/B test, admittedly on a moderate sample size but the point Julia is making is that the amount of abuse differed depending on the gender of the name and photo. All other variables, such as "annoyingness factor of having a pop-up chat box", were the same in all cases. Even if the abuse were partly caused by the chat box popping up annoyingly, one could still conclude that a chat box popping up with a woman's name and face annoys people more than with a man's or a generic "team" one.