Comment by kimixa
6 hours ago
Statistically you'd expect some "random" samplings of the population to show significant deviations from the mean just by chance.
Roll enough different sets of dice and you'd expect some to end up being all sixes - that doesn't mean that set is rigged. Yeah, they're the ones you'd do further tests on, but it's not evidence in itself.
If the amount of individuals in your sample is large enough, it shouldn't be the case.
Relevant XKCD: 882
But that's why you do multiple testing correction