The evidence "group A was hired more than than group B, and the effect was statistically significant" doesn't imply we should find a significant effect when comparing any other set of groups A, B, ~A, C, C^B, etc.

It's generally hard to prove things with statistics without a randomized experiment.