According to psychological essentialism, people divide the world into categories that are seen as possessing deep, underlying properties that account for what is common in members of the category. I examine two ways in which this phenomenon has been used either to debunk or to vindicate essentialism about natural kinds. I argue that neither way affects the essentialist thesis, since they depend on other types of evidence that independently reject/support the thesis. I contend that research on psychological essentialism may play a more direct role in a different argument, which addresses the reality of certain natural kinds. To this end, I will revise the issue of the mind-independence of natural kinds through the concept of unification principle. This concept can offer a criterion of natural kindness that allows certain sorts of mental dependence as constitutive of an objective mind-independent kind. I will apply this idea to the case of race, examining some ways in which findings about psychological essentialism could either debunk or vindicate the existence of such a natural kind, and extend it other putative kinds.