Current models of scientific inquiry assume scientists to all share the same evaluative standards. However, science is often characterised by multiple ones, that is by evaluative diversity. We investigate how scientific success is affected by evaluative diversity through computer-based simulations. Our results show that communities with diverse standards profit immensely from scientists sharing all the approaches they explored, regardless of whether they considered them valuable. Moreover, we find that even a moderate degree of evaluative diversity can, under certain conditions, lead scientists to reach more satisfying results than those they would reach in homogeneous communities.