If every group is a bell curve, then the curves have similar variance and similar mean as well, and the variance within each group is larger than the difference between the means, and the differences are not big enough to say in a meaningful way that one group is smarter than another.