At a time when only 4% of the population had graduate degrees, if 25% of married grad-degree-holders were married to another grad-degree-holder, that would constitute a large preference in favor of other graduates.