Two randomly chosen white students interact three times more often than do a black student and a white student. However, placing the black and white student in the same freshman dorm increases their frequency of interaction by a factor of three.
Stumbled upon this old paper. Interesting read.
