The authors actually expand on this reason at length, trying to figure out how they could test not whether crime is a cause of the lower social interaction, but presuming that crime is a necessary consequence of density and testing by what mechanism it makes people interact less.