This paper argues that research on segregation has had an over-reliance on statistical measures of outcomes, which limits our understanding of segregation processes and impacts.