This article proposes a new method for measuring school desegregation in multiracial districts, and uses the new method to measure the desegregation effects of magnet schools in Los Angeles. Rather than measuring desegregation between only two groups at a time, I compute the index of interracial exposure for Blacks, Whites, Latinos, and Asians. Each of these figures is then combined to form a composite desegregation score. I find that magnet schools in Los Angeles are less segregated than nonmagnet schools, but that Whites and Asians are the most segregated groups and have become more segregated over time. (Contains 5 tables and 5 notes.)