Ordinal data can be rank ordered but not assumed to have equal distances between categories. Using support by judges for civil rights measures and bussing as the primary example, this paper indicates how such data can best be analyzed.