This population-based case-control study examined the risk of multiple myeloma in relation to socioeconomic status. Subjects included 689 cases with newly diagnosed multiple myeloma during 1977-1981 from four U.S. populations and 1,680 controls selected from residents of these same populations. We collected lifetime occupational histories and coded them according to the 1970 Duncan Socioeconomic Index and Nam-Powers Socioeconomic Status scores. We classified scores for the occupations held the longest, highest ever held, and held most recently into quartiles based on the distribution among controls. After adjusting for age group, race, and study site, risk of multiple myeloma was inversely associated with socioeconomic status scores in both men and women. Risk among persons in the lowest quartile of scores was 63% higher (95% confidence interval 21%-119%) than that among those in the highest quartile when the highest Nam-Powers score was used. Similar trends were evident for all three methods of classifying occupational history and for both Duncan and Nam-Power scores. These results changed little after removing from analyses occupations previously associated with increased risk. The occupation-based scores were stronger predictors of risk than years of education. As a proxy measure of occupational, environmental, or life-style factors, socioeconomic status may be a clue to etiologic factors for multiple myeloma.