A topic of increasing concern to scholars and policy makers is the degree of bias in estimates of the socioeconomic returns (occupational status and earnings) to schooling. This paper, based on a re-analysis of published data from seven sources and on new data, attempts to quantify the specification bias due to the omission of socioeconomic origins and mental ability from status and earnings functions. Conclusions are based on three sets of statistics: (1) the amount of explained variance in occupational status and earnings uniquely attributable to schooling: (2) the proportional reduction in socioeconomic returns to schooling controlling SES origins and/or mental ability; and (3) the reduced form and structural estimates for the two data sets available with indices of parental wealth. The results suggest that estimates of the occupational status returns to schooling are not seriously inflated, while those of earnings are overestimated by around 40 percent. Thus, the practice of calculating economic returns without adequate controls for ability and social origins is seriously questioned. This paper also suggests that an over concern with rates of returns to schooling--to the exclusion of other ascribed and achieved effects--obfuscates understanding of the socioeconomic achievement process. (Author/JM)