Citation

Goodman LA. Am. J. Sociol. 1953; 58(5): 503-510.

Copyright

(Copyright © 1953, University of Chicago Press)

DOI

10.1086/221203

PMID

unavailable

Abstract

On the basis of the analysis of experience tables alone, the use of a prediction instrument as a decision-making procedure which replaces the procedure in use is found to be unjustified. The experience tables are determined in part by the actual decision-making procedure in use. Hence any analysis of a prediction instrument by experience tables alone must be made relative to the actual decision-making procedure. This calls for reformulating the use of prediction. The proposed reformulation involves treating prediction as a "red flag" in the final stage of the decision-making process. It also introduces explicitly the determination of social costs in the use of prediction.