In classification research reports, psychologists should inform policymakers of methodological limitations. The possibility of prediction error must be underscored, and information should be given about the proportion of the variance that models can account for or not explain. Psychologists must also be clear about differentiating the context of classification studies as regards basic research and policy research. Empirical classification models provide analytical frameworks for extending understanding of offender behavior (basic research). Given the just-deserts debate and the serious consequences of prediction errors, the researcher must be cautious in not prematurely extending basic research results to the formation and implementation of policy decisions. This caution is particularly crucial for research settings where work on offender classification is funded by Federal sources that promote policy applications as explicit goals of the grant. Footnotes are provided. (Author abstract modified)