This paper reviews the question of the cost and benefits of data for the analysis of state policy questions, and concludes that in the abstract the question cannot be answered; rather, it must be considered on a continuing basis whenever data is to be collected in the support of baseline studies or special projects. The paper suggests six criteria which should be considered in determining how an agency will go about acquiring the necessary information to accomplish its mission, and makes the points that addressing these considerations in an open and candid manner and involving the various parties to the data flow should significantly improve the relationship of benefits to costs in data collection. (Author/DAG)