A "good but discouraging" book of readings in policy evaluation. Neither of the two key essays, by Rossi and Williams, is optimistic about either the present or short-run future of evaluation. Although both are written in the spirit of "Well, at least we are moving in the right direction," the evidence of the book does not really bear out even the qualified optimism for current direction of movement. The readings as a whole indicate that the gap between the abstract techniques and the actual applications is so wide that it is difficult to see how the techniques can be applied directly to policy. Because there are always technical flaws in any evaluation, it is extremely difficult for evaluation results to have a major effect on policy. Analysis can be an important part of the advocacy process leading to the making of policy, but more important than the style of analysis in government is the existence and style of analysts in government. For analysis to be effective, analysts must be listened to by those making policy. Current evidence is that such an audience does not exist, and that is something that Rossi and Williams and their authors can do little about. 5 pp.