Evaluation is an act of persuasion directed to a specific audience concerning the solution of a problem. The process of evaluation is prescribed by the nature of knowledge--which is generally complex, always uncertain (in varying degrees), and not always propositional--and by the nature of logic, which is always selective. In the process of persuasion one must ascertain who the audience is and find a basis of agreement on premises, both of facts and values, and on presumptions. Two criteria for evaluation are: the most efficient way to a given end, or the most effective use of available resources. Quantitative evaluation methods involve three stages: (1) substantive definition of the problem and its translation into a formal, mathematical model; (2) compilation of information in terms of the formal model and its formal, logical analysis; and (3) translation of the formal conclusions back into substantive terms. Both formulation and interpretation require good intuitive judgment. The evaluator and the audience must employ their reasoning in a dialogue, and both must assume responsibility, since evaluation is never completely convincing nor entirely arbitrary. The logical arguments used in two works are discussed. The works--Gene V. Glass' review of Michael Scriven's instructional cassette lecture on "Evaluation Skills;" and Scriven's reply--are appended. (Author/CTM)