A pretest-posttest design for measuring the effects of educational programs uses comparison groups consisting of pupils like those in the treatment group but not getting that particular treatment. Although the design is geared primarily to evaluation of Title I programs in large cities, it should also apply to other situations. The plan for measuring treatment effects is to (1) identify the major objectives and the most important side effects of the programs, and (2) develop measures of the objectives and side effects. When several programs are being evaluated, careful selection of the comparison group is necessary to avoid the statistical problems of confounding and interaction of treatment effects. Evaluation of treatment effects is the process of judging the value of a treatment. In education, where the consequences of programs have much greater importance than the programs themselves, program evaluation is consideration of the consequences of the program. Therefore, program ratings should be based on how favorable and important the consequences are to the users. (LN)