The three RMC models endorsed by the U.S. Office of Education for the evaluation of Elementary and Secondary Education Act Title I programs are based on narrowly conceived approaches to evaluation--the measurement of cognitive achievement gains. Each model requires the comparison of observed student performance with an estimate of what level of performance would have been achieved without the Title I project. The models differ in the way the non-participation estimate is made. Under Model A the estimate is based on test norms, assuming equal mean pretest and mean posttest percentile ranks. Model B requires the use of an experimental design including a control group. Model C is described as a regression-discontinuity analysis. Only Model A appears to be feasible for most Title I programs. However, data are presented to challenge the plausibility of the assumption that non-participation would result in the percentile rank of the posttest mean being equal to the percentile rank of the pretest mean; and additional data are presented to challenge the assumption that the use of an equating test might ameliorate this problem. It is proposed that educational evaluation is too immature for fixation on a small set of models. Flexibility is needed. (Author/CTM)