An approach for clearer observation of differences when evaluating educational programs is presented. The standardized tests utilized in evaluating programs are designed to measure topics commonly taught and to maximize individual differences. This masks between-program differences and the unique aspects of different programs. Analysis of item difficulties for an achievement test relative to the program mean difficulty assists in identifying program strengths and weaknesses. The correlations of programs (as variables) using item difficulties as observations indicate the degree of communality between programs. These techniques should assist in evaluation of innovative educational programs. (Author)