This paper presents a new technique in which covariates are controlled and evaluation data still are presented graphically in a manner that makes sense to policy makers. An illustration is provided using data drawn from an evaluation by the author of a major National Science Foundation (NSF) university funding program. (Author)