A two-level data set can be structured in either long format (LF) or wide format (WF), and both have corresponding SEM approaches for estimating multilevel models. Intuitively, one might expect these approaches to perform similarly. However, the two data formats yield data matrices with different numbers of columns and rows, and their "cols :…

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)