Directors of educational projects need to be aware of the consequences their decisions may have for evaluation and appreciate the need for working closely with their evaluators from the earliest planning stage. Attempting to address the needs of project directors and evaluators, this guidebook deals with one central aspect of project evaluation--measuring congnitive achievement gains. Its purpose is to provide the tools needed to conduct technically sound, interpretable evaluation studies. It covers the entire evaluation process from the administrative decisions in selecting an evaluation design to the details of collecting, analyzing, and reporting the data. After the introduction, Chapter 2 describes 12 hazards commonly encountered in evaluations which may invalidate otherwise sound studies. The hazards are discussed, and ways to avoid the hazard are outlined. Chapter 3 presents a procedural guide, in decision-tree form, for selecting a suitable evaluation model given a particular set of constraints. Chapter 4 presents the five evaluation models referred to in Chapter 3. Each model is summarized describing its characteristics, strengths, weaknesses, and considerations relating to its implementation. Chapter 5 deals with the details of data collection and Chapter 6 with summarizing and reporting of impact data. Appendices contain characteristics of some commonly used standardized tests and analysis of covariance worksheets. (RC)