In this paper the authors illustrate the use of a strict standard for evaluating evidence on programs and strategies designed to improve outcomes for high school students. They explain what they mean by solid evidence, and present examples from multi-site evaluations of three programs. After that they examine some of the evidence on high school size, and explain why clear inferences about cause and effect remain elusive. They also look at examples of studies that use data for the whole student population in large districts, as an approach to reduce possible selection bias. They conclude with a predictable recommendation for more rigorous evaluation, and a programmatic suggestion. (Contains 21 footnotes.)