Meta-analysis is an analysis technique that combines the results of a given body of empirical evidence to determine the size of an effect that exists for a given approach, or the strength of the association between two predictors. It represents a category of research syntheses that bases reviews of the literature on explicit, shared, statistical assumptions (Cooper & Hedges, 1998).
 Extensive references for each kernel described may be found in Embry and Biglan, 2008.
 There are many ways to organize this process. We employ logic models because they provide a familiar and intuitive way to organize these tasks. Logic models are typically developed by program staff, often in conjunction with researchers, board members, and program designers, to depict in detail the inputs, activities, outputs, and short and longer-term outcomes that a program is expected to produce. For example, a program might train program staff (input) to work with parents (activity) to reduce harsh and neglectful parenting (output) in order to reduce early sexual behavior (short-term outcome). However, this framework can similarly be used to develop, assess, revise, and test the effectiveness of evidence-informed programs.