This report identifies and synthesizes what is known about specific program features for children and youth aged 6-17 that might be manipulated or regulated-components such as staff wages, group size, activities, and theoretical approach that make up the "how" of program implementation. The authors have culled information from the limited extant collection of "gold standard" research--experimental studies that systematically manipulated program components and recorded their apparent effects on participant, provider, and program outcomes; this is presented in Chapter II. They have also synthesized findings from non-experimental sources, such as quasi-experimental studies, multivariate analyses, meta-analyses, and literature reviews (Chapter III). Finally, they have drawn on the wisdom of practitioners in the field who daily experience the benefits and detriments of various program components; this perspective is shared in Chapter IV. Their goal is to summarize and, where possible, synthesize information from these three types of information. Throughout the document and in the selected findings, they identify specific program features as "positive", "neutral", or "negative". These assessments reflect whether the component appears to benefit or detract from positive program, participant, or provider outcomes. In general, an assessment of neutral indicates that a particular component does not appear to "matter" relative to program outcomes. Specific assessment definitions differ slightly according to the type of information from which they are drawn; they encourage the reader to review definitions in the methodology sections of Chapters II, III, and IV. Altogether, they reviewed hundred of studies, meta-analyses, literature reviews, and provider writings, of which roughly 130 were selected for inclusion in this report. These sources address nearly 50 different program features. The authors present a small, and particularly rich, subset of findings from this collection. These fifteen components, discussed in more detail in their concluding chapter (V) and in Tables D1 and D2, comprise those findings informed by at least two, and sometimes all three, sources of information. Appended are: (1) Source Information from Chapter II (Experimental Research); (2) Source Information from Chapter III (Non-Experimental Research); (3) Features of Developmental Settings; (4) Presentation of Findings; and (5) Valuable Resources for Program Design and Implementation. (Contains 25 footnotes.) [This paper was originally prepared for the John S. and James L. Knight Foundation (August 1, 2003) and was updated, revised, and expanded for The Atlantic Philanthropies (October, 2006).]