Research synthesis is sometimes viewed as an end point, a cumulation of scientists' knowledge about a topic. But it's really a sort of way-station in a continuing investigation of a field. It allows science to take stock of what is known in order to identify the most promising directions for future research. This is the generative role for research synthesis. Wood and Eagly (in press) explain how syntheses can be generative in this way, and Wood and Christensen (2003) provide a basic how-to guide for conducting a meta-analytic review.

Wood, W., & Eagly, A. H. (in press). Testing theories and conducting research from meta-analytic syntheses. In H. Cooper, L. Hedges, & J. Valentine (Eds.), Handbook of research synthesis (2nd ed). New York: Russell Sage.

Wood, W., & Christensen, P. N. (2003). Quantitative research synthesis across studies, paradigms, and time. In C. Sansone, C. C. Morf, & A. T. Panter (Eds.), Handbook of methods in social psychology (pp. 335-356). Thousand Oakes, CA: Sage. [request paper]