ABSTRACT
For statisticians, the term ‘meta-analysis’ refers solely to the statistical technique of combining effect sizes in order to test a specific research hypothesis. In this sense, a researcher could carry out a number of studies using the same variables and combine the results within a single research paper. This practice is extremely rare, and I have yet to see such a paper published in a psychology journal. By far the most common use of meta-analysis in the psychological literature is to combine data from a large number of published studies by different authors. This is a very different approach from the usual application of statistics in psychology. However the use of published literature for theory-building is very common practice. Indeed, two of the most prestigious journals in psychology (Psychological Bulletin and Psychological Review) are devoted mainly to scholarly reviews of the literature on a specific topic, where the author has written a timely round-up of related studies in that field, or has advanced theory by identifying a neglected link between different research literatures, or has devised a new theory or model based on published work.
