It appears that relatively little thought has been given to the methods for reviewing, synthesizing, and reporting the results of a set of empirical studies on a given substantive topic. The question is raised as to whether the studies that have been reviewed constitute a population or a sample, and if a sample, whether they are representative or biased or random so that inferences can be tested with confidence. The purposes of this study were to: describe the various methods used for specified aspects of reviews and syntheses, determine the frequency of the alternative methods, critically evaluate the strengths and weaknesses of these methods, and suggest ways in which more powerful and valid reviews and syntheses can be done. The primary source of data was a content analysis of a random sample of reviews and of a sample of allegedly exemplary reviews. The author concludes that although such reviews are important to science and social policy-making, many integrative reviews are done less rigorously than is currently possible. This paper and the recent work of Gene Glass suggest several ideas for improving the prevailing methods for reviews. (Author/CTM)