It is 20 years since conservation biologists were intro-duced to the merits of meta-analysis by Fernandez-Duque and Valeggia (1994) in this journal. Over 600 conserva-tion meta-analyses have since been published, and the number is increasing steadily. Meta-analysis has a longer history of use in psychology and social science (Glass 1976), and in medical literature its use predates that in conservation biology or ecology by at least a decade (Fig. 1). Meta-analysis is a powerful tool that allows similar in-dividual studies to be combined for analysis of a much larger effective sample size. Conservation researchers work with limited resources (McCarthy et al. 2012) and large-scale well-replicated studies are often not feasible with the available, relatively short-term funding. Due to their large sample sizes, meta-analyses can identify sig-nificant relationships that were undetected in individual studies. For example, due to the absence of a meta-analysis, the efficacy of streptokinase in the treatment of myocardial infarction went undetected for 10 years. The cumulative meta-analysis of 30 years of studies on the topic identified a significant effect of streptokinase after 14 years (Lau et al. 1992). The potential loss of life resulting from this missed relationship highlights the im-portance of meta-analysis in informing medical policy and practice. Although ecological research is inherently more variable, the utility of meta-analysis in identifying signifi-cant relationships across small yet comparable studies is clear. Furthermore, systems may be less readily replicated in conservation than in medicine, but meta-analyses still provide an opportunity to combine studies with some replication to powerfully examine relationships across like studies. A further value of meta-analysis stems from the ability to statistically investigate the influence of heterogene-ity, which allows additional relationships, such as study latitude or local climate, to be assessed. Individual stud-ies vary in methods used and systems investigated, and although such differences may hamper efforts to pool studies into a meaningful analysis, these differences can be investigated as sources of variability in the results and often include factors that were not a focus of the original studies.