The objective of this state-of-the-art review was to quantify the replicability of pesticide studies using micro/ mesocosms. Low interpretability of micro/mesocosm studies, and inconclusive and highly variable data, resulted in a discontinuation of the use of these studies for the registration of pesticides. Coefficients of variation, CV%, were calculated on the basis of data tables as a measure of statistical 'effectiveness' taken from the literature. The average CV in the investigated studies was 45%; larger out-door mesocosms averaged 51%, and smaller indoor micro/mesocosms averaged 32%. CVs on variables involving animals were higher than CVs on plant end-points, which in turn were higher than abiotic variables for all experiments. However, to enhance the interpretability and implementation of micro/mesocosm studies for pesticide registration, a number of context-dependent steps could be incorporated; 1) determine the appropriate experimental design and number of replicates by using power analysis, 2) Utilise advanced statistical analysis, such as probabilistic effect distribution and principal response curves, 4) report, preferably in quantitative terms using power analysis, the risk of Type II error. The author's primary conclusion is that the level of CVs is context dependent and, therefore, it is not possible to suggest a generally acceptable level of CVs for all experiments. This has been suggested both directly and indirectly in the literature. Moreover, the number of insignificant (p > 0.05) results is high, 88% of all test biotic variables had no statistical significance. The average number of replicates were 3-4, which theoretically should yield significant effects at least at the highest test-concentration, then resulting in 75-66% insignificant results.