The use of statistical tools in field testing of putative effects of genetically modified plants on nontarget organisms.

摘要: To fulfill existing guidelines, applicants that aim to place their genetically modified (GM) insect-resistant crop plants on the market are required provide data from field experiments address potential impacts of GM nontarget organisms (NTO's). Such may be based varied experimental designs. The recent EFSA guidance document for environmental risk assessment (2010) does not clear and structured suggestions statistics trials effects NTO's. This review examines practices in plant testing such as way randomization, replication, pseudoreplication. Emphasis is placed importance design features used which NTO's assessed. statistical power positive negative aspects various models discussed. Equivalence difference compared, checking distribution stressed decide selection proper model. While continuous (e.g., pH temperature) classical approaches – example, analysis variance (ANOVA) appropriate, discontinuous (counts) only generalized linear (GLM) shown efficient. There no golden rule test most appropriate any situation. In particular, block designs covariates play a role GLMs should used. Generic advice offered will help both setting up interpretation obtained this testing. combination decision trees checklist trials, provided, analyses assess whether were correctly applied. We offer generic assessors