Meta-analysis in plant pathology: synthesizing research results

Abstract

Meta-analysis is a set of statistical procedures for synthesizing research results from a number of different studies. An estimate of a statistical effect, such as the difference in disease severity for plants with or without a management treatment, is collected from each study along with a measure of the variance of the estimate of the effect. Combining results from different studies will generally result in increased statistical power so that it is easier to detect small effects. Combining results from different studies may also make it possible to compare the size of the effect as a function of other predictor variables such as geographic region or pathogen species. We present a review of the basic methodology for meta-analysis. We also present an example of meta-analysis of the relationship between disease severity and yield loss for foliar wheat diseases, based on data collected from a decade of fungicide and nematicide test results.