Meta-analysis and Reporting Guidelines

Reporting Guidelines for landscape/study site parameters

Meta-analyses (see below) and systematic reviews are crucial for research and practice, because they synthesize evidence from various studies on one subject. The big advantage of meta-analyses and systematic reviews is not only their high transparency but particularly the large sample sizes and thus high statistical power.

Yet, poor reporting in primary studies too often hinders the conduction of meta-analyses and systematic reviews. Especially the characteristics of study sites often are poorly reported. A main focus of landscape ecology lies on the influence of the landscape and the spatial arrangement of landscape elements on wildlife species. Precisely this influence cannot be analyzed in a meta-analysis or systematic review, because the data are not available.

I am developing a proposal of reporting guidelines for landscape/study site parameters in order to facilitate future work on meta-analyses and systematic reviews. The “black box” of study site characteristics hopefully will become a bit more transparent.

Meta-analysis – Usage and definition in ecology

Meta-analysis is a statistical method to synthesize research results across various studies. The data of the primary studies are transformed into a common effect size, are weighted and pooled to a new overall effect. In the medical sciences and in psychology meta-analysis is used since decades and the term is well-defined. In ecology the method became popular in the beginning of the 1990s.

The term “meta-analysis” is not well-defined in the ecological sciences and often does not refer to the clear and definite methodology that is established in the medical sciences and in psychology. In our article “Meta-analysis: A need for well-defined usage in ecology and conservation biology” (Vetter et al. 2013) we evaluate 133 papers from the area of BIODIVERSITY CONSERVATION that had “meta-analysis” in their titles. To be rated as a meta-analysis a paper had to fulfill several criteria which we extracted from the standard meta-analysis literature in the medical sciences (Borenstein et al. 2009). One third of the papers were no meta-analysis and even 25% did not fulfill a single criterion for a meta-analysis. Only one of 133 papers fulfilled all criteria, including e.g. the consideration of heterogeneity.