Systematic literature survey
On March 23, 2018 we executed search queries in the Web of Science and google scholar with the search term (meta-analys* OR meta-regression*) AND ecolog* . Google scholar results were compiled with the software Publish or Perish 6 . The 2,626 publications we identified were screened for the following inclusion criteria: i) the research field was ecology (excluding medical, social, financial and ecosystem service studies), ii) the authors conducted an original meta-analysis that was based on summary statistics from previous publications (excluding theoretical, methodological, commentary, raw data analysis and update studies), iii) the study quantified effect sizes as either response ratios, mean differences or correlation coefficients and iv) the authors could or should have applied a weighting scheme to summarize those effect sizes.
The 505 studies that met these criteria were then screened in order to extract: i) the year of their publication, ii) the effect size applied (response ratio, mean difference or correlation coefficient), iii) whether or not the authors encountered primary studies with missing variance or sample size information and iv) how the authors dealt with this missing information. Cases where the authors were vague with stating how they dealt with missing data (e.g. statements such as “we extracted all available data”) were classified as missing data encountered . The Literature search, inclusion criteria, data extracted and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) are reported in the Supplement S2 of the Supporting Information.