Appendix A. A figure showing meta-analysis results calculated using the log response ratio effect size metric and a single randomly selected interaction per predatorprey species pair per published source.
FIG. A1. Results of a meta-analysis using log response ratio effect sizes to examine the effect of predator hunting mode and predator habitat domain on the strength of non-consumptive effects for a variety of prey response metrices. For papers in which the same predatorprey species pair was measured under different conditions (i.e., high vs. low predator density), a single interaction was randomly chosen for analysis. Mean effect sizes were generated using a random-effects model, and a bootstrapping routine was used to generate the accompanying 95% confidence intervals. Numbers in parentheses at the base of each bar represent the number of published papers (regular text) and the number of independent experiments (bold text) contributing to a given mean. ND = insufficient data (fewer than five independent experiments). For initial values, *P < 0.05. For adjusted values, § = significant at α = 0.05 after step-up FDR Bonferroni-type correction (Benjamini and Hochberg 1995).
Benjamini, Y., and Y. Hochberg. 1995. Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society, Series B 57:289300.