Predictor variables
For each case, data for model moderators, such as the taxonomic group, body size, and habitat type were collected. Taxonomic names were harmonised to those used by the Global Soil Biodiversity Atlas (GSBA; (Table S2). From the GSBA harmonisation, each taxonomic group was assigned to a body-size category (micro-, meso-, macro-fauna) if that had not already been done by the original author. Body-size categories were also based on . Microarthropods were assigned to meso-fauna, macro-arthropods and macro-invertebrates were assigned to macro-fauna. Any cases where authors did not classify the data to a taxonomic or size unit (e.g., described as ‘All soil fauna’), were classed into one category, “All Sizes”. The type of habitat the samples were from was also classified into one of six predefined classifications; agriculture, grassland, woody, cold/dry, wetlands, or artificial.
To help ensure consistency in data extraction, all project members followed a data extraction protocol created at the start of the project. In addition, data extracted from nearly all articles were checked by a second individual.