Column_Name,Description
model_comparison_ID,An ID that links each model in model comparison series to each other.
dataset,"The dataset the model was trained on, either Plant_Abundance.csv or Plant_Diversity.csv"
Variable,"The variable tested in the model comparison series. These include landform, environmental variables, functional traits, methodological factors, as well as coevolutionary history and phylogenetic and functional novelty. See Column_Metadata.csv for a description of each of these variables, which occur in the Plant_Abundance.csv and Plant_Diversity.csv datasets."
Analysis,"The analysis series, indicating whether the model comparison is concerned with ‘Environmental Factors’, ‘Functional Traits’, ‘Methodological Factors’, ‘Megafauna Diversity’, ‘Coevolutionary History and Novelty’, and ‘Responses to Native and Introduced Megafauna’"
Sensitivity,"This column specifies whether the model is a sensitivity analysis. Main text results are designated as ‘Primary Analyses’. Sensitivity analyses include, ‘Megafauna Studied in Native and Introduced Range’, ‘Simple Random’, ‘Final in Time Series / Largest Scale Measure Only’,  ‘Unspecified Nativeness Plants Excluded’, ‘Species-level Responses Only’,  and ‘Pure Megafauna Nativeness’"
exclusion,A executable formula to exclude data from datasets for the model comparison group. This can be executed using eval(parse(text = guide$exclusion)) in R.
random,The random effect formula for the model comparison series
formula_base,The main effect formula for the individual ‘base’ model. This is then compared to the null model to understand whether the ‘Variable’ of interest improved model quality.
formula_nativeness,The main effect formula for the individual ‘nativeness’ model. This is then compared to the base model to understand whether megafauna nativeness improved model quality relative to the ‘Variable’ of interest.
formula_null,"The main effect formula for the individual ‘null’ model. This is generally an intercept only model used to test whether the ‘base’ model, and thus ‘Variable’ of interest, improves model quality. Note, that some model comparison IDs lack a null model, as the base model itself is an intercept-only model. In this case (as for model_ID_null and dir_null) this will have an NA value"
model_ID_null,"A unique ID for the null model, corresponding to filename in models.zip"
model_ID_base,"A unique ID for the base model, corresponding to filename in models.zip"
model_ID_nativeness,"A unique ID for the nativeness model, corresponding to filename in models.zip"
dir_null,The directory path for the null model of model comparison series.
dir_base,The directory path for the base model of model comparison series.
dir_nativeness,The directory path for the nativeness model of model comparison series.
base_LRT,The log-likelihood test statistic for the comparison of the base model and null model of the model comparison series.
base_LRT_pval,The log-likelihood p-value statistic for the comparison of the nativeness model and base model of the model comparison series.
nativeness_LRT,The log-likelihood test statistic for the comparison of the nativeness model and base model of the model comparison series.
nativeness_LRT_pval,The log-likelihood test p-value for the comparison of the nativeness model and base model of the model comparison series.