Data extraction and descriptive statistics

To evaluate the use of hierarchical modeling, we extracted several metadata for each relevant study, i.e., year of publication, location (country/countries/territories where that study was carried out), taxa (i.e., mammals, amphibians, reptiles, invertebrates, fish, birds), type of study (i.e., single-season occupancy, single-season N-mixture, multiple seasons (dynamic) occupancy, multiple seasons (dynamic) N-mixture, multi-state (single or multiple seasons), multi-scale (single season), multi-species (community) occupancy single-season, multi-species (community) occupancy multiple seasons, co-occupancy), type of data (i.e., camera traps, tracks, acoustic surveys, visual surveys, eDNA, traps, interviews, online databases), study design (i.e., grid, transect, feature-based such as survey at ponds or other discrete patches in the environment, territorial units such as counties or game management units, opportunistic collection), number of sampling units, and duration of study.
We analyzed article metadata using descriptive statistics, including the frequency of articles per metadata category. Chi-square tests were performed to determine if observed frequencies in a category matched the expected frequencies. Furthermore, changes in frequencies across categories (taxa, data, and modeling choice) were visualized multi-dimensionally using alluvial plots. These analyses were performed using the base, dplyr, and ggalluvial R packages \citep{RN765,RN767,RN766,brunson2020ggalluvial}. Graphs, excluding network and alluvial plots, were generated using the ggpubr package \citep{kassambara2020ggpubr}.