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}.