<b>Data and code for</b>: Seagrass fisheries expose the poverty dimensions of marine conservation

This repository contains the R code and associated input datasets used to reproduce the analyses and figures for the manuscript:<b> </b><b>“Seagrass fisheries expose the poverty dimensions of marine conservation”</b><br>The analysis investigates relationships between household poverty, fishing livelihoods and seagrass dependence across 1,705 households from 156 coastal communities in five Indo-Pacific countries: Cambodia, Indonesia, the Philippines, Sri Lanka and Tanzania.The R script reproduces:Descriptive statistics on household composition, fishing households, poverty status and seagrass use.Comparisons of income between fishing and non-fishing households.Mixed-effects models testing differences in household income between fishing and non-fishing households.Multi-model averaged mixed-effects models examining predictors of household dollars per day among fishing households.A complementary binary mixed-effects model examining whether complete seagrass dependence is concentrated among lower-income households.Country fixed-effect robustness checks for the main household income model and seagrass-dependence model.Model-averaged coefficient tables and robustness summary tables.Figures used in the manuscript, including the Indo-Pacific seagrass fishing map, household income predictor plots and predicted probability of complete seagrass dependence.The household income variable is expressed as International dollars per day, adjusted for inflation, purchasing power parity and household economies of scale using a square-root equivalence scale. Complete seagrass dependence is defined as households using seagrass meadows as their only habitat for fishing or gleaning.<b>Files</b><b>R analysis script</b>: Contains all code required to clean the household data, run statistical models, conduct robustness checks, generate summary outputs and produce manuscript figures.<b>household_data_final_edit.csv</b>: Household-level survey dataset used for the socioeconomic and fishing livelihood analyses. Anonymised.<b>global_seagrass_fishing.csv</b>: Dataset of seagrass-associated fishing observations used to map the wider distribution of seagrass fisheries.<b>villages_fishing.csv</b>: Locations of surveyed fishing communities and seagrass-fishing observations used to produce the Indo-Pacific map.<b>Reproducibility notes</b>The analysis was conducted in R using packages including <code>tidyverse</code>, <code>lme4</code>, <code>lmerTest</code>, <code>MuMIn</code>, <code>broom.mixed</code>, <code>car</code>, <code>sjPlot</code>, <code>patchwork</code>, <code>rnaturalearth</code>, <code>viridis</code> and <code>ggthemr</code>.The script assumes that all input files are stored in the same working directory. Users should update the <code>setwd()</code> path at the start of the script to match their local directory before running the analysis.The main income model uses multi-model averaging of mixed-effects models, with country, region and area included as nested random effects. Robustness checks are included where country is treated as a fixed effect and either forced into all candidate models or included as a candidate fixed effect. These checks were added to assess whether focal household-level associations were sensitive to country-effect specification.