County-level estimates of moral concerns have the potential of being misused for targeting in influence campaigns. To mitigate this risk, we do not share the exact estimates of moral concerns. Instead, we used the `synthpop` R package to create synthetic data which mimic the original estimates of moral concerns and preserve the relationships with the other variables in the dataset—but sever the link between each county and its residents' moral concerns (`d_synth.rds`). All other variables, sufficient to reproduce Models 1 and 2, are provided unaltered.