Very preliminary analysis for an effort to see whether newspaper employment/existence/journalism is correlated with arrests for public corruption. We can see that as jobs decline, so do arrests, but there's a bunch more to fix on this

The question becomes: what's the outcome variable of interest - like, do we want to see that newspaper employment is associated with more arrests for public corruption? If there were fewer, newspapers would also be doing their job as implicit watchdogs.

There's an enough problem, too. What constitutes success? Right now, we have BLS data on newspaper employment per county and county per court district public corruption prosecutions. We need to standardize a bit, too.

Would be curious what else might be tested to see if there is an effect on the main relationship here (I think that's called a confounder?) - easiest if available at county level. Obviously vote shifts/partisanship. What else?

Yes, I know we should standardize for population.