Election Incident Reporting III: The importance of disaggregation

Just one more set of observations from the excellent Election Incident Reporting System, before I head off to election night parties.

Last time, I blogged about how it would he helpful to adjust these figures by population size or number of jurisdiction; essentially, make them into “per capita” (or per voter or per jurisdiction) measures.

It’s also critical that any user of the system click through to the county level. For instance, let’s take the three “red” states on this map: California, Pennsylvania, and Ohio.

California currently sits at > 250 incidents. But more than half the incidents are from one county: Los Angeles. Otherwise, California looks to be conducting a relatively unprobematic election, at least according to this website.

Pennsylvania similarly reports two counties with a high level of incidents: Philadelphia and Allegheny (not surprisingly, the two most populous counties in the state).

Ohio, in contrast, also falls into the > 250 incidents category, but four separate counties (Franklin, Cuyahoga, Hamilton, and Montgomery) show greater than 70 incidents.