Previously, we examined marked drops in voter registration totals as a result of voter roll cleanup in Kansas counties.
Further analysis of this data reveals that voter roll “churn” — the overall change due to a combination of cleanup (subtraction) vs. registration (addition) — is not evenly geographically distributed.
See the chart on the second tab of this spreadsheet. This chart shows the average population size vs % average change (adjusted by number of days since previous file upload).
It is important to consider the difference in magnitude of scales in each county when considering the size of the reported change. For example, a 1% change in a populous county like Johnson County accounts for more churn simply due to the raw number of voters that would affect than if this same 1% churn happened in a less populous county. For that reason, it’s important to establish a baseline.
This data looks at the changes in voter registration totals between reporting dates for our entire dataset. (Looking across 40 voter files, roughly 70 million rows of data.)
Column H shows the change in registered voters per day between two reports submitted by a given county. Paired with Column J, this metric highlights time periods during which counties enrolled or unenrolled significant numbers of registrants.
Column I shows the change in registered voters as a percent, which adjusts the reports based on each county’s population. This is another helpful metric for identifying time periods during which significant numbers of voters were added or dropped from the rolls.
You can see that average daily churn for all counties in Kansas is .01291%. The lowest churn is found in Ellsworth at .00496%. The highest is in Labette at almost .037%– 7.5X higher than Ellsworth and 2.9X higher than average.
These differences in churn rates might be explained by looking at the general stability or transient nature of the voting population in a particular county; low rates of churn might indicate low rates of movement among registered voters. There could also be additional explanations, such as county election officials engaging in efforts to clean up their lists to better reflect those residing in their county. High levels of churn caused by additional people joining the voter rolls might measure successful voter registration efforts by organizations within communities. Or, there could be other explanations.