Areas with high in-person voting rates in Wisconsin’s April 7 election saw increases in positive COVID-19 test results in the weeks that followed, according to a working paper published this week by economists from UW-Oshkosh and Ball State University.
Municipalities across Wisconsin were forced to consolidate polling places for a variety of reasons — including a shortage of poll workers and the need to move crowds away from locations with vulnerable people such as senior centers.
Consolidation of polling places led to more people at the polls that remained open, causing more infections, the paper states. Alternatively, counties with higher rates of absentee voting participation saw lower rates of COVID-19 detection in the weeks after the election.
“The decision to consolidate polling locations poses a unique problem for these clerks: closing locations can create some insulation to the relatively vulnerable, but it also increases the likelihood of infection at the remaining locations due to the increase in voter density,” the paper states. “Collectively, these figures show increases in positive test results are much higher in locations with higher in-person voting.”
More than six weeks after the election, the Wisconsin Department of Health Services (DHS) has linked more than 50 confirmed cases of COVID-19 to in-person voting through testing and contact tracing. The paper states that these direct efforts show a link between voting in person and contracting the virus, but are not nearly enough to show the full effect.
“Our results confirm the Wisconsin Department of Health Services findings on the link between the spread of COVID-19 and voting using testing and tracing methods,” the paper states. “However, the tracing investigation undertaken was not comprehensive, and our results indicate a much larger potential relationship. Specifically, results show that counties which had more in-person voting per voting location (all else equal) had a higher rate of positive COVID-19 tests than counties with relatively fewer in-person voters.”
Because it is a working paper, the findings have not yet been peer-reviewed — a process through which academic studies are subjected to analysis and scrutiny. Chad Cotti, one of the paper’s authors, said in a series of tweets that the team released the results so the evidence could be weighed as the state and country head toward important elections in the fall.
For non-academics, this is a working papers series, designed to solicit feedback on preliminary work & improve the final analysis. It is not complete. That said, we decided to discuss it publicly, although we would greatly prefer to wait until it had gone through peer-review.1/2 https://t.co/PZOcyczzIJ
— Chad Cotti (@Chad_Cotti) May 19, 2020
“In this case, however, we decided to share the information publicly because the length of time to complete the peer-review process is typically years and will likely exceed the epidemic,” Cotti wrote. “So, as we will have more elections this year, on balance, we wanted to share the early findings. Any single piece of evidence should be weighed against other pieces of evidence when determining a final conclusion. These early findings provide one additional piece of evidence, which is consistent with the WDHS findings that in-person voting contributed to spread of COVID19.”
Jeffrey Harris, a physician and economist at MIT, previously released a working paper on COVID-19 mitigation efforts that included an analysis of several Wisconsin counties. In an email, he said the paper’s use of a positivity rate could be misleading.
“The authors appear to rely instead on the positivity rate — that is, the number of newly diagnosed infections per test performed,” Harris said. “In my working paper, I expressed concern that the positivity rate can give misleading results.”
The authors do give a reason for using the positivity rate, saying that using only confirmed cases can also be misleading because it relies on who is being tested for the virus — which isn’t randomized.
Harris also pushed back on the paper’s overall findings, saying there wasn’t evidence of the election causing an increase in infections.
“The authors appear to rely on test results three weeks after the April 7 primaries,” Harris said. “In my working paper, I detected no increase in COVID19 incidence in Milwaukee and Waukesha counties during weeks 0, 1 and 2 after the primaries, and an increase only after week 3. I wrote, ‘This is certainly not strong evidence in favor of a large effect of the primary elections on the incidence of new cases. It does not, however, rule out a smaller effect detectable only through detailed case finding.’”
The paper was released as debates over voting-by-mail have become increasingly polarized. A number of states have mailed absentee ballot applications to all registered voters.
President Donald Trump mischaracterized the effort in a (since-deleted) tweet, saying Michigan was illegally sending actual ballots to voters and threatening to pull federal funding. Trump and other Republican politicians have often suggested, without evidence, that mail-in voting increases voter fraud.
On Wednesday, the Wisconsin Elections Commission discussed a staff proposal to mail ballot applications to all of Wisconsin’s registered voters — but ultimately did not vote on the proposal.
But Republicans in the state Legislature have already balked at that idea.
“It does not make sense to centralize absentee voting in Madison when that funding could be better utilized by the clerks themselves,” Senate Majority Leader Scott Fitzgerald said in a statement. “If anything, WEC should consider using federal dollars to boost municipal clerks’ budgets so they can hire extra staff and expand hours for early voting. I also do not support a mass mailing to a voter list that we know has not been cleaned up, in defiance of state law.” (Fitzgerald was referring to an ongoing legal battle over whether the Elections Commission should be forced to purge more than 200,000 voters who may have changed address from the voting rolls.)
With Wisconsin set to hold a primary election in August and the country preparing for a presidential election in November, the paper’s authors believe voting by mail is the best way to keep voters safe during a pandemic.
“Given these results, it may be prudent, to the extent possible, that policy makers and election clerks take steps to either expand the number of polling locations or encourage absentee voting for future elections held during the COVID-19 pandemic,” the paper states.