The roar of Harleys announced the determination of 460,000 people to ignore public health warnings and attend the 10-day Sturgis Motorcycle Rally in August, largely without masks or social distancing.
The silent pings of their cellphones enabled a team of economists to trace these reckless souls back to their home counties and statistically link the rally to at least 250,000 COVID-19 infections.
The results are reported in a new study, “The Contagion Externality of a Superspreading Event: The Sturgis Motorcycle Rally and COVID-19” from the Center for Health Economics & Policy Studies at San Diego State University. The introduction quotes what the lead vocalist of Smash Mouth told the barefaced crowd packed before the stage at a concert during the rally:
“Now we’re all here together tonight. And we’re being human once again. F*ck that COVID sh*t.”
As is taught in every introductory statistics class, there is a critical difference between correlation and causation. And the four authors of the study are economists, not epidemiologists.
But someone who is an epidemiologist, the highly regarded disease specialist Dr. Jeffrey Klausner, described the results of the study as “plausible” in an interview with The Daily Beast. Prominent virologist Dr. Angela Rasmussen used the very same word in describing the study to The Beast. Both Klausner and Rasmussen noted the size of the rally and the crowd’s apparent failure to take mitigating precautions.
Klausner did take issue with the study’s estimate that the additional cases generated public health costs of $12.2 billion. He termed the number “absurd,” observing that it was generated by what the study terms that “the statistical cost of COVID-19 case of $46,000.” Klausner figures that a fair number of those who tested positive would have been either asymptomatic or in need of only minimal medical care.
“Generally, in epidemiology a case is somebody who has symptoms,” Klaunsner said.
Rasmussen focused on the human cost. She lamented that America is so lacking in contact tracing that we may never have more than a correlative estimate of how many were ultimately infected as people came to Sturgis from all over the country and then returned to whence they came with whatever COVID-19 they may have contracted. She said that South Korea was able to test 45,000 people immediately after several night-clubbers tested positive in Seoul, thereby identifying 160 cases and preventing a serious outbreak.
In the absence of such contact tracing, we at least have cellphone tracing that the team of economists—Dhaval Dave, Andrew Friedson, Drew McNichols and Joseph Sabia—employed. They used it to produce what may well be the best estimate they will ever get of the Sturgis rally’s impact on public health. And they seemed to have made every possible effort to generate the most accurate number.
As recounted in the study and in a Daily Beast interview with Dave on Tuesday, the team began by analyzing “anonymized cell phone” data from July 6 to Aug. 30 that was publicly available from SafeGraph Inc. They studied movement patterns in Sturgis and surrounding Meade County before, during and after the rally. They also traced the participants back to their home counties.
In counties that had a significant number of Sturgis attendees, the economists studied the number of COVID-19 infections in the days afterwards, from Aug. 28 up to Sept. 2. They compared that to the number for the proceeding days and to infections in countries that were similar by various measures—including mitigation policies—but had no rally-goers.
“We are very careful,” Dave said. “You have a treatment group and compare it to a control group.”
He was deliberately using the terminology of medical researchers.
“Multiple placebos,” he said.
The ultimate effort was to create what is known in the team's field of study as a counterfactual. They constructed what Sabia called “an alternate universe” of counties similar to their counterparts in every way save for one all-important factor.
“What would they have looked like if this rally had never taken place?” Sabvia said. “It’s upon us to estimate, to give our best estimate of what that counterfactual would have looked like.”
As part of their effort to be as detailed and complete as possible, the team determined that the residents of Sturgis and surrounding Meade County spent an average of 10.1 hours a day at home prior to the rally, but only 8.8 hours during it.
The team had done similar calculations with an opposite outcome when studying COVID-19 spread during Black Lives Matter protests and the Trump rally in Tulsa. The average time at home had increased significantly during both and the team theorized that what was in effect sheltering in place in the general locale may have offset any increase in virus infections resulting from the events. The team determined that neither the protests nor the rally appeared to have caused a spike in the number of COVID-19 cases in those general locales. They thereby thwarted expectations on both sides of the great American divide.
“The data was the data,” Sabia siad. “Nothing political about applying economy theory. Just study the data and see where it goes.”
The data regarding Sturgis showed that the number of pings by people from outside the community increased by 92.5 percent compared to a two-week period prior to the rally. The crowd there happened to peak on Aug. 8, the night Smash Mouth played and lead singer Steve Harwell imparted from the stage the same “f*ck COVID” message offered by T-shirts a number of people in the crowd were wearing.
The study’s subsequent data indicates that the rally increased COVID-19 infection in Meade County by 6.3 to 6.9 cases per 1,000 of the population. Infections in South Dakota were calculated to have risen by up to 3.9 cases per 1,000 as a result of the rally.
“This represents an increase of over 35 percent relative to the 9.7 cases per 1,000 population in South Dakota on July 31, 2020,” the study noted.
The study further found that the rally caused COVID-19 infections to rise by 10.7 percent in counties with the highest number of Sturgis attendees. The counties with few attendees were also factored in. The final tally was astonishing, but deemed plausible by the experts: 266,796 coronavirus infections.
“Or 19 percent of 1.4 million new cases,” the study says.
However accurate the number of COVID-19 infections caused by the rally may or may not be, nobody should contest the team’s ultimate goal concerning those of us who still shun masks and ignore social distancing.
“Getting individuals to internalize the ultimate cost they’re imposing on others,” Sabia said.