SIOUX FALLS, S.D. (KELO) — The COVID-19 case count may not be completely accurate but the study is “directionally correct,” a South Dakota State University math professor said of the recent IZA Institute of Labor Economics’ paper that links 266,796 COVID-19 cases to the Sturgis Motorcycle Rally.
Kurt Cogswell teaches in SDSU’s data science department. The IZA paper is “directionally correct in terms of no spread to some spread,” Cogswell said. As to 266,796 cases that number seems high, Cogswell said.
“I don’t know that I can buy in to their total number,” Cogswell said.
The IZA paper said it is a discussion paper and is provisional. Updates to such papers can be made at a later time.
Although the determined number of Sturgis-related cases may be high, Cogswell pointed out that the IZA researchers were transparent and clear in the assumptions they used.
“They were very clear about their data. They were very clear about their assumptions,” Cogswell said.
He’d venture that if any scientist or economist asked the researcher to discuss a certain point, the IZA researchers would be willing to discuss it.
Cogswell said the paper used the best reliable data available to make its projections.
As a data scientist, he describes high quality data as a specific number of guests at a wedding in Maine who got COVID-19 and spread it to a specific number of people, Cogswell said.
The only way to get that type of data from Sturgis would have been to test people at the rally and then test them again when they returned to their home county, Cogswell said.
Gov. Kristi Noem has called the IZA paper “fiction” conducted by academics and in other media reports, said it was math done on the back of napkin. Dr. Joshua Clayton, the epidemiologist for the South Dakota Department of Health, said on Tuesday that the paper had not been reviewed by peers and the information did not correlate with data within the state. DOH secretary Kim Malsam-Rysdon said using cell phone data to try and project COVID-19 cases is not an accurate correlation.
The most recent criticism is from the city of Sturgis, which issued a news release late Wednesday calling the IZA paper “blatantly faulty.” The city news release from city manager Daniel Ainslie said health departments across the U.S. have been working to identify COVID-19 cases related to the Sturgis Rally and so far, have identified fewer than 300 cases.
But is the paper fiction and faulty?
The IZA research was completed by economists and part of the IZA Institute of Labor Economics’ discussion series. According to the IZA website, the organization “disseminates high-quality research by IZA network members to the scientific community and the interested public worldwide before they are later published in academic journals.”
Cogswell said it’s a classic economic analysis. Researchers work with available data and then make assumptions, Cogswell said.
Some of the assumptions may be large, he said. One example is looking at the COVID-19 case increases in a county from which a rally attendee comes from and comparing that increase to a county that didn’t have a rally attendee, he said.
Assumptions don’t make the paper fiction or irresponsible, Cogswell said.
“I’ve done back of the napkin math and this is better than that,” Cogswell said.
It’s meant to contribute to public discussion, Cogswell said of the paper.
Clayton said the paper had not been peer reviewed.
IZA’s approach with discussion papers makes it different than traditional scientific papers published in journals with peer review.
And this is included on the paper itself “IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character.”
Malsam-Rysdon discounted the researches use of cell phone data to try and project COVID-19 cases.
Using cell phone or mobile devices in COVID-19 research is common.
The South Dakota DOH uses a Care19 Diary App, which is a tool for mobile devices used for contact tracing.
The Centers for Disease Control uses cell phone data and has encouraged the use of that information during the pandemic.
The IZA researchers used smartphone data from SafeGraph Inc. to track the mobility of non-resident Sturgis attendees during the 10 days of the Sturgis Rally.
“SafeGraph provides census-block-group-level data from 45 million anonymized cell phones that
allow us to measure the residence of individuals and the jurisdictions to which they travel,” according to the IZA paper.
Kyle Cronin, a professor at Dakota State University, said date provided through mobile phones can provide many details about the user.
“The data can identify the location and how long the user is that location, Cronin said.
The data can even be used to determine where bar patrons may have been before entering another bar, Cronin said.
Cronin said companies such as SafeGraph generally provide some information to researchers. But there is often a fee for a lot of details, he said.
An editorial published online in ScienceAdvances called “Mobile phone data for informing public health actions across the COVID-19 pandemic life cycle” said there needed to be more consistent and widespread use of mobile phone data in fighting the pandemic.
The June 5 editorial also said that “findings must be shared quickly–there will be time for peer-reviewed publications later.” That is in contrast to some other reports about the danger of releasing non-peer reviewed information.