Oct 07

Tracking COVID-19: ‘We are in an unknown world’

In this new world of COVID-19, we’re all in uncharted territory, according to Dr. Martin Andersen, assistant professor of economics at the Bryan School of Business and Economics.

“Our behavior (social distancing and other preventative precautions); our health and the health of those most vulnerable in society, the economy – our incomes and savings accounts; our children’s education, and so much more, are unusually uncertain now as we navigate the pandemic world,” said Andersen.

Andersen is one of the health policy researchers sifting through a myriad of data and working to understand what is happening, why it’s happening, and what we can learn from predictable and unpredictable human behaviors. He’s looking at the consequences of these behaviors, as well as the ever-changing behavior of the spread of the coronavirus to assess, and possibly predict numerous possible outcomes.

“We’re in an unfamiliar world right now, doing our best to use traditional research methods – as well as finding innovative ways of collecting and analyzing data – in order to make sense of what’s going on now, in the recent past, and as we project into the future,” Andersen said.

Three recently posted pre-prints illustrate some of his work:

1. Early Evidence on Social Distancing in Response to COVID-19 in the US investigates the importance of communicating the threat posed by COVID-19 since most changes in social distancing appear to be voluntary, possibly reflecting beliefs about disease risk. (April 5; revised June 25)

2. The Effect of Federal Paid Sick Leave Mandate on Working and Staying at Home: Evidence from Cellular Device Data suggests that the paid sick leave mandate decreased full-time work and increased at-home work. Given that up to 47% of employees are covered by the federal mandate, the effect sizes are significant. (May 2020)

3. Impacts of State Reopening Policy on Human Mobility harnesses cellular device signal data to examine the effects of the timing and pace of reopening plans in different states. It evaluates observations to come to understand the extent to which people are resuming movement and physical proximity as the COVID-19 pandemic continues. (May 2020)

With numerous articles in the hopper, Andersen says this is an incredibly exciting time for his research because he was among the first to study utterly new phenomena early on; there are many researchers in this space now. His work is being cited in papers far more than ever expected because he’s one of the early health economists researching these topics.

“Being an early adopter of studying social distancing comes with a great deal of responsibility. We’ve got to get it right, to lay out the basic facts upon which further research can be built in looking at human behavior, health, and the economy – and how they are intertwined.”


Andersen enjoys working with a huge variety of data vendors to figure out what’s being measured: people staying home, distances traveled, mask use, and more.

“What we are seeing in the data now is the breaking apart of the relationships that existed early in the pandemic,” he said. “There’s a correlation between increased social distancing and the case counts getting weaker. This a good thing. With this model in mind, the measurements I use are good proxies for social distancing. This means people are doing things to reduce the risk of transmission. Or not. Either way, there are consequences.”

Take Texas for example, which opened prematurely with hopes of activating the economy. Not long after opening, the state experienced seven weeks of skyrocketing coronavirus cases.

“Our data shows what we would expect to see when states don’t stay shut down long enough. We can see that it’s dangerous to ignore the precautions,” he said. “Now Texas has reversed itself and is having cities mandate the mask use. This is a big deal.”

Using the data to look ahead, Andersen sees more coronavirus cases a few weeks out where people aren’t staying home and fewer cases in places where people are responding cautiously.

His research explores another dimension of social distancing: the correlation between how well-connected different (geographic) areas are to one another and the spread of the coronavirus. The degree to which counties are connected (through human mobility) corresponds with coronavirus case growth. This is a new way of looking at distancing data, and part of a broader Indiana University study of the network of connections affecting disease spread.

“What makes the coronavirus so tricky is that many infected people don’t feel bad,” he said. “They have no idea that they are sick, and so they interact with people in the population, many of whom may be more susceptible. Then, it can take five days before an infected person shows symptoms. If infected people felt sick right away, we could eliminate this five-day window. We are trying to find a way to catch the virus before it is transmitted by mitigating or containing it once we know where it is.”

New research methods include the use of cell phone data, credit card transactions, payroll records, and Google search trends.

“We know there are lots of ways to look at things, and we are learning every day,” said Andersen. “With the data that’s accessible, we can track trends in mobility and see spikes in Google searches for things like ‘Coronavirus Symptoms.’ Looking at credit card transactions reveals that in places pushing to open the economy quickly, such as Texas, California, and Florida, there is an increase in mobility, which correlates to an increase in disease transmission.”

Andersen’s challenge is to find the best and most useful applications for so much data.


“We can use Sweden as a cautionary tale,” Andersen said. “Sweden and Denmark share borders but have vastly different policies resulting in vastly divergent health and economic outcomes; Sweden didn’t impose lockdowns at all and is now paying a heavy price as thousands more people have died from the coronavirus than in its neighboring countries. The Swedish economy has not experienced any economic gains from staying open. Denmark did impose a lockdown, has curbed coronavirus deaths, and now is slowly opening schools and businesses.”

Andersen reflects: “We were lucky in 2003 with the first round of SARS. We were lucky with MERs. But now our luck has run out and we are scrambling to adapt. Even though things haven’t been as good as we hoped with the first round of re-openings, we can study the data to look at what works and what doesn’t.”

“We may all need to adjust to wearing a mask at the grocery store, or to having stores limit the number of people that enter. I don’t think the social distancing precautions will be going away. We might even learn to like this new way,” he said. “For stores, observing these precautions could become a selling point since social distancing measures allow for a more sedate shopping experience and (potentially) more fruitful interactions with salespeople.”

Andersen acknowledges that this is a crazy time – and this drives him to work extra hard to understand things through his research. With each new piece of awareness, a new question will arise that will lead to even more understanding.

“We are in an unknown world,” he said. “Five years from now I believe we won’t recognize the world we are in right now.”

This article first appeared in the Fall 2020 Minervanomics newsletter. It has since been updated.