BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Study: Trump Rallies Linked To Over 700 Covid-19 Coronavirus Deaths, Here Are The Caveats

Following
This article is more than 3 years old.

Gee, what’s the best thing to do during the Covid-19 coronavirus pandemic? How about crowd together people who are not wearing face masks and have them scream and shout? That may not be the best thing for people, but from the virus’s perspective, it could be a veritable buffet.

Indeed, public health experts have expressed concerns that President Donald Trump’s campaign rallies have resulted and will continue to result in more and more Covid-19 coronavirus cases and deaths. Just listen to what Anthony Fauci, MD, who is known as the Director of the National Institute of Allergy and Infectious Diseases (NIAID) and not as a neuroradiologist, said when Trump resumed his campaign rally tour after the President’s recent bout with Covid-19:

Umm, “asking for trouble” is typically not a good thing to do, unless you happen to be requesting the song by Pink during a Saturday Zoom Halloween party. So exactly how much trouble have Trump’s campaign rallies already caused?

That’s not an easy question to answer since the U.S. doesn’t have a great national Covid-19 coronavirus surveillance system. In this case, “doesn’t have a great” means “doesn’t really have” or “has a WTH is happening.” Infections and resulting deaths could easily go undetected. Nevertheless, a team of economists (B. Douglas Bernheim, Nina Buchmann, Zach Freitas-Groff, and Sebasti´an Otero) have given it the old college try. The team from Stanford University has tried to answer this question, at least for some of the President’s campaign rallies. They have estimated that Trump’s rallies held between June 20 and September 22 have already produced over 30,000 more infections and over 700 more deaths from the severe acute respiratory syndrome coronavirus 2 (SARS-CoV2). That sounds potentially super, as in super spreader-ish. Take these estimates with some upside-down baseball caps full of salt though.

First of all, the study has not yet been published in a peer-reviewed scientific journal. Instead, it’s been posted on a server, which basically means that the authors had Internet access and knew how to upload a document. Therefore, it has not yet been reviewed and vetted by experts in public health, epidemiology, and infectious diseases to determine how valid the study’s methods, results, and conclusions may be. There is a huge difference between a study that’s been posted on a website somewhere versus one published in a reputable scientific journal. In this case, huge means 6900-pound-cheesecake-huge. It’s like being told that a necklace is from a person named Tiffany versus from Tiffany & Company. Or comparing some random dude with an cowboy hat singing “Rolling in the Deep” in a karaoke club at 3 am with Adele singing the same song in the studio for her official record. Or a Zoom beard versus a real beard.

Plus, while not unusual, the approach and methods used for the study do have a number of inherent weaknesses and limitations that may not be readily evident like the Invisible Man trying to get a license at the DMV. So before drawing any stronger conclusions, take a deeper dive into how exactly this study was done.

Let’s start with a snapshot of the data that was used. The list of Trump campaign rallies came from a Wikipedia. While Wikipedia may be a great place to learn about the Dog-Fart Roller Coaster as well as fart lighting, which is not related to the roller coaster and you should never try, it is not exactly the most accurate scientific reference. Nonetheless, for the purposes of this study, a Wikipedia list is probably a reasonable source since Trump doesn’t tend to hide where he is holding his rallies and how people can sign up to attend or, in the case of Tik-Tok teens and K-pop fans, sign up and not attend.

Included in the study were three indoor rallies held by the Trump campaign in Tulsa, Oklahoma, on June 20, Phoenix, Arizona, on June 23, and Henderson, Nevada, on September 13. Then there were the following 18 outdoors campaign rallies (yes, 18) that Trump held during that time period:

  • Mankato 8/17/2020
  • Bemidji 9/18/2020
  • Oshkosh 8/17/2020
  • Fayetteville 9/19/2020
  • Yuma 8/18/2020
  • Old Forge 8/20/2020
  • Londonberry 8/28/2020
  • Latrobe 9/3/2020
  • Winston-Salem 9/8/2020
  • Freeland 9/10/2020
  • Mosinee 9/17/2020
  • Swanton 9/21/2020
  • Vandalia 9/21/2020
  • Pittsburgh 9/22/2020
  • Minden 9/12/2020
  • Jacksonville 9/24/2020
  • Newport News 9/25/2020
  • Middletown 9/26/2020

As Yoda might say, lots of rallies are those. This is certainly not a list of all Trump rallies that have been held during the pandemic. As you can see, the entire month of October was not included. This is noteworthy since Covid-19 coronavirus transmission has been picking up as the weather has gotten colder and less humid. At the same time, Trump has been intensifying his campaigning like a hot dog vendor near the end of a football game because rumor has it that Trump is interested in getting re-elected. Therefore, the study may be missing a sizeable percentage of the SARS-CoV2 transmission that may have occurred this year from Trump’s rallies.

The next step was to pull data on subsequent reported Covid-19 cases and deaths in those areas from the Johns Hopkins University Coronavirus Resource Center website, which in turn compiles data from Centers for Disease Control and Prevention (CDC) and U.S. state and local governments. Note that the key word here is “reported.” Reported does not mean actual. Just like the number of people who actually admit that they read 50 Shades of Grey may not quite match up with the number of book sales, reported Covid-19 numbers likely underestimate the true numbers of deaths and cases. After all, Covid-19 testing in the U.S. has been as haphazard as a puppy in a blindfold with a martini in hand driving a race car in the woods in the dark. The federal response hasn’t exactly been like the Men in Black. There’s still no federal system tracking exactly what percentages of different populations have been tested and how often.

The team also obtained data on county-level demographic and election data, testing, and what they deemed “COVID-related” policies, such as county-level policies, such as shelter-in-place orders and mask mandates from different sources.

Armed with this data, what did the team of economists do? Let’ see. It rhymes with the “stan rum expressions” or “fan some Jeff Sessions.” The team, drum roll please, ran some regressions. That basically meant comparing the increases in cases and deaths in counties after they had Trump rallies with those in “similar” counties that didn’t have such rallies. They adjusted this comparison by factoring in some of the other aforementioned data such as the presence and absence of mask mandates and then came up with the associated or correlated increases in reported Covid-19 cases and deaths. That’s how they came up with an estimated average of 261 more reported Covid-19 cases per hundred-thousand residents in the counties that had held Trump rallies. Multiplying this number by the total number residents in all of the counties that had Trump rallies and the infection fatality rate for the Covid-19 coronavirus then yielded the over 30,000 cases and over 700 deaths numbers.

So should you hang your baseball cap or beret on these numbers? Again, keep in mind that the research team dealt only with reported data on Covid-19 coronavirus cases, which may be nowhere close to the actual number of Covid-19 coronavirus cases. Remember that the next time your friend knows way too many details about “merengue with some salsa moves” from 50 Shades of Grey but continues to deny ever reading the book. Moreover, one must “regress” to inform you that these numbers stemmed from merely associations, correlations. Regressions can only show correlation, not cause-and effect. Remember, correlation does not mean causation. Repeat, correlation does not mean causation. Say it like Yoda would. Causation means correlation does not.

Think about how many different things can be associated with one not really causing the other. For example, the rise of the Kardashians may have correlated with the rise of the stock market. Does that mean that Kim Kardashian and her sisters can claim credit for the booming stock market throughout the 2010’s? Probably not. Neither can the increasing use of the words “YOLO” or “stan” or the increasing number of pictures of Russian President Vladamir Putin going topless. The increasing availability of avocado toast, of course, is a different story.

Such regression models (not to be confused with other types of computer models or Gigi Hadid or Putin going topless) tend to oversimplify situations and not take into account the complexities of the Covid-19 coronavirus and its spread and impact. Many other factors such as changing temperature and humidity, compliance with different state and country recommendations regulations, and social mixing patterns can affect the transmission of the SARS-CoV2 in numerous complex ways not captured by this regression model. In other words, this study is like trying to find one word to describe the Michael Sorrentino character on the Jersey Shore TV reality show. It’s an oversimplification of the situation.

So take this study’s numbers like most numbers that you see on a Tinder profile. You can say “interesting” and rub your chin while furrowing your eyebrow, as long as you have washed your hands before doing so. But before you bank on the specific numbers listed wait until there is more confirmation of their accuracy and validity, through scientific peer-review and by other methods besides regression. Treat any study that claims that something is causing the spread of the Covid-19 coronavirus like hearing that your significant other is on a call with Jake from State Farm at 3 am in the morning. Double-check that it is actually Jake from State Farm and that he’s wearing khakis. When it comes to the actual number of Covid-19 coronavirus cases and deaths caused by Trump’s campaign rally, to quote a recent Saturday Night Live (SNL) sketch, we don't know dis.

That being said, it wouldn’t be surprising if the Trump rallies were to have increased the number of Covid-19 coronavirus infections and deaths. After all, such rallies have gone against a number of different public health recommendations during this pandemic. People including Trump have bragged about how many people have attended these rallies:

Some of the rallies have been indoors. People have stood or sat within six feet or one Kid Rock (since Kid Rock is approximately six feet tall) of each other for an extended period of time. Many people have not worn face masks, meaning effectively covering their mouths and noses and not just wearing a Halloween mask. Among those who have worn masks, a number have left their noses exposed, which is a bit like wearing underwear but leaving, well, you get the picture. People have cheered and shouted things like “lock her up” and thus potentially spreading into the air virus-laden respiratory droplets from their noses and mouths. With all of this going on, who knows what kind of rally the SARS-CoV2 may have had or be having?

The rallies themselves aren’t the only things that may have promoted the spread of the Covid-19 coronavirus. Political rallies are like baked lasagna. They may involve layering things on and may be cheesy. They also involve preparation and an aftermath. Unless the Space Force has developed a transportation machine, for each rally, attendees have had to travel to the rallies and wait in lines to enter the venues with each other. After the rallies they’ve had to wait together for transportation for at least a little while. Sometimes, as this Sky News Australia report showed, for a very long while, perhaps even an oh-bleep-it’s-cold-need-to-go-to-the-hospital while:

This pandemic ain’t no party, this ain’t no disco. This ain't no fooling around, in the words of the Talking Heads (the band and not political commentators on TV). It’s just ain’t wise or responsible to hold such mass gatherings during the pandemic, especially as the number of reported Covid-19 coronavirus cases per day continues to rise. That’s like going to an active fire wearing a Count’s or Contessa’s outfit made out of toilet paper and dryer lint and saying, “this is going to go real well.” There’s really no justification for such events when either remote speeches or at least events that enforce more social distancing will still allow you say things such as “in California, you have a special mask. You cannot, under any circumstances, take it off. You have to eat through the mask.” Continuing to say that our country is "rounding the turn" or “rounding the corner” when the turn may be next to a cliff or finding others to blame for the pandemic during rallies is not going to make the Covid-19 coronavirus go away. In fact, the Covid-19 coronavirus likes campaign rallies but is probably not listening to what’s being said.

Full coverage and live updates on the Coronavirus

Follow me on Twitter or LinkedInCheck out my website