Regular readers might remember how in May tech bro turned “Debate me, bro!” antivaxxer Steve Kirsch estimated that COVID-19 vaccines had killed a half a million people, all based on his survey. The survey itself was one of the most risible surveys that I’ve ever seen—a “simple survey of my readers,” as Kirsch put it—in which he just posted an Internet survey to his Substack that anyone could access and answer and then extrapolated from ~400 respondents to claim that COVID-19 vaccines have killed a half a million people. I kid you not. Yesterday, BMC Infectious Diseases published a paper that gave me very strong Steve Kirsch vibes. It’s by a Michigan State University economist named Mark Skidmore, who is the Morris Chair in State and Local Government Finance and Policy, Department of Agricultural, Food, and Resource Economics; Professor, Department of Economics. Entitled, The role of social circle COVID-19 illness and vaccination experiences in COVID-19 vaccination decisions: an online survey of the United States population, it reminded me, more than anything else, of a Steve Kirsch “study” in which a fundamentally bad research design is tarted up with (some) seemingly legitimate social science research methodology and then.dishonestly spun to produce a fake estimate of 278,000 fatalities due to COVID-19 vaccines, which is then “validated” by an incompetent dumpster dive into the Vaccine Adverse Events Reporting System (VAERS) database of the sort that I’ve been writing about since 2006.
Let’s dive in.
A “bait and switch” to promote an antivax narrative
The first thing that hit me about this paper is that it’s a bit of a bait-and-switch. On the surface, the primary aim of Prof. Skidmore’s survey is not at all unreasonable, as related in the abstract:
The primary aim of this work is to identify the factors associated by American citizens with the decision to be vaccinated against COVID-19.
Nor is Prof. Skidmore’s methodology—on the surface:
An online survey of COVID-19 health experiences was conducted. Information was collected regarding reasons for and against COVID-19 inoculations, experiences with COVID-19 illness and COVID-19 inoculations by survey respondents and their social circles. Logit regression analyses were carried out to identify factors influencing the likelihood of being vaccinated.
However, the true purpose of Prof. Skidmore’s survey rapidly becomes apparent, from what reads like a tacked-on analysis:
In addition, the proportion of fatal events from COVID-19 vaccinations was estimated and compared with the data in the VAERS database.
As Scooby-Doo would, say, “Ruh-roh!” I sense…an antivax assumption behind the survey, the same antivax assumption behind deceptive dumpster-diving into VAERS, namely that every report of a death or adverse event that occurs after vaccination must have been caused by vaccination.
And that appears to be exactly the assumption that Prof. Skidmore had. First, let’s look at an entirely unsurprising finding of the survey:
A total of 2840 participants completed the survey between December 18 and 23, 2021. 51% (1383 of 2840) of the participants were female and the mean age was 47 (95% CI 46.36–47.64) years. Those who knew someone who experienced a health problem from COVID-19 were more likely to be vaccinated (OR: 1.309, 95% CI 1.094–1.566), while those who knew someone who experienced a health problem following vaccination were less likely to be vaccinated (OR: 0.567, 95% CI 0.461–0.698). 34% (959 of 2840) reported that they knew at least one person who had experienced a significant health problem due to the COVID-19 illness. Similarly, 22% (612 of 2840) of respondents indicated that they knew at least one person who had experienced a severe health problem following COVID-19 vaccination.
It makes intuitive sense that people who knew someone who had experienced a significant health issue after COVID-19 would be more likely to be vaccinated, while those who know someone who experienced a severe health issue after COVID-19 vaccination—whether that health problem was caused by the vaccine or not, I hasten to add—would be less likely to be vaccinated. There is, I suppose, utility in demonstrating this point about vaccine hesitancy yet again and perhaps attempting to quantify the effect. I have no problem with that sort of survey. What I do have a problem with is when such a survey is used as a jumping-off point to support conspiracy theories involving vaccine “depopulation” and “died suddenly” narratives. Here’s what I mean:
With these survey data, the total number of fatalities due to COVID-19 inoculation may be as high as 278,000 (95% CI 217,330–332,608) when fatalities that may have occurred regardless of inoculation are removed.“Fatalities that may have occurred regardless of inoculation are removed”? How, pray tell, did Prof. Skidmore figure this out?
And the conclusion:
Knowing someone who reported serious health issues either from COVID-19 or from COVID-19 vaccination are important factors for the decision to get vaccinated. The large difference in the possible number of fatalities due to COVID-19 vaccination that emerges from this survey and the available governmental data should be further investigated.
See the bait-and-switch? There’s a reasonable conclusion (health issues after COVID-19 or COVID-19 vaccination influence people’s decision to get vaccinated) that’s used to buttress an antivax conclusion, namely that vaccines killed nearly 300K people. Prof. Skidmore then then JAQs off about the supposed discrepancy between this estimate and government statements about the safety of the vaccines, using his JAQing off to call for an “investigation.” True, at the very end Prof. Skidmore concedes that there are “limitations with using a survey to collect COVID-19 health information, particularly for a politicized health issue” and respondents “often interpret events with bias due to perceptions based on history, beliefs, culture and family background.” (Well, duh.) Unfortunately, these limitations are only mentioned in a paragraph near the very end of the paper and do not dispel the overall narrative.
I was also immediately struck by the phrase, “when fatalities that may have occurred regardless of inoculation are removed,” which made me immediately ask how Prof. Skidmore estimated this number. I’ll get into this in a minute. First, however, let’s look at how antivaxxers are spinning this survey.
Antivaxxer’s react to Prof. Skidmore’s survey
Let’s just say that…they got the message. They completely ignored the mundane finding that knowing someone who had serious health problems after COVID-19 or COVID-19 vaccination influences the decision to vaccinate and focused like a laser on the true message that Prof. Skidmore intended to convey, that vaccines are killing lots of people.
For example, here’s Dr. Aseem Malhotra:
And more antivax spin. I’ll restrain myself to just a few examples, because antivaxxers have been Tweeting links to this study since yesterday morning:
OK, I’ll stop. With Twitter. Meanwhile, on antivaccine Substack, which is linking to the study far and wide as slam-dunk “evidence” that COVID-19 vaccines are killing and injuring:
- New peer-reviewed study: >217,000 Americans killed by the COVID vaccines in just the first year alone! (From—who else?—Steve Kirsch)
- New survey finds shockingly high rate of side effects
- Covid Vaccines Killed 278,000 Americans by the end of 2021, Peer Reviewed Study Finds (the author will be familiar to regular readers)
- National Survey Finds finds 28% Personally know of a Death Caused by COVID-19 Vaccines (Dr. Peter McCullough, of course)
- On “excess deaths,” deadly crashes, fatal falls: Introduction to this week’s summary of those who have “died suddenly” (Mark Crispin Miller, unsurprisingly)
You get the idea. There is, of course, a fatal flaw in the very conception of the study’s “secondary”—really primary, but not represented as such—aim, which is to claim that the survey shows that there are lots of vaccine deaths that “they” don’t want you to know about.
The fatal and likely intentional flaw in Prof. Skidmore’s study
Regular readers know that the most significant limitations of the VAERS database are that anyone can report anything and that correlation does not necessarily equal causation. That’s why antivaxxers have long dumpster dived into the VAERS database and represented every report as being slam dunk evidence that the adverse event reported was due to the vaccine. Again, I’ve been discussing this issue 2006 and was incredibly frustrated how unprepared the CDC, public health authorities, and the government were to counter disinformation about vaccines based on VAERS. Indeed, it was entirely predictable to those of us who had been following the antivaccine movement for a long time that they would immediately weaponize VAERS against COVID-19 vaccines by ignoring the base rate of the adverse events reported to VAERS and portraying raw numbers as evidence that vaccines were killing and injuring hundreds of thousands, if not millions, and that’s exactly what they did throughout 2021 and beyond, starting within weeks after the emergency use authorization for the Pfizer vaccine. I started trying to explain very early after the vaccines rolled out and continue to do so, but the narrative has stuck, because a beleaguered band of skeptical bloggers and social media mavens really have little chance against the tsunami of antivax disinformation firehosed by the antivaccine movement—and people like Prof. Skidmore.
As antivaxxers do with VAERS, Prof. Skidmore appears to accept nearly every respondent’s report of a death from the vaccine as having been because of the vaccine. First, though, let me note that the criticism that this is just an “Internet survey” that I’ve seen from some critics on Twitter is a bit off-base. Yes, it is a survey administered online, but it’s not just a survey that anyone can fill out and self-selects for antivaxxers, like Kirsch’s pseudosurvey. Rather:
The sample was obtained by Dynata, the world’s largest first-party data platform, and is representative for the US American population . The sampling using Dynata is based on opt-in sampling, respondents deliver high quality data, they are diverse and have community norms of honesty and accuracy . The survey was opened to the Dynata panel until the required number of responses was obtained from each category of the stratification variables age, sex, and income, as required for a balanced response set. With opt-in sampling there is no response rate as classically defined in survey research.
We can argue about whether companies like this provide samples that are truly representative of the sociodemographics of the US population or not; for my purpose discussing this study it doesn’t matter much. Even if Dynata’s sample is as close to a perfect reflection of the US population as is humanly possible to attain, the spin being put on this survey (that it’s evidence of hundreds of thousands of vaccine deaths) and the use to which it is put (that the government is “hiding” hundreds of thousands of vaccine deaths) would still be utter dishonest nonsense. Here’s what I mean.
The first problem to deal with is that a survey like this is just not a valid method to calculate a valid and accurate estimate of the number of deaths due to…well, almost any cause.It’s even less so in this case, given that deciding whether any given death is plausibly due to a vaccine is not a straightforward determination. When someone in a person’s circle of relatives, friends, and acquaintances dies, generally that person will just take the family members’ word for what the cause of death was, regardless of whether the family member is correct or not. The authors even note this in one of their results, sort of:
Estimated nationwide COVID-19 vaccine fatalities based on the Democrat, Republican and Independent subsets are 109,564, 463,444 and 247,867, respectively. With the vaccinated and unvaccinated subgroups, estimated COVID-19 vaccine fatalities are 110,942 and 659,995.
There’s actually an interesting observation there, namely that political and ideological orientation has a huge effect on whether a person attributes the death of a loved one, friend, or family member to COVID-19 vaccines.This observation, however, says absolutely nothing about whether that death of a loved one, friend, or family member was actually caused by COVID-19 vaccines. Indeed, the huge discrepancy in the numbers based on ideological groups should tell you that these numbers are perceptions, not reality. In the hands of a researcher not hellbent on spinning these results into antivaccine disinformation, these results could have served as the preliminary data from which to develop interesting hypotheses to test in further studies. Prof. Skidmore is an economist, however, not a social scientist, and instead he uses these numbers to argue that we should “investigate” the apparent discrepancy between these numbers and the government’s figures. In other words, he uses them to bolster the antivax “died suddenly” conspiracy theory that claims the vaccines are killing huge numbers of people but “they” are “covering up” the evidence.
Don’t believe me? Check out this passage:
The ratio of COVID-19 vaccine deaths to COVID-19 illness deaths of the people respondents knew best who had health problems is 57/165 =0.345, whereas the ratio of vaccine-associated fatalities to COVID-19 fatalities from government sources is 8023/839,993 = 0.0096. The null hypothesis (H0) that the true ratio, X, is equal to the CDC ratio which is also equal to the survey ratio: X = CDC Ratio = Survey Ratio.
This hypothesis is tested using state-by-state VAERS data on reported COVID-19 vaccine-associated deaths and COVID-19 illness fatalities. The alternative hypothesis (Ha) is: X = CDC Ratio < Survey Ratio. The mean (u) and standard deviation (σ) of the ratio of vaccine fatalities to COVID-19 fatalities from the state-by-state data are u = 0.0136 and σ = 0.0111. The probability that the Survey Ratio > CDC Ratio = X is P(CDC Ratio > 0.345). With P(CDC Ratio > 0.345) = 0 and a Z-score = 28.86; the null hypothesis is rejected.So science-y! So statistics-y! So wrong!
So basically, Prof. Skidmore, for reasons known only to him, is testing a fantasy hypothesis that his estimate of how many deaths have resulted from COVID-19 vaccines based on a survey that that reflects perceptions of the respondents, not actual medical reality should match government statistics and declares the hypothesis falsified when the two numbers are very different from each other. Then he does the same thing with VAERS. Truly, this is fractal pseudoscience whose abuse of frequentist statistics is truly impressive.
That’s not all. It gets even more depressingly hilarious when Prof. Skidmore tries to estimate how many people would have “died anyway” not necessarily because of the vaccines, in order to subtract that number from the total and seem more “reasonable.” Let’s just say that he gives a master class in how not to avoid the base rate bias, in which people tend to ignore the base rate of a particular adverse event and attribute all such events to the cause they are examining, in this case vaccines.
I explained this in detail almost two years ago and then again in October, but, in brief, according to the CDC the yearly number of deaths in the US in 2019 was 2,854,838, for a rate of 869.7 deaths per 100,000 population. In other words, less than 1% of the US population dies every year, but that’s still a huge number. Immediately before the pandemic around 7,821 people died each and every day. If you start vaccinating millions of people in a short period of time, by random chance alone and the law of large numbers, there will be a large number of people who die within, say, 30 days of being vaccinated even if the vaccine has absolutely nothing to do with their deaths. That’s why VAERS numbers, divorced from baseline rates, are utterly meaningless, and, as the VAERS website itself states, VAERS cannot reliably determine incidence and prevalence of an adverse event (AE) after vaccination. That’s not what it was designed to do. Rather, it was designed as an early warning system to generate safety signals that produce hypotheses of specific possible AEs after vaccination to be investigated further, and that’s how the CDC uses it.
The same reasoning applies to Prof. Skidmore’s survey. By random chance alone a lot of people will know (or hear of) other people in their extended social social circle who died within a timeframe after being vaccinated against COVID-19 that makes it plausible to a human pattern-seeking brain to think that the vaccine caused it. Again, he just assumes that if a respondent attributed a death to the vaccine, that it must have been the vaccine, with a small “correction”:
This evaluation is conducted under the assumption that the reported vaccine-related fatalities and injuries are caused by the COVID-19 vaccine but is now relaxed by reducing the number of reported fatalities by the fatalities due to other causes that would be expected to have occurred anyway. An estimated 2.43 fatalities might have occurred from heart attacks, strokes and blood clots within the survey sample regardless of vaccination status. Subtracting these fatalities from total estimated vaccine fatalities generates a nationwide estimate of 278,000 fatalities, which is 4.1% smaller. Estimated total adverse events are correspondingly reduced by 4.1%.
Let’s just say that this is not how you correct for the base rate of these events in the population! Actually, it’s impossible to “correct” for anything without the medical records and death certificates of all the people in the respondents’ social circle who died. Again, a survey of lay people—and even of medically trained people—is not a reliable method to estimate deaths due to pretty much any cause. For that, you need systematically collected data analyzed by epidemiologists and statisticians who know what they are doing, a description that most certainly does not apply to Prof. Skidmore.
After reading Prof. Skidmore’s survey results, I was left with three questions:
- Why did Prof. Skidmore do this study?
- Who funded Prof. Skidmore’s survey?
- How on earth did BMC Infectious Diseases publish such dreck?
The first question was answered easily with bit of Googling quickly led me to a website maintained by Prof. Skidmore called Lighthouse Economics, where he maintains a personal blog. Just a brief perusal of his blog demonstrates that he’s a hard core antivaccine conspiracy theorist. If you don’t believe me, just peruse some of these entries for yourself, such as:
- Technocrats-At-Work: Medical Profession Implements WHO’s Digital Diagnosis Code For The Unvaxxed
- All Hell Breaks Loose: Elon Musk Tweets He Felt He Was Dying After Booster, Young Cousin Hospitalized With Myocarditis; Scott Adams Says Anti-Vaxxers Were Right, Protests Outside BBC, Fox News VP Dies
- Africa Is Starkly Unvaccinated, and Starkly Unvanquished by COVID
- MRNA Vaccine is a “Shotgun Blast” to Brain and Reproductive Organs
- “Fully Vaccinated” 27 Year Old Model Dies Suddenly
You get the idea. Prof. Skidmore did this survey because he’s antivax.
More interestingly, the Funding section reveals that this entire survey was funded by one person, Catherine Austin Fitts. I had never heard of Ms. Fitts; so Google was my friend. It didn’t take long to find out that she’s worked with Robert F. Kennedy, Jr. before, with a Washington Post article noting:
The viral clip from “Planet Lockdown” depicts an interview with Catherine Austin Fitts, who served as assistant secretary of housing and urban development under President George H.W. Bush and has since worked in finance. Fitts, who has no background in medicine or public health, has worked with anti-vaccine activist Robert F. Kennedy Jr. to promote unfounded claims about the pandemic and to oppose lockdown measures put in place to slow the spread of the virus.Quelle surprise! The funder of Prof. Skidmore’s “study” is an antivax loon.
So this survey was done by an antivaxxer and funded by…an antivax conspiracy loon!
It’s also apparently been in the work for a year. Buried in a list of papers in a Steve Kirsch Substack entitled License to kill (and how to redpill patients) there’s a link to a paper published as a PDF on Prof. Skidmore’s site as a “working paper” entitled How Many People Died from the Covid-19 Inoculations? An Estimate Based on a Survey of the United States Population*. Interestingly, the primary and secondary aims are identical to the aims in the version published yesterday in BMC Infectious Diseases, but flipped. The primary aim in the old “working paper” dated February 28, 2022 appears to have been to examine “the potential fatalities and injuries from the COVID-19 inoculation using an online ‘Covid-19 Health Experiences Survey’ administered to a representative sample of the United States (US) population,” while almost as an afterthought Prof. Skidmore writes that “I also analyze the factors that influence the likelihoods of being inoculated, experiencing an adverse event, and knowing someone who was injured by the COVID-19 inoculation.” (Emphasis mine.)
Looking at the paper in light of this nearly year-old precursor (which quotes antivax luminaries like Sen. Ron Johnson and Informed Choice Australia—even the execrable Stephanie Seneff “review” article on vaccine harm), I’m even more convinced that it is a total bait-and-switch. I really do think that Prof. Skidmore did the survey to come up with an inflated estimate for the number of COVID-19 vaccine “deaths” but also did an analysis of factors leading to COVID-19 vaccine hesitancy, which would be much more palatable and likely to be published in a peer-reviewed journal. He then shoehorned in the inflated, implausible, and unproven death estimates, whose successful publication in a reputable peer-reviewed journal was the real purpose of this survey. The “working paper” version of the manuscript is a “greatest hits” collection of antivaccine tropes, distortions, and misinterpretations of scientific studies; the published version is cleaned up just enough to make it past peer review by people clearly not familiar with antivaccine disinformation but still keep the claim that COVID-19 vaccines caused 278,000 deaths in the final version, so that antivaxxers can point to it instead of the correlations between various responses and likelihood to accept the COVID-19 vaccine.
My final thought involves peer review and ethics. First, how did this awful manuscript get through BMC Infectious Diseases peer review? This is one of the cases where an open peer review process is quite…illuminating. I perused the peer review reports and was amazed and disappointed. I’ll give Prof. Skidmore credit for one thing; he’s persistent. He submitted a total of five versions of his manuscript, and it appears to have been reviewed by only two peer reviewers, one of whom allowed himself to be identified. This reviewer was Yasir Ahmed Mohammed Elhadi, who appears to be a Sudanese researcher who is now student in the Department of Health Administration and Behavioral Science, High Institute of Public Health, Alexandria University. Let’s just say that his comments on the manuscript and its revisions were not exactly rigorous, either, nor were the other reviewer’s comments. Neither of them noted the fatal flaw at the heart of the assumptions underlying this study, specifically how you cannot infer causation from the perceptions of respondents reported on a survey like this. Neither noted the base rate fallacy at the heart of this study. Neither noted that you can’t infer causation from raw VAERS data. In short, neither appeared to know what they were talking about.
Finally, how did this study manage to get approval from MSU’s IRB, anyway? My guess is that the survey looked looked like a straightforward assessment of factors that affected people’s decision to be vaccinated against COVID-19, without any obvious indication how the authors would use the results to deceptively claim that close to 300,000 people had been killed by COVID-19 vaccines and then bolster a conspiracy that the government was somehow hiding this carnage. Again, the whole study strikes me as a bait-and-switch to get antivaccine disinformation published in a peer reviewed journal. It might be worth asking the chair of the IRB.
The bottom line is that this survey, although seemingly a bit more rigorous than the typical antivax survey only in that it used a company that provides a respondent pool designed to mimic the demographics of the US, is at its heart no different than the sorts of bizarre Internet surveys about vaccines that antivaxxers have been doing for years and years and then spinning as “proof” that vaccines are harmful. MSU should be embarrassed that it has on its faculty such a conspiracy monger willing to use bogus research techniques to promote antivaccine disinformation.