Taxpayer-Funded Study Pushes False Narrative about Amish and Mennonite Excess Deaths During COVID-19

An interesting challenge was posed recently by Steve Kirsch, an entrepreneur and founder of the COVID-19 Early Treatment Fund and the Vaccine Safety Research Foundation, who asked in a tweet if anyone could name just five unvaccinated Amish people who died from COVID-19.¹ Anecdotally, a couple of people who say they work closely with the Amish shared that they knew few to no Amish people who died from COVID-19. Thinking that this is a great case study, and since I also happen to live in northeast Indiana where a number of Amish communities reside, I responded to his tweet by saying that I would be willing to assist in any sort of study looking at the Amish and vaccines. Shortly thereafter, someone replied to my response with a link to a study where the authors conclude that a large spike in "excess deaths" among the Amish and Mennonite communities in 2020 was due to COVID-19. The study is titled, "Closed but Not Protected: Excess Deaths Among the Amish and Mennonites During the COVID-19 Pandemic,"² and upon reading it, it immediately set off a number of red flags. Since this study keeps cropping up every time the question is raised about how the largely unvaccinated Amish fared during COVID-19, I decided to take a closer look at this study. Here I present what is ultimately a thorough debunking.

At the center of the study is a small Ohio newspaper called The Budget, which prints, among other things, obituaries for Amish communities all over the country. Situated in the heart of a small town called Sugarcreek, Ohio, The Budget sends out about 20,000 copies of its paper every week to Amish and Mennonite communities in about 40 different states. About 1,100 Amish "scribes," each representing one or more community, are known to correspond at times with The Budget, either through post, fax, or email (the latter two methods presumably via an intermediary). While official death records of closed religious communities like the Amish are incomplete at best, the obituaries sent in by the scribes to this newspaper were declared "an efficient index of mortality" by the authors of this study in question. Using this as the sole source of data for the study, the authors go on to calculate "excess death" percentages for 2020 and declare that these were due to COVID.

Before we get into more detail about their methodology here, I want to take a moment to establish some pertinent facts and general context to set the stage. Obituaries are not, never were, and never will be a thorough, comprehensive record of all-cause mortality, and the authors' description of them as an "efficient" record of mortality seems disingenuous on its face. A rough estimate is more than sufficient to illustrate this crucial context: if we estimate that The Budget covers 75% of the total Amish/Mennonite population,³ or 375,000 out of about 500,000 total people, and we use an average death rate of 9 per 1,000 people,⁴ we can then establish that roughly 3,375 people die per year in The Budget's coverage area. On average, only about 400 obituaries are printed by The Budget per year,⁵ which is only about 12% of the total deaths in its coverage area (this rose to only about 16% in 2020). In developed countries with reliable census data and vital statistics, records of all-cause mortality approach 100% of total deaths, and this is the only kind of data upon which we can reliably calculate excess mortality. Realize that this is damning to the entire premise of the authors' taxpayer-funded study about "large numbers" of Amish/Mennonite "excess deaths," but let us entertain them for a little while longer, shall we?

In their paper, which was published in the Journal of Science and Religion in June of 2021, the authors explained their methodology as follows:

"We coded all obituaries published in The Budget from 2015 through January 2021 for the [deceased’s] death date, age at death, sex, and the state in which the death occurred. Our dataset excludes deaths occurring before January 1, 2015, and after December 31, 2020, yielding 2,438 cases across 34 states."

Following that, the authors compared the average number of obituaries in 2020 to a 5-year baseline average (2015-2019). The change was represented as an "excess death" percentage, which was calculated on a month-to-month basis (they did this both for the entire dataset and for the state of Ohio exclusively). So for example, they calculated that in November of 2020, "excess death" was up by 125% compared to the baseline average. The authors actually highlighted this particular figure, which happens to be the highest calculated "excess death" rate, in the abstract of the paper. The authors also pointed out in their study what they saw as "similar patterns" between their month-to-month "excess death" calculations and the three waves of COVID-19 that spread across the U.S. in 2020.

The authors eventually conclude that "The large number of excess deaths among the Amish and Mennonite community (. . .) indicates not only the presence, but the impact of COVID-19. . .” They remark that this is "concerning," but interestingly, they do not contextualize that concern in such a meaningful way as to compare their calculated "excess death" percentages to the national average. We will take a look at that later. Another main takeaway that the authors want you to have from this paper is, as stated in the abstract, that "Amish/Mennonite excess death rates are similar to the national trends in the USA." This is another point of contention that I have with this paper that we will get into shortly.

So even before we get to crunching some numbers here, it's plain to see that the only evidence presented in this paper is an increase in reports of deaths in 2020—there is no evidence at all that the actual number of deaths increased, nor any evidence that any of these additional reported deaths had anything to do with COVID. Other perfectly logical possibilities for why obituary reporting went up, in a year of great social upheaval mind you, are not even considered in the paper (we will look at this later). But the only thing that can be confidently concluded here is that there was a rise in death REPORTS in 2020—that's it. As previously mentioned, The Budget normally prints obituaries for about 12% of the total deceased among its representative population, and in 2020 that went up to about 16%. Clearly there is still a lot of room for an even greater rise in obituary reporting before we can even begin to talk about "excess deaths." For the authors to misconstrue this rise in reports as "excess deaths" and go on to conclude that it "indicates (. . .) the impact of COVID-19 on this community" is not only intellectually dishonest, but as will become more and more apparent, brazenly misleading.

Consider the repeated misuse of the phrase "excess deaths," even in the very title of the Stein et al. paper. Excess death or excess mortality calculations require thorough, regularly-updated data on all-cause mortality⁶ for a target population that is clearly defined,⁷ yet neither of these criteria were met by the authors. They had a small newspaper that only prints obituaries for an estimated 12% of the deaths among the population that it represents, and according to Milo Miller, The Budget's publisher, he never "bumps obits,"⁸ meaning that he never doesn't print obituaries that he receives. So The Budget is hardly a thorough, comprehensive record of all-cause mortality, neither for the total Amish and Mennonite population nor for the population that The Budget represents. The latter, being precisely unknown, is the other unmet criterion for calculating excess mortality. The authors simply DO NOT HAVE THE DATA to calculate "excess death" or "excess mortality" in the literal, commonly understood sense of the phrase (and well-established in the literature. See here and here). Their attempt to conflate "excess deaths" with "reported deaths" strikes me as being deliberately misleading; something too about the way that they use the same calculation with the same parameters (using a five-year baseline average) as a legitimate excess mortality calculation, yet none of the necessary data. This could all be an honest mistake on the part of the authors, but then again, I'd have to bring up the fact that this is a government-funded study—to the tune of $258,719 via an NSF grant for a group of four studies—that just so happens to fall in lockstep with the government's own messaging, true or not, of this being an indisputably "deadly" pandemic that we have to "vaccinate our way out of." There is a whole host of groups and actors and motivations (other than the truth) that we could unpack here, but I will leave that for another time.

So let us pretend for a moment that we are the authors and we have this dataset from The Budget. The honest and straightforward way to talk about this data is to say that there is a rise in REPORTED deaths. After that, we are all free to speculate as to why there was a rise in death reports, as long as we admit that that's all it is: speculation. I have no problem with speculating that some amount of these deaths were from COVID-19, and I have almost no doubt that some amount of these reported deaths were, in fact, due to COVID-19. But to be intellectually honest, we cannot stop there. We have to consider other possible factors as well, because again, I stress, we do not have the data to conclude anything with certainty, despite what was falsely projected by the authors of the study.

Bearing in mind that the year 2020 was a time of great social upheaval, let us now consider some other possibilities for why there was a rise in REPORTED deaths to The Budget during this year. The Amish and Old Order Mennonites use limited-range horses and buggies for transportation closer to home, yet they allow and rely on transportation from others when they need to traverse express highways, enter high-traffic, inner-city areas, or travel distances longer than about twenty-five miles. Not only did everyone face government-imposed lock-downs and travel restrictions during the pandemic, social distancing guidelines and ultimately self-imposed restrictions by those who did not want to travel with passenger(s) in their cars (considering too that the Amish are averse to wearing masks) all contributed to a reduction in travel for the Amish in 2020. This led to an overall decline in social interactions. For example, the ability to attend larger events and gatherings, especially distant ones, was significantly hampered in 2020.

Overall, face-to-face interactions, where news and information about other, particularly more distant communities is shared, declined. Now keep that in mind when considering what Amish scribes actually do—they do not just send obituaries to The Budget, but also local news, marriage reports, births, illnesses, etc. In fact, the scribes actually produce the content of the newspaper, while Milo Miller and his team at The Budget are basically only charged with editing and assembling the final package.⁹ Therefore, it is only logical to speculate that the Amish and Mennonites would want to step up other communicative efforts to regain what is lost via the decline in in-person social interactions, and the scribes perfectly fill this role. Let us also consider that obituaries are particularly compelling information to share and seek out during a pandemic, when fear and concern for distant friends and relatives are heightened. Obituaries may also provide a small sense of closure to those who may not be able to travel to attend a funeral. All of these are logical, common sense considerations that were completely absent from the study, and it doesn't stop there.

Consider the simple fact that since many special events and gatherings were cancelled in 2020, people in the Amish and Mennonite community, including scribes, had more time on their hands. Keep in mind that a lot of the dispatches sent out from the scribes come in the form of hand-written letters. Remarkably, the same authors found in a separate study about Amish/Mennonite coping strategies during COVID-19, which was produced under the same quarter-million dollar grant as the study in question here, that 24% of scribe content and other entries to The Budget between March and April 2020 mentioned "having more time and a slower pace of life as a benefit of the pandemic." So the authors' own study, published three months prior to the one in question here, contains valuable context for interpreting the rise in obituary reporting, yet they fail to even mention it here. Considering the gross mischaracterization of their core data, it seems this latest bit of information fits a pattern not of incompetence, but of malfeasance.

The bottom line here is that there's simply not enough evidence to conclude why the number of obituaries went up in 2020, neither for the authors' conclusion of COVID-19's sole responsibility nor for my more logical conclusion that it was probably some combination of all of the factors described above—in conjunction with COVID. The degree to which each played a part, we will probably never know.

Let me ask an interesting question: if the authors of the study truly believed that all or most of what amounted to a roughly 4% rise in obituary reporting was "excess deaths" indicative of the "impact of COVID-19," then why didn't they sound the alarm that the "excess death" percentage among the Amish/Mennonites would be MANY TIMES HIGHER than the national average? The excess death percentage for the United States was already preliminarily calculated as early as January of 2021 at about 12% (see here), and it continued to be refined in the following months, well before the June, 2021 publication of the Stein et. al paper in question. From their perspective, if they even bothered to do the full-year calculation from their data, the Amish/Mennonite "excess death" percentage for 2020 was almost four times higher than the national average—sitting at 47%. While the national excess death percent for 2020 has since been revised to about 15% (see here), that’s still significantly less than 47%. Let's not lose sight of what this would actually mean—many more people would be dying according to the authors' own interpretations and conclusions from the data. How many? Using the combined population of Amish and Mennonites of about 500,000, a figure which appeared in the Mennonite World Conference's World Directory of 2018,¹⁰ the average yearly deaths (using the previously-established crude rate of 9 per 1000) for the total Amish/Mennonite population would equal 4,500. 15% excess deaths would mean that 675 more people died in 2020, but 47% excess deaths would mean that 2115 more people died in 2020, an increase of 1,440 people above and beyond what one might expect even after COVID-19! Yet rather chillingly, the authors don't make such qualitative and quantitative assessments as these, and contrary to their own interpretation of the data, they even state in the abstract that "Amish/Mennonite excess death rates are similar [emphasis mine] to the national trends in the U.S.A." It seems rather bizarre and contradictory, but then again, by now you understand why they probably don't want to draw too much attention and scrutiny to their faulty, misleading study.

All of this does point to the need for further research, for at this point in time there is little to no definitive conclusions we can draw from such incomplete data. Unfortunately, it would be too late to help the Amish and Mennonites to know if their population was indeed more susceptible, for whatever reason, to one of earlier, more damaging variants of COVID-19. Still, I would like to pursue further avenues of research, such as amassing individual case studies involving direct interviews with individual community scribes and a thorough examination of their records. What's more, The Budget itself is an immense resource chocked full of weekly first-hand accounts of life in Amish and Mennonite communities in 2020. All 52 editions for that year, along with the every edition of the paper going back to its inception in 1890, are all digitized and searchable at the Ohio Amish Library inside the Amish & Mennonite Heritage Center outside of Berlin, Ohio.¹¹ Why not utilize this wonderful resource to research the full context—what was really going on in Amish and Mennonite communities during 2020 and beyond?¹² Extending this research into 2021 is another avenue that would help to answer the many open questions and hypotheses that one could investigate, with the ultimate goal, as we will soon discuss, of greater health and greater protection for potentially large swaths of the U.S. population. I hope to write a continuation of this article where I personally look into this further.

Before concluding, there is one final matter that I would like to address. That is, the argument in the Stein et al. paper that the month-by-month pattern of excess reported deaths in The Budget during 2020 was "similar" to the overall U.S. trend of excess deaths that corresponded to the three so-called "waves" of COVID-19 that swept through the country. You can compare this yourself using the two pairs of graphs below. Note that for each pair of graphs, I could not plot both lines on the same graph because we are not comparing apples to apples here. The authors are basically using this as a way to say, "Look, this really IS an apple," and by "apple" they mean "excess deaths." However, such side-by-side comparisons as you will see below were not offered in their study, and I believe you can see why. This is extraordinarily weak evidence, if it qualifies to be called "evidence" at all. Figures 1 and 3 below were recreated from this graph, although the values were averaged by month to better compare with the Stein et al. data in Figure 2, which comes from Table 1 in their study. In Figures 1 and 2, which are scaled differently on the y-axis, there are similarities to be noted like the upward trend line in both graphs and the presence of distinct "waves," but there are differences as well, such as the fact that the two graphs generally don't align, especially in months like June, August, and December. For Figure 3, I merely readjusted the y-axis scale of Figure 1 to match that of Figure 2. Another difference emerges here: Figure 2 appears as a rather erratic, "high peak-to-peak amplitude signal" when compared to the much smoother shape of Figure 3. Maybe if you squint your eyes really hard, you can see a similar pattern in the graphs below, but I think it's pretty clear that this is evidence of absolutely nothing. If the authors made this argument at the same place that I am presenting it—towards the end—then it would appear as though they were making a desperate hail-Mary pass.

Fig. 1

Fig. 2

Fig. 3

At the end of the day, we must consider the ultimate point of all this, for which I will zero-in on the word "protected." The title of the Stein et al. study begins with the phrase, "Closed but Not Protected," and their overall conclusion, to paraphrase, is that the Amish/Mennonite population was similarly unprotected as the rest of the United States, with a sort of unspoken subtext indicating that they were actually LESS protected (per their own interpretation of the data, but they probably don't want you to scrutinize it too much). It is a noble pursuit indeed to want to protect people—to protect life—and comparing how different groups and their respective lifestyles fared against the same harmful forces is a great start to that pursuit. We begin broadly by learning which lifestyles are generally healthier, which in this case may translate into better protection against viral outbreaks and other bodily hazards. Later, we can narrow down which particular lifestyle choices--be it vaccine uptake, use of pharmaceuticals, utilization of other modern medical practices, diet, even exposure to electromagnetic radiation—have the greatest bearing our health, either positively or negatively. So two extremely broad potentials exist at the outset of this noble pursuit: to better protect (i.e. make healthier) the Amish/Mennonite population, or to better protect the general population of the United States and beyond, which is achieved by more closely aligning one's lifestyle to the healthier lifestyle of the other. Yet understand that the potential to help the vastly higher population of the United States is unfairly precluded by the evidence-free assertions about "excess deaths" made by the authors of this study. It's irresponsible. Their study actively hinders the noble pursuit heretofore described. Even if we grant that the viable potential is to help the Amish/Mennonite population, the mealy-mouthed way that the authors went about it would have almost been a disservice. Altogether, their efforts ARE a disservice, and I am happy to have set the record straight. I only regret that the taxpayers cannot be refunded for this atrocity.

1. See here.

2. Hereby referred to as the Stein et al. paper. See here or here for full text.

3. This is a more-than-reasonable estimate considering the number of scribes, their respective community populations on average, and comments from The Budget’s publisher, Milo Miller, who has stated that his paper circulates in "nearly all known Amish settlements" and has "pretty good saturation."

4. 2015-2020 average. The World Bank; Death rate, crude (per 1,000 people) - United States; See here.

5. See here. Estimate derived from 2015-2020 yearly totals.

6. “Measuring and monitoring excess mortality across different countries requires, first and foremost, a comprehensive and regularly-updated dataset on all-cause mortality.” See Tracking excess mortality across countries during the COVID-19 pandemic with the World Mortality Dataset

7. Consider, for example, Human Mortality Database’s methods, which can be viewed or downloaded from their website (from the main menu). Their full protocol mentions “population” or “populations” 175 times in 78 pages.

8. Per our own personal communication.

9. See here.

10. Click here for pdf link.

11. See here.

12. Our now infamous authors actually began this process, and while they made some progress, they really only scratched the surface in terms of what's really important to research here, and again, they did not even apply their own findings to help interpret the data from this study in question. See the links to their additional studies at the bottom of this page.