The Numbers Nobody Is Running: Examining the Statistical Relationship Between COVID Mortality and Influenza Vaccination Rates
For four years, the American public was told to trust the science. Trust the institutions. Trust the models. And above all, do not ask uncomfortable questions about the data underlying the decisions that shuttered businesses, closed schools, and mandated injections for millions of citizens. Now that the dust has begun to settle on the COVID-19 era, a quieter but increasingly urgent conversation is taking place among independent researchers, biostatisticians, and public health skeptics: what does the correlation coefficient actually tell us when we map influenza vaccination rates against COVID-19 death counts across counties, states, and nations?
The mainstream press has shown little appetite for this question. That, in itself, should tell you something.
What a Correlation Coefficient Actually Measures
Before diving into the controversy, a brief statistical primer is warranted. A correlation coefficient — most commonly expressed as Pearson's r — is a numerical measure ranging from -1 to +1 that describes the strength and direction of a linear relationship between two variables. A coefficient near +1 indicates that as one variable increases, the other tends to increase as well. A coefficient near -1 suggests the opposite. A coefficient near zero implies little to no linear relationship.
The critical caveat — one that statisticians repeat endlessly but that public health communicators conveniently forget when it suits their purposes — is that correlation does not imply causation. Two variables can move in tandem for any number of reasons: shared underlying causes, confounding variables, data collection artifacts, or simple coincidence across a limited sample.
With that foundational principle in mind, the question becomes: when researchers calculate the correlation between flu shot uptake rates and COVID-19 deaths across comparable geographic units, what do they find — and why has that finding received so little scrutiny from the institutions we fund to scrutinize it?
The Research That Raised Eyebrows
Several peer-reviewed and preprint studies published between 2020 and 2023 identified a statistically notable positive correlation between higher influenza vaccination rates and higher COVID-19 case or mortality counts in certain datasets. One frequently cited analysis — drawing on data from 39 countries — reported a Pearson correlation coefficient suggesting that nations with higher flu vaccination coverage among elderly populations did not demonstrate the protective COVID-19 outcomes that public health models might have predicted.
These findings were met not with rigorous scientific engagement, but with reflexive dismissal. Papers were flagged on social media platforms. Researchers who raised the question found their work buried in search results. Journalists who might have asked follow-up questions appeared to receive the memo: this line of inquiry was not to be pursued.
It is worth being precise about what responsible analysts have argued here. The more credible voices in this conversation are not claiming that flu shots cause COVID deaths. They are asking whether the correlation data, once properly adjusted for confounders — age distribution, comorbidity prevalence, healthcare access, testing intensity — tells a different story than the one the public health establishment has authorized. That is a legitimate scientific question. The refusal to engage with it seriously is not.
The Confounding Variable Problem
Any honest analysis of this correlation must grapple seriously with confounding variables — and this is precisely where the mainstream dismissal falls short. Critics of the correlation studies often point out, correctly, that higher flu vaccination rates tend to cluster in older populations, who also faced dramatically elevated COVID-19 mortality risk. In other words, the positive correlation may simply reflect the fact that older, more medically vulnerable people were both more likely to receive annual flu shots and more likely to die from COVID-19.
This is a reasonable objection. But it is not a conversation-ending one. The proper response to a confounding variable concern is to control for it — to run the analysis again with age-stratified data, to examine the correlation within comparable age cohorts, to build a multivariate model that isolates the variables of interest. What the public health establishment and its media allies have largely declined to do is fund or publicize that more rigorous analysis.
The question is not whether confounders exist. They always do. The question is whether the institutions charged with protecting public health are genuinely interested in understanding the data — or whether they are interested in protecting a predetermined conclusion.
A Pattern of Selective Skepticism
What makes this episode particularly instructive is the asymmetry of scrutiny applied to different lines of research. Studies suggesting that masks, lockdowns, or vaccines reduced COVID mortality were amplified, cited by government officials, and treated as settled science with remarkable speed. Studies raising inconvenient questions about the same interventions were subjected to a level of methodological criticism rarely applied to the favored findings.
This is not how science is supposed to work. The correlation coefficient does not care about the political preferences of the researcher running the calculation. It does not consult the editorial standards of the New York Times before returning a value. It simply describes the relationship between two variables in a given dataset.
Conservative and independent media outlets have been among the few spaces willing to host this conversation without immediately reaching for the "misinformation" label. That should not be the case. A free and functioning press — regardless of ideological orientation — ought to be capable of asking: what does the data show, who collected it, who funded the analysis, and what are the methodological limitations?
What Accountability Would Look Like
The American public deserves a genuine accounting of pandemic-era data — not a curated summary approved by the same agencies whose decisions are under examination. That accounting should include:
- Full public release of county-level and state-level datasets linking vaccination records, flu shot uptake rates, COVID case counts, and mortality figures, with sufficient granularity to allow independent analysis.
- Pre-registered replication studies in which researchers with no institutional stake in the outcome run correlation and regression analyses on the available data.
- Congressional scrutiny of the editorial and funding decisions that shaped which pandemic research received amplification and which was quietly sidelined.
- Honest engagement from public health officials with the methodological objections raised by credentialed critics — not dismissal, not appeals to authority, but actual engagement with the numbers.
None of this is radical. It is the basic apparatus of scientific accountability that the public was promised and never received.
The Deeper Issue
The controversy over the correlation between flu vaccination rates and COVID mortality outcomes is, at its core, a controversy about institutional trust — and whether that trust has been earned. Americans were told repeatedly that the science was settled, that the data was clear, and that questioning the consensus was tantamount to endangering lives.
But science that cannot withstand questions is not science. It is orthodoxy. And orthodoxy, as this publication has documented across a range of policy domains, tends to serve the interests of those who enforce it.
The correlation coefficient is a humble tool. It does not lie, and it does not spin. What happens to the findings it generates, however — which ones get published, which ones get funded, which ones get reported, and which ones get buried — that is a deeply human and deeply political process. Americans deserve to understand both the math and the machinery behind it.
