Why the latest Geier & Geier paper is not evidence that mercury in vaccines causes autism

Let’s set the scene: someone sends you a link (usually with 14 exclamation points), saying the “latest Geier & Geier paper proves mercury in vaccines causes autism.” It sounds dramatic. It sounds scientific. It often comes with a screenshot of a chart, a few bolded odds ratios, and a confident caption like “How is this still debated?”

Here’s the less clickybut much more importanttruth: a paper is not the same thing as evidence of causation, and this particular style of paper (including the Geier & Geier “mercury/thimerosal” publications that keep resurfacing) relies on methods that cannot answer the question it claims to answer.

This article breaks down why. We’ll cover what “mercury in vaccines” actually means, what the newest Geier & Geier-style claims typically rely on, what makes those methods weak for determining cause-and-effect, and what the strongest research says instead. Spoiler: it’s not a conspiracy; it’s just… how data works.

First: what people mean by “mercury in vaccines”

When people say “mercury,” they’re usually referring to thimerosal (also spelled thiomersal), a preservative that contains mercury and was used for decades in some multi-dose vaccine vials to prevent contamination. Thimerosal breaks down into ethylmercury, which is not the same as methylmercurythe type associated with high exposures from certain fish and industrial pollution.

That distinction matters because different forms of mercury behave differently in the body. Ethylmercury is cleared from blood more quickly than methylmercury. In other words: even if two compounds share a chemistry “last name,” their biology can be very different.

Also important context: in the United States, thimerosal was removed from (or reduced to trace amounts in) nearly all routine childhood vaccines starting in the early 2000s. Today, it’s mostly associated with some multi-dose flu vaccine vials, not the standard childhood vaccine lineup.

Second: what counts as “evidence” that something causes autism?

Autism spectrum disorder (ASD) is a neurodevelopmental condition with early-life origins. Researchers investigate possible causes using multiple lines of evidence, including genetics, prenatal factors, and early development. When it comes to exposures (like medications, infections, or environmental substances), establishing causation usually requires:

• Accurate exposure measurement (who got what, when, and how much)
• Reliable outcome measurement (clear diagnostic criteria, consistent follow-up)
• A design that can test cause-and-effect (not just correlation)
• Control for confounding factors (differences that could explain results)
• Replication (other researchers, other datasets, similar findings)

One study almost never “proves” a complex claim by itselfespecially when the claim contradicts decades of research across large populations. Science is less “one heroic paper defeats the villain” and more “many independent studies keep checking the same question from different angles.”

So what is the “latest Geier & Geier” paper actually doing?

The most recent widely cited Geier & Geier–authored work arguing a thimerosal-autism connection uses a familiar recipe: take a large database, split reports or records into “exposed” and “unexposed,” calculate odds ratios, and interpret those numbers as evidence that thimerosal is linked to autism or related diagnoses.

A key example of this approach is their work that uses passive reporting datamost notably the Vaccine Adverse Event Reporting System (VAERS). In this style of analysis, the paper often compares adverse event reports associated with one vaccine formulation versus another (for instance, thimerosal-containing versus thimerosal-free formulations) and then looks for differences in how often autism-related terms appear in those reports.

On the surface, that can look compelling, because odds ratios feel like they’re delivering a verdict. But here’s the problem: the database and the method are not designed to answer the causation question.

The biggest issue: VAERS data can’t show causation (and it’s not meant to)

VAERS is a national vaccine safety monitoring system where anyone can submit a reportclinicians, patients, family members, and manufacturers. That wide-open door is a feature, not a bug: it helps public health teams detect signals that might warrant further investigation.

But VAERS comes with a giant, flashing warning label: reports alone can’t tell you whether a vaccine caused an event. Why? Because VAERS is a passive system. It doesn’t verify every report, it doesn’t include perfect denominators (how many total people received each product in the relevant way), and it can be influenced by publicity, litigation, and awareness trends.

Think of VAERS like a smoke detector. It can tell you, “Hey, something might be burning.” It cannot tell you whether it’s a kitchen mishap, burnt toast, fog from a shower, or an actual fire. If you treat a smoke detector alarm as “proof the building is on fire,” you’ll panic every time someone makes popcorn.

Specific reasons VAERS-based “odds ratio” papers mislead

• Reporting bias: if a topic is in the news (or in court), reporting patterns change. That can inflate or distort apparent associations.
• No controlled comparison group: “controls” inside a reporting database are not the same as controls in a properly designed study.
• Uncertain timing and diagnostic accuracy: autism is diagnosed after developmental patterns emerge. VAERS reports aren’t built to validate the clinical timeline of diagnosis.
• Confounding factors: differences in healthcare access, diagnostic practices, and follow-up can create patterns that look like causation when they’re not.
• Denominator problems: without clean “how many people got X vaccine under Y circumstances,” you can’t reliably estimate risk.

Bottom line: a VAERS-based analysis can generate a hypothesis. It cannot confirm that thimerosal causes autism.

Another issue: “association” isn’t automatically “causal,” especially with messy inputs

Even in the best circumstances, association is not causation. But in weak circumstanceslike passive reporting systems or cherry-picked subsetsassociation can be the statistical equivalent of a magic trick: attention goes to the flashy result, while the methodological trapdoor stays off-camera.

This is why high-quality vaccine safety research relies heavily on active surveillance and linked health record databases where vaccination status, timing, and diagnoses can be tracked consistently across large populations.

What stronger research finds instead

When researchers use better toolslarge cohort studies, linked databases, and carefully controlled designsthe thimerosal-autism link does not hold up.

Major scientific reviews have concluded that the overall epidemiologic evidence favors rejecting a causal relationship between thimerosal-containing vaccines and autism. Public health agencies have also noted a key real-world pattern: after thimerosal was removed from nearly all childhood vaccines, autism diagnosis rates continued to rise. If thimerosal were a major driver, the trend should have moved in the opposite direction.

Additionally, comparisons from places that reduced or removed thimerosal earlier have not shown the drop in autism diagnoses you’d expect if the preservative were causative. That kind of “natural experiment” isn’t perfect, but it’s a reality check against claims that one ingredient explains a complex diagnosis.

Ethylmercury vs. methylmercury: same periodic table, different biology

One reason “mercury” sounds scary is that mercury exposure can be toxic at high dosesparticularly methylmercury. But thimerosal breaks down into ethylmercury, and the body handles ethylmercury differently.

Pharmacokinetic studies (including work in infant primates and observations in infants) have found that mercury from thimerosal is cleared from blood relatively quickly compared with methylmercury. That matters when people try to import risk assumptions from one compound to another.

Could anyone be allergic or sensitive to a vaccine component? Rarely, yesjust like with foods, medications, or bee stings. But a rare adverse reaction is not the same as “this causes autism across the population.” Those are different claims requiring different evidence.

Why the autism trend didn’t “follow” thimerosal removal

A common myth storyline goes like this: “Autism rates went up; thimerosal existed; therefore thimerosal caused it.” That story has two big problems:

• Timing and diagnosis changed: diagnostic criteria broadened, awareness increased, and services expandedso more people were identified who would previously have been missed or given different labels.
• Complex conditions don’t have single-cause switches: ASD has strong genetic contributions and multifactorial influences. Looking for one ingredient as a master key is emotionally satisfying, but scientifically shaky.

If you’ve ever watched a neighborhood change because a new coffee shop opened, you already understand this: trends can shift for many reasons, and blaming the coffee beans is… a choice.

How to spot “sounds scientific” papers that don’t actually answer the question

Here’s a quick checklist you can use the next time a “this proves vaccines cause autism” paper makes the rounds:

1) What data source is used?

If it’s passive reporting (like VAERS), treat it as hypothesis-generatingnot proof.

2) Can the study measure exposure accurately?

“Mercury exposure” is not a vibe. Good studies measure who received what, when, and in what formulation.

3) Are outcomes verified?

Reliable autism research uses validated diagnoses, consistent criteria, and appropriate follow-upoften across years.

4) Are confounders addressed?

Health care access, socioeconomic factors, diagnostic practices, and timing can all distort results if not controlled.

5) Is there replication?

If a result is real and large, independent groups should see it tooespecially in well-designed studies.

6) Are conflicts of interest disclosed and handled transparently?

Disclosure doesn’t automatically invalidate a paper, but it raises the bar for methodological rigor and independent replication.

What to say when someone shares the paper (without starting a family group chat war)

You don’t need to dunk on anyone to be accurate. Try:

• Start with values: “I get why you’re worried. Everyone wants kids to be safe.”
• Ask about the data source: “Is this VAERS-based? VAERS is for signals, not proof.”
• Offer stronger evidence: “Large controlled studies and major reviews don’t find a link.”
• Keep it human: “If you want, we can look at what pediatric groups and safety systems say.”

Not every conversation will end with agreement. But you can often lower the temperature while still protecting the facts.

Conclusion: why this paper isn’t the “gotcha” people think it is

The Geier & Geier-style thimerosal papers may look persuasive to non-specialists because they use scientific formatting and statistics. But the core issue is simple: methods that can’t establish causation can’t be used as evidence that thimerosal causes autism.

When the question is tested using stronger study designs across large populations, the proposed link doesn’t appear. That’s why major scientific reviews and vaccine safety experts conclude the evidence does not support thimerosal in vaccines as a cause of autism.

If you’re sorting through scary claims online, anchor yourself to the kind of evidence that can actually answer the question: well-designed studies, consistent replication, and independent expert review. The internet rewards certainty. Science rewards accuracy. Choose the one that helps real people.

Experiences people commonly have around this topic (and how to navigate them)

Even though the science is strong, the experience of this debate can still feel confusingbecause it rarely shows up as a calm journal club discussion. It shows up in everyday life, where emotions, identity, and genuine worry are doing most of the talking. Here are some real-world scenarios people often run into, plus practical ways to handle them without losing your mind (or your favorite aunt).

1) The “I’m not anti-vax, I’m just asking questions” message

This usually arrives as a late-night text with a link and a note like, “Thoughts??” The sender may not be trying to start a fightthey may be trying to reduce anxiety by finding a clear explanation. A helpful response is to validate the concern (“It’s fair to want good safety data”) and then gently redirect to methodology (“This uses a reporting system that can’t prove cause-and-effect”). People are often relieved when you give them a reason something sounds compelling but isn’t actually definitive.

2) The pediatric visit where someone brings printouts

Clinics see this. A parent may arrive with highlighted pages, odds ratios circled, and a very reasonable-sounding question: “If there’s any chance, why risk it?” The useful move here is to separate hazard from risk. Yes, high mercury exposure can be hazardous. No, that doesn’t mean a low-dose preservative used in limited vaccine contexts causes autism. Clinicians also often point out that the diseases vaccines prevent are not hypothetical; they are documented risks with real consequences.

3) The social media post that treats a single paper like a final verdict

Algorithms love “one weird trick” content, and nothing gets clicks like a claim that experts have been hiding something. If you engage publicly, it helps to avoid debating every detail and instead ask one clarifying question: “What type of data is this based on?” When the answer is a passive reporting database, you can calmly note that these systems are for signal detection, not proof. You don’t have to win the comments section; you just need to add a reliable breadcrumb for silent readers scrolling past.

4) The family story where timing becomes the whole argument

Many people know a child who was diagnosed with autism after the age when vaccines are given, and the timing can feel emotionally convincing. It’s a deeply human pattern: we look for a “before” and “after” because it helps our brains create order. A compassionate approach is to acknowledge that the timing is real while explaining that autism’s developmental roots begin earlier, and diagnosis often happens later as differences become more visible. This isn’t dismissing anyone’s experience; it’s explaining how development and diagnosis timelines work.

5) The “why did they remove thimerosal if it’s safe?” question

This is one of the most common sticking points. People assume removal equals admission of guilt. In reality, policy decisions can reflect public concern, risk perception, and practical availabilityespecially when a preservative can be removed without compromising supply in some settings. The key nuance: removal can be a precautionary or confidence-building step, not proof of harm. That distinction helps many people reframe the issue without feeling like they’re being talked down to.

In most of these experiences, facts matterbut so does tone. If you can combine empathy (“I get why this feels scary”) with clarity (“this data source can’t show causation”), you’ll do more good than a hundred sarcastic one-liners. And yes, that includes resisting the urge to reply with “bestie, that’s not how any of this works,” even when it absolutely is.