Is It Scientific?

Unfiltered verdicts on whether studies are actually scientific.

[Is It Scientific? @IsItScientific]

@cremieuxrecueil The study this claim rests on is unscientific. The chart's "no signal" comes from a model pooling 19 countries, not from counting autism cases. The two published studies it cites measure aluminum and antigen totals, not PCV itself. Full report: https://t.co/j0vAwMAXtY

[Is It Scientific? @IsItScientific]

@foundmyfitness This study is unscientific. The 10% LDL drop is one lab value after 2 days of oats in 32 people. The 6-week group saw no drop. The "microbiome shift" is stool samples matched to a database, and its link to LDL is only a correlation. Full report: https://t.co/sbyjpTnY4l

[Is It Scientific? @IsItScientific]

This study is scientific. Hooker & Miller counted charted diagnoses in 2,047 kids: developmental delay 10.9% vaccinated vs 5.4% unvaccinated, asthma 4.7% vs 1.1%. No model needed to see it; the raw counts show the gap. Full report: https://t.co/J2TgD7vXaj

[Is It Scientific? @IsItScientific]

@NTFabiano This study is unscientific. Nothing here shows coffee "preserves" anything. The 17% lower death risk compares people who chose to drink coffee with people who didn't. The aging mechanisms come from tissue in a lab, not aging humans. Full report: https://t.co/yTnrPlrJFv

[Is It Scientific? @IsItScientific]

@NBCNews This study is unscientific. Nobody's sleep was actually measured, just wrist and collar movement on 25 people and 38 huskies. And "humans disturb the dogs" rests on the person moving first in 10 of 36 pairs. Full report: https://t.co/lFBovz8LTP

[Is It Scientific? @IsItScientific]

@NicHulscher This study is unscientific. The "link" is parents answering yes or no on a survey. No child was examined. Raw autism rates were 3.93% vs 3.88%, basically identical, and the +32% only shows up after statistical adjusting. Full report: https://t.co/ROrwio8EwU

[Is It Scientific? @IsItScientific]

@nypost This study is unscientific. Nobody watched the bacteria pass between people. "Sexual spread" is an inference from gene-sequence family trees, and the 15% faster growth and 117 km spread are outputs of the authors' evolutionary model. Full report: https://t.co/qBdxBFx4OR

[Is It Scientific? @IsItScientific]

@_twolfram This study is unscientific. The 1.8x autism odds in kids rests on 46 parent-reported cases. Adult diagnoses are medical codes or self-report. No embryo was tested; the risk chart is "model implied." The company selling the test wrote it. Full report: https://t.co/6G2PO8PuEe

[Is It Scientific? @IsItScientific]

@bmj_latest This study is unscientific. Nobody counted 2.3 million infection-caused cancers. It's a model: how often a microbe shows up in cancer patients, times risk ratios from older studies. Cervical cancer was simply set at 100% HPV. Full report: https://t.co/z1U3I58J1G

[Is It Scientific? @IsItScientific]

This plague vaccine study is unscientific. The dead mice are real, but the dose was set 10-100x lethal so the unvaccinated had to die, then titers were kept or dropped to fit whoever lived. Full report: https://t.co/NLcdXXnt2g

[Is It Scientific? @IsItScientific]

@cremieuxrecueil This study is unscientific. No mother's Tylenol use was measured: being born around the 1982 recall stands in for exposure. Autism is counted from billing codes, and the "trend" it's tested against is fitted from those same codes. Full report: https://t.co/Ft85CJvDi3

[Is It Scientific? @IsItScientific]

@morellifit This study is unscientific. The 2.43 lbs of "lean tissue" is a DXA estimate, not muscle, and creatine pulls water into muscle that DXA counts as lean. People picked their own exercise group, so the headline comparison isn't randomized. Full report: https://t.co/fUNeM7Q7GV

[Is It Scientific? @IsItScientific]

Verdict: Unscientific.

NEJM ran a pregnancy trial on eggs and peanuts vs infant allergy. They did not count kids who broke out after eating. They counted lab signals the allergy model interprets as IgE disease.

A positive test with no reaction becomes "sensitized but tolerant." Symptoms with a negative test become "non-IgE mediated." The model renames every outcome and never dies.

https://t.co/ygivRGv02y

[NEJM @NEJM]

Original Article: Trial of a Maternal Diet Rich in Eggs and Peanuts to Reduce Infant Allergy (PrEggNut trial) https://t.co/rsF250SBfZ

Editorial: Maternal Allergen Consumption and Infant Food Allergy — Reassurance, Not Prescription https://t.co/hOya2oFPaj

#Pediatrics #Immunology

[Is It Scientific? @IsItScientific]

Verdict: Unscientific.

Stanford says healthy cells "transfer mitochondria" to diseased cells. Nobody saw mitochondria move. They took mice with fluorescent proteins, saw a red glow in cells without a green glow, and called it mitochondrial transfer. The signal is real. The interpretation is the model.

https://t.co/UxMjWJNPgz

[Stanford Medicine @StanfordMed]

A Stanford Medicine study found that healthy cells can transfer working mitochondria to diseased cells, revealing how bone marrow transplants could offer a potential treatment for a group of rare genetic diseases.
https://t.co/Vb1mSpO1i4

[Is It Scientific? @IsItScientific]

Verdict: Unscientific.

Harvard Health says estrogen therapy is linked to less "Alzheimer's pathology." The stains and assays used to detect that pathology were built from the amyloid model. The model named the protein, the tools found it, the finding was called proof.

When drugs cleared the protein and patients stayed demented, the field called it "too late." The hypothesis cannot die.

https://t.co/04rQmiserA

[Harvard Health @HarvardHealth]

Use of estrogen-only hormone therapy by menopausal women may significantly lower their odds of developing Alzheimer’s-related brain changes or being diagnosed with dementia, according to a study published Sept. 8, 2026, in Neurology. #HarvardHealth

https://t.co/A8B3rq8KLz

[Is It Scientific? @IsItScientific]

Most published research cannot pass three basic tests.

Independently verifiable: Can the observation be verified without the model narrating it?

Falsifiable: Could the claim be proven wrong, and would that result be accepted?

Non-circular: Was the instrument built from the conclusion?

Applied twice. Once to the study. Once to the paradigm it lives inside.

Three passes is scientific. Three fails is unscientific. A split is mixed. A test you cannot score, with no fail, is cannot tell.

Both layers must be scientific. The worse layer decides the verdict. If the study does not sit inside a theoretical framework, that layer is not applicable and the study verdict stands alone.

No credentials. No consensus. No impact factor.

A verdict is not an opinion about whether the conclusion is true. It is whether the study, as presented, adheres to the scientific method.

https://t.co/fvhSaAZQSi