How to Read a Scientific Claim: Six Questions
Reading level: Introductory · Foundation: Science & Evidence · Evidence: Established
Every day we meet claims dressed as science: “studies show”, “experts agree”, “research proves”. Some are solid; many are not. You do not need a laboratory to tell the difference. You need a few good questions, asked calmly and in the right order.
Why this matters even in respected science
In 2015 a large team of researchers, the Open Science Collaboration, tried to repeat 100 published psychology experiments. Only about 36% of the repeated studies produced a statistically significant result again, and the effects they found were on average about half as large as the originals. This did not show that psychology is worthless. It showed that a single study, however interesting, is a starting point rather than a conclusion. (Established)
The physician and researcher John Ioannidis had already argued in 2005 that small studies, small effects, flexible analyses and strong incentives to publish exciting results make false positives more likely. Science corrects itself, but slowly, and the corrections rarely make headlines.
Six questions to ask of any claim
- What exactly is claimed? “Linked to”, “associated with” and “causes” mean very different things.
- Who was studied? Mice, thirty students or a nationally representative sample of thousands?
- How large is the effect? A “significant” result can still be tiny in everyday terms.
- Has it been repeated? One study is a clue; several independent ones are evidence; a careful review of many is stronger still.
- Who benefits from the claim? Funding and sales are not proof of bias, but they are reasons to look closer.
- What would change my mind? If nothing could, the claim has become a belief rather than a finding.
How we label evidence at PMA Science University
To make these distinctions visible, every claim in our courses and essays carries one of five labels:
- Established: supported by mature, published evidence.
- Evidence-informed: built from established concepts, but not itself validated.
- Emerging: promising but incomplete; taught with uncertainty.
- Interpretation: historical or philosophical reasoning, not scientific proof.
- PMA Science Original: our own frameworks and exercises, identified as such.
These labels are a promise to readers: we will tell you not only what we think, but how sure anyone can reasonably be.
The Stoic connection
Epictetus taught his students to pause before accepting any impression, and in effect to say to it: wait a moment, let me see what you are and put you to the test. Checking a claim before believing it is the same discipline, applied to information rather than to emotion. (Interpretation)
Try it this week. Take one health or psychology headline you see. Find the original study it refers to and answer just three of the six questions above: what was claimed, who was studied and how large the effect was. Notice how often the headline says more than the study does.
Read more about the five foundations and our evidence standard on the Foundations page.
Sources and further reading
- Open Science Collaboration (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716.
- Ioannidis, J. P. A. (2005). Why most published research findings are false. PLoS Medicine, 2(8), e124.
- Epictetus, Discourses 2.18; see “Epictetus” in the Stanford Encyclopedia of Philosophy.
- Spiegelhalter, D. (2019). The Art of Statistics: Learning from Data. Pelican.




Leave a Reply
Want to join the discussion?Feel free to contribute!