Smaller Than Aspirin: What Effect Size Reveals About Intuition

In the winter of 1988, a major heart-disease trial did something studies almost never do: it stopped early. The evidence that a low dose of aspirin prevented heart attacks was judged so clear that the researchers decided it was no longer ethical to keep giving half the participants a placebo. Aspirin worked. The result became a textbook landmark of preventive medicine.

Here is the part almost nobody mentions: by the standard yardstick scientists use to measure how strong an effect is, that landmark result was tiny. And some of the most controversial experiments in science — laboratory studies of intuition, of "knowing" things the senses can't reach — have produced effects several times larger, only to be set aside.

This isn't an argument that aspirin doesn't work, or that intuition is settled science. It's a comparison the researcher Dean Radin has pressed for years, and it's really about consistency: judging like evidence by like standards.

The result medicine trusts

The Physicians' Health Study followed 22,071 male U.S. physicians, randomly assigned to take either aspirin or a placebo every other day. The outcome was decisive enough to halt the trial: 104 heart attacks in the aspirin group versus 189 in the placebo group — about a 44% reduction in risk, with a p-value below 0.00001. The odds that this was a fluke were vanishingly small.

So far, so familiar. Now the twist. "Statistically significant" answers one question — are we sure there's an effect? It says nothing about how big the effect is. For that, scientists use effect size. Convert this trial's outcome into a correlation — the kind of "binomial effect size display" popularized by Rosenthal and Rubin — and the aspirin effect comes out around r ≈ 0.03. (You can check it yourself: the 104-versus-189 split across the two arms works out to a correlation of about 0.03; the figure isn't reported in the original paper, it's derived from its numbers.) Squared, that's roughly 0.001 — about one tenth of one percent of the variance in who had a heart attack and who didn't.

That is not a knock on aspirin. A tiny effect spread across millions of people prevents a great many heart attacks — which is exactly why the trial was stopped. The point is the number itself: 0.03 is the size of an effect modern medicine treats as a triumph.

Now measure intuition

Hold that 0.03 in mind, and look at the experiments mainstream science is most reluctant to take seriously — the laboratory studies of what parapsychologists clinically call non-local information transfer: getting accurate information without a known sensory channel.

In the ganzfeld telepathy experiments, a "receiver" in mild sensory isolation tries to describe an image a distant "sender" is viewing, then picks it from four options. Chance is 25%. Across the Bem & Honorton autoganzfeld sessions, receivers scored about 32% — an effect size of roughly 0.16, several times the aspirin figure. (Critics rightly note that the edge came mostly from dynamic video targets; static images sat closer to chance.) A larger 2010 meta-analysis by Storm and colleagues, pooling studies through 2008, put the broader ganzfeld effect a little lower — around 0.14 — still well above the aspirin benchmark.

The most scrutinized body of work is the U.S. government's own. For two decades the CIA and Defense Intelligence Agency funded remote viewing research, the program later declassified as STAR GATE. In 1995 the CIA commissioned statistician Jessica Utts to evaluate it, and her report was blunt:

Using the standards applied to any other area of science, it is concluded that psychic functioning has been well established. The statistical results of the studies examined are far beyond what is expected by chance.

Utts put the magnitude of the remote-viewing effect at "between what social scientists call a small and a medium effect" — on the order of 0.2 — and noted it was consistent across independent labs.

Same standards, opposite verdicts

Line the numbers up:

These correlations come from different measurement traditions, so the table is about order of magnitude, not decimal-point equivalence. Even read that loosely, the intuition effects are several times larger — up to roughly an order of magnitude — than the aspirin effect we hold up as decisive. Judged by the same statistics, they should be at least as interesting. Yet one is in every cardiology textbook and the others are treated as fringe.

The honest part

If the comparison stopped there, it would be propaganda. It doesn't.

The intuition research has real problems, and the first is replication. In 1999, Milton & Wiseman pooled 30 ganzfeld studies from seven labs and found an effect size of about 0.013 — essentially zero, and not significant. Proponents and skeptics have been trading meta-analyses ever since.

The second problem is what the comparison leaves out. Effect size is one axis, not the only one. Aspirin has a known biological mechanism; "non-local information transfer" has none — no accepted account of how it could work. That matters. An extraordinary claim with no mechanism starts from a much lower prior than a drug whose chemistry we understand, so identical effect sizes don't automatically earn identical confidence. This is exactly the objection the skeptic Ray Hyman, who evaluated STAR GATE alongside Utts, pressed — and, unlike Utts, his bottom line was that the case was not established without independent replication and a mechanism. (The same 1995 review also judged that the program had produced no demonstrated operational intelligence value, whatever the lab statistics showed.) The units differ, too: one large, pre-registered drug trial is not the same kind of object as a meta-analysis of many small, uneven studies, even when the headline r matches.

And psychology has since lived through a replication crisis — catalyzed in part by a 2011 parapsychology paper — that taught everyone to distrust small effects mined from flexible analyses and unpublished null results (the "file drawer"). That problem bites the psi literature especially hard, because amateur null results rarely get published at all.

So this is not a claim that intuition is proven. It's a narrower, more durable claim: the reflexive dismissal isn't consistent with the evidence. Effects this size, replicated imperfectly, are normal across the soft sciences and much of medicine. We investigate them. We don't laugh them out of the room because of what they'd imply.

Why this matters to us

Intuito exists because that question deserves better than a shrug. Notice, though, what the studies above don't tell you: they're group effects, averaged across many people in a lab. They say nothing about your own number. So we built a place to find out — run your own trials, log your hits and misses, and watch your personal hit rate take shape over time. The data is yours, and it answers to no one's prior, believer or skeptic.

The aspirin study is a reminder that a small, real effect can still change the world. The only way to learn whether yours is real is to start measuring.

Start measuring your intuition →

Sources

1. Steering Committee of the Physicians' Health Study Research Group. "Preliminary Report: Findings from the Aspirin Component of the Ongoing Physicians' Health Study." New England Journal of Medicine 318:262–264 (1988).
2. Rosenthal, R. & Rubin, D. B. "A Simple, General Purpose Display of Magnitude of Experimental Effect." Journal of Educational Psychology 74:166–169 (1982); and Funder, D. C. & Ozer, D. J. "Evaluating Effect Size in Psychological Research." Advances in Methods and Practices in Psychological Science 2(2):156–168 (2019) — source of the r ≈ 0.03 aspirin figure (a binomial-effect-size-display conversion, not a number reported in the trial itself).
3. Radin, D. The Conscious Universe: The Scientific Truth of Psychic Phenomena. HarperEdge (1997) — origin of the aspirin / effect-size comparison.
4. Bem, D. J. & Honorton, C. "Does psi exist? Replicable evidence for an anomalous process of information transfer." Psychological Bulletin 115(1):4–18 (1994).
5. Utts, J. "An Assessment of the Evidence for Psychic Functioning." Report to the American Institutes for Research / CIA (1995); with Hyman, R. "Evaluation of Program on Anomalous Mental Phenomena" (1995). (Commissioned government reports, not peer-reviewed.)
6. Storm, L., Tressoldi, P. E. & Di Risio, L. "Meta-Analysis of Free-Response Studies, 1992–2008." Psychological Bulletin 136(4):471–485 (2010).
7. Milton, J. & Wiseman, R. "Does psi exist? Lack of replication of an anomalous process of information transfer." Psychological Bulletin 125(4):387–391 (1999).
8. Bem, D. J. "Feeling the Future: Experimental Evidence for Anomalous Retroactive Influences on Cognition and Affect." Journal of Personality and Social Psychology 100(3):407–425 (2011).