The Control Group Has Been Controlled

TL;DR

• Quantum mechanics really does pressure naïve realism, but that is not the same as evidence that human thought steers quantum outcomes Chalmers & McQueen (2021), Okon & Sebastián (2018).
• Double-slit “observer consciousness” findings looked suggestive, then weakened under reanalysis and sham controls; the cleanest read today is method fragility, not mind-over-matter Tremblay (2019), Walleczek & von Stillfried (2019).
• RNG/QRNG studies produced a tiny meta-analytic signal, but strong replications and large Bayesian tests leaned null or artifact Bösch et al. (2006), Jahn et al. (2000), Maier et al. (2018).
• Direct consciousness-causes-collapse tests exist, but the classic Hall line was null, Bierman’s EEG follow-up was explicitly inconclusive, and recent Lucido studies are too niche and under-stressed to carry much metaphysical weight Bierman (2003), de Barros & Oas (2017).
• Serious idealist-friendly physicists usually argue about interpretation and first-person structure, not about proven lab psychokinesis Müller (2021), Catani (2023), Adlam (2023).

“The only question is whether it’s a signal of psi, or a signal of poor experimental technique.”
— Scott Alexander, “The Control Group Is Out Of Control” (2014)

What exactly is being claimed when people say “thinking affects quantum states”?#

This phrase usually smuggles three different theses through the same customs gate:

1. Quantum theory is hard to interpret without consciousness.
2. Consciousness plays a role in collapse.
3. Human intention can measurably bias quantum outcomes in the lab.

These are not equivalent. The first is interpretive, the second is a dynamical hypothesis, and the third is a straight empirical claim. The first two can remain live even if the third faceplants. Formal work by David Chalmers, Kelvin McQueen, Adrian Kent, Elias Okon, and Miguel Ángel Sebastián shows that consciousness-related collapse ideas are a real research program, not just incense with equations; but those same authors also frame the issue as speculative, constrained, and testable rather than established fact Chalmers & McQueen (2021), Kent (2020/2021), Okon & Sebastián (2018).¹

(For broader context on how consciousness theories have developed, see the historical evolution of consciousness and Sam Harris on consciousness.)

It is also easy to confuse observer-dependence with mind-over-matter. Extended Wigner’s-friend experiments do put pressure on simple notions of observer-independent facts, but they do not show that your thoughts can nudge a detector into a preferred outcome. They show that quantum theory and classical assumptions about objectivity do not fit together as tidily as we once hoped Proietti et al. (2019).

So the article’s target is narrow: has human thinking been shown to influence quantum states, rather than merely interpret them? On that question, the evidence is much weaker than the rhetoric around it.

A compact map of the evidence#

What would convincing evidence that thought influences quantum states actually look like?#

Scott Alexander revived Allan Crossman’s rude but useful line that parapsychology functions as “the control group for science” Alexander (2014). The point is not that anomalous researchers are fools. Most are not. The point is that if weak methods can repeatedly generate exciting positives in a domain where the effect size is tiny and the priors are hostile, then the methods are the main character.

For this literature, a genuinely convincing result would need at least four things:

1. A predeclared analysis that does not depend on fringe choice, lag choice, trimming choice, or post-hoc metric migration.
2. Matched sham conditions run through the same hardware and timing pipeline, so the instrument can betray itself before metaphysics gets blamed.
3. Independent labs—including skeptical labs—getting the same sign and comparable effect size.
4. Public raw data, code, and hardware diagnostics so that subtle preprocessing or drift can be audited.

That is not overkill. In the double-slit line, the claimed effect size is on the order of 0.001%; in the sham-based critique of that same paradigm, the identified false-positive effect was 0.0159%, roughly an order of magnitude larger than the claim it was meant to test Walleczek & von Stillfried (2019). In that regime, the universe is perfectly capable of manufacturing fake profundity out of signal processing lint.

Did double-slit consciousness experiments survive scrutiny?#

The ordinary double-slit experiment is already weird enough to ruin a perfectly good afternoon. But it does not by itself imply that consciousness is the operative variable. Physicist Massimiliano Sassoli de Bianchi’s critique of the Radin line is blunt: quantum mechanics does not need a “psychophysical ingredient” to explain measurement, and double-slit interference experiments do not automatically test the mind’s role in collapse Sassoli de Bianchi (2013).

Dean Radin and colleagues nevertheless ran an ambitious program. In their 2012 paper, participants were periodically instructed to focus attention toward a sealed double-slit optical system or to relax. Across six experiments with 137 people and 250 sessions, the authors reported that a preplanned spectral ratio shifted in the predicted direction, with 250 no-observer control sessions showing no such effect Radin et al. (2012). That is exactly the sort of result that makes otherwise sensible people start talking like 1930s mystics with oscilloscopes.

Then the follow-up weather moved in.

Nicolas Tremblay independently reanalyzed a two-year dataset from this program and found that the headline significance depended on an erroneous trimming procedure that could severely underestimate p-values. He also showed that the result was not robust to apparently discretionary choices involving fringe selection, year aggregation, sign inversion, lag selection, and multiple-testing correction. The directional shifts did not disappear entirely, but after conservative correction they did not count as evidence Tremblay (2019).

The sham critique was harsher still. Walleczek and von Stillfried used what they call an advanced meta-experimental protocol and found a statistically significant false-positive effect in a sham condition—without test subjects present. They reported a 50% false-positive detection rate in the commissioned replication, with a sham effect size of 0.0159%, while the claimed consciousness effect for the paradigm had been about 0.001% Walleczek & von Stillfried (2019). That is not a peccadillo. It means the instrument-plus-analysis stack was capable of generating the very sort of signal the theory wanted to see.

Radin’s group pushed back. In their response, they argued that the sham critique mishandled multiplicity: after false-discovery-rate correction, none of the eight mean-based sham tests remained significant. They also argued that the broader literature matters more than one commissioned study, pointing to 28 related experiments, 11 individually significant, with a tiny cumulative binomial probability Radin et al. (2020). Later, Radin and Delorme proposed that the relevant signal may appear more cleanly as a variance shift than as a mean shift, reporting significant variance differences in a two-year online dataset Radin & Delorme (2022).

This reply deserves to be heard, but it does not really rescue the evidential situation. When a putative signal migrates from one metric to another, from direct effect to cumulative corpus, from mean to variance, from prediction to post-hoc reinterpretation, that may reflect genuine subtlety—or it may reflect a literature learning to metabolize failure. At present the double-slit line looks less like a clean discovery and more like a high-end confusion engine: interesting, not decisive, and still too friendly with measurement bias.

Can people bias quantum random number generators?#

This is the largest and statistically most developed branch of the literature. It is also the one that most resembles the user’s intuitive example: a machine emits allegedly quantum randomness, a person intends “up” or “down,” and researchers ask whether the output leans accordingly.

The benchmark summary is Bösch, Steinkamp, and Boller’s 2006 meta-analysis of 380 RNG studies. They found a significant but very small overall effect Bösch et al. (2006). That sounds promising until one reads the companion reply, where the same authors emphasize that the dataset also showed a small-study effect and extreme heterogeneity, and that publication bias remained the most parsimonious explanation unless stronger prospective controls were used Bösch, Steinkamp, & Boller (2006b).

These devices were not always literally Geiger-counter decay rigs. PEAR-era systems and their descendants used several physical noise sources, including quantum tunneling and thermal noise in resistors Jahn et al. (2000).² For the present question, that is close enough: the claim is still that mental states bias a physically random device.

The PortREG consortium replication is one of the most informative documents in this literature. Three labs, identical noise-source equipment, 227 operators, 750 experimental series, and a pre-agreed primary criterion. The mean deviations went in the intended direction—but the overall size was about an order of magnitude smaller than the prior Princeton benchmark and did not reach persuasive significance Jahn et al. (2000). The authors then explored “structural anomalies” in the data. Interesting, perhaps, but also a very standard move in controversial literatures: the headline effect weakens, and subtler patterns are invited to dinner.

A later Bayesian test went harder in the other direction. Maier and colleagues ran an online micro-PK experiment with 12,571 participants and found BF01 = 10.07, strong evidence for the null on the aggregate mean Maier et al. (2018). Rather than stop there, they proposed that the sequential data showed an oscillatory pattern that might reflect a different sort of PK. Physicist Hartmut Grote answered that the main result was simply strong evidence against micro-PK and that the oscillation idea was post-hoc and unsupported Grote (2018).

Then the literature produced one of its purest little morality plays. A preregistered follow-up aimed to test a correlational micro-PK effect discovered that the earlier signal had depended on a coding error involving eight participants. Once corrected, the original correlation vanished (BF01 = 49.48), and the new data again supported the null Maier & Dechamps (2022). Even so, the authors continued discussing experimenter effects, expectation effects, and decline effects as possible ways to understand the pattern Maier & Dechamps (2022).

That is the RNG story in miniature: a small positive literature, then heterogeneity, then weak replications, then nulls, then post-hoc structure, then experimenter effects, then a theory flexible enough to survive on volatility if not on means. This is not nothing. But it is not what robust discovery looks like.

Have direct consciousness-causes-collapse experiments worked?#

If the double-slit and RNG literatures are indirect, the Hall/Bierman line is more direct. It asks, in effect: if consciousness is the thing that collapses the wavefunction, can we design a delayed-observation setup where this matters?

The classic Hall-style experiment was simple. A radioactive decay event was measured; observer 1 sometimes saw it first, and observer 2 later had to guess whether that prior observation had occurred. Result: chance performance. Observer 2 guessed 50% correct. On a simple consciousness-causes-collapse reading, that is a failure Bierman (2003).

Dick Bierman’s 2003 conceptual replication was more sophisticated. He increased the delay to around one second and used EEG signatures of early brain processing rather than verbal report. The paper reported significant differences in observer-2 brain responses depending on whether observer 1 had already looked, with an exact binomial p <.02 Bierman (2003). That is genuinely interesting. But the same paper then refuses to overclaim: the author says the result is not enough to unequivocally accept the hypothesis and adds, correctly, “Strong claims need strong evidence” Bierman (2003). That sentence has aged better than most of the field.

There is also a conceptual problem here. De Barros and Oas argued that attempts to decisively falsify consciousness-causes-collapse are often flawed, and that under reasonable assumptions the hypothesis may even become effectively unfalsifiable de Barros & Oas (2017). That does not make the hypothesis false. It does make it a slippery empirical object. A theory that can always retreat to “you did not isolate consciousness correctly” is a theory that is hard to reward for surviving.

Recent work by Richard Lucido tries a clever side door. Instead of asking observers to guess whether collapse occurred, Lucido uses subliminal priming. Random fluctuations in radioactive decay generate primes that are either consciously observed in advance or left unobserved; later reaction-time differences are then interpreted as indirect evidence about whether collapse had already happened. The 2023 paper reported support for the CCC interpretation but explicitly said replication would be needed for confidence Lucido (2023). A 2024 follow-up reported similar support Lucido (2024), and a 2025 extension asked, with commendable zoological ambition, whether cats count as collapse-inducing observers Lucido (2025). Essentia promoted this line sympathetically in 2025 Essentia Foundation (2025).

My verdict here is the same as with the other literatures: clever, imaginative, not yet load-bearing. These papers live in niche venues, are mostly self-replicated so far, and have not yet been put through the kind of independent, adversarial, hardware-aware audit that the claim demands. A neat experimental angle is not the same thing as a settled result.

What did the Bem saga teach this literature?#

Daryl Bem’s precognition program is not a quantum-state-control literature, but it belongs here because it is the nearest cousin in evidential ecology.

In 2011, Bem reported nine experiments suggesting anomalous retroactive influences on cognition and affect Bem (2011). Wagenmakers and colleagues replied that the analyses were partly exploratory and that one-sided p-values overstated the evidence Wagenmakers et al. (2011). Bem, Utts, and Johnson responded that the Bayesian critique embedded unrealistically skeptical priors and that Bayesian methods themselves contain traps if used carelessly Bem, Utts, & Johnson (2011).

Then came exact preregistered replications: three attempts, combined n = 150, combined p = .83, no support Ritchie, Wiseman, & French (2012). Later, Bem and collaborators countered partly with a meta-analysis of 90 experiments from 33 labs, claiming an overall effect above 6 sigma and a Bayes factor above 10^9 Bem et al. (2016).

Why drag this into a quantum article? Because the same pattern recurs:

• an initially exciting small effect,
• critiques focusing on analysis and design,
• failed or weak direct replications,
• then rescue by broader meta-analysis or more elaborate post-hoc interpretation.

The moral is not “psi is impossible.” The moral is that meta-analytic significance is not a spell. It does not automatically convert flexible, weakly controlled, or subtly confounded paradigms into strong evidence. Alexander’s methodological point bites hard here: exact replication of a confounded design can replicate the confound Alexander (2014).

That is the real relevance of Bem to the quantum-consciousness debate. Not because precognition and wavefunction collapse are the same claim. Because they inhabit the same inferential biome.

What do proponents say in reply?#

The fairest skeptical review has to state the best comeback before dissecting it.

Proponents usually say some combination of the following:

• The effects are ultra-weak, so harsh controls can wash them out.
• Mean shifts are the wrong statistic; the real signal may appear as variance, oscillation, decline effects, or contextual dependence Radin & Delorme (2022), Maier & Dechamps (2022).
• Broad literatures matter more than single critiques, so the cumulative meta-analytic or corpus-level signal should dominate local failures Radin et al. (2020), Bem et al. (2016).
• Skeptics underestimate experimenter effects, expectancy effects, or subtle observer-dependent context that the phenomenon itself may require Maier & Dechamps (2022).

None of these replies is absurd. The trouble is cumulative. Each one makes the target harder to hit and easier to retrofit. A phenomenon that can disappear in the mean, reappear in the variance, decline under replication, depend on the right experimenter mindset, and survive mainly at the level of cumulative anomaly is a phenomenon that has become epistemically gelatinous. It may still be real. But it is not in good evidential shape.

What do physicists—including some on Essentia—actually argue?#

This is the part often mangled online. Serious idealist-adjacent or consciousness-friendly physicists absolutely exist. They are not interchangeable with the meme version of “quantum proves mind magic.”

On Essentia, Markus Müller explicitly says his goal is not to interpret quantum mechanics in the usual picture-book way of telling us what is “really going on in the quantum world.” Instead, he treats quantum states as a catalog of probabilities of private future experiences and argues that Bell-type results trouble a naïve “container view” of an objective material world Müller (2021). He also says there is currently no hope of direct experimental test for his broader approach Müller (2021). That is a first-person reconstruction of physics, not evidence that concentration changes interference fringes.

Also on Essentia, Lorenzo Catani argues that the double-slit experiment does not capture the essence of quantum theory Catani (2023). Emily Adlam frames the issue in terms of confirmation and intersubjective confirmation Adlam (2023). Again: interpretation, epistemology, observer-dependence—not proven psychokinesis.

The formal literature has the same shape. Chalmers and McQueen explicitly say simple consciousness-collapse models are already ruled out by the quantum Zeno effect, while more complex versions remain compatible with evidence and worth exploring Chalmers & McQueen (2021). Okon and Sebastián go further and build a consciousness-linked collapse model they say is fully consistent with materialism Okon & Sebastián (2018). Adrian Kent reviews the recent upsurge of interest in tying collapse to measures of consciousness Kent (2020/2021). Meanwhile, collapse models in general are increasingly constrained by actual experiments Carlesso et al. (2022), Bassi, Dorato, & Ulbricht (2023).

That last point matters. Even if some consciousness-linked collapse model were eventually vindicated, it would not automatically hand analytic idealism the crown jewels. It might fit dual-aspect monism, neutral monism, or even a broadened materialism just as well. A positive result here would be philosophically explosive, yes. But it would not come with a prepaid metaphysics.

So what should a careful reader believe right now?#

Here is the short, unscented verdict.

There has been intriguing evidence. It mostly did not hold up.

The strongest positive findings in this area tend to shrink, split, migrate, or soften under stricter controls, broader reanalysis, sham protocols, preregistration, and hostile follow-up. The most likely explanation today is not that every positive result is fraud or idiocy. It is that the effects are either nonexistent or so tiny that they are easily mimicked by measurement bias, preprocessing discretion, publication bias, expectancy effects, and the general perversity of low-signal experiments Tremblay (2019), Walleczek & von Stillfried (2019), Bösch et al. (2006), Maier et al. (2018).

That conclusion has an important boundary condition. It does not show that consciousness is epiphenomenal. It does not prove physicalism. It does not bury idealism. It only tells us that double-slit attention studies, QRNG intention studies, Hall-style delayed-observation tests, and current subliminal-priming collapse paradigms are not yet strong empirical evidence that thinking changes quantum states.

What would change my mind? A multicenter preregistered protocol shared by sympathetic and skeptical labs; matched sham conditions; hardware audits; public data and code; and an effect that survives neighboring analytic choices while remaining the same sign and scale across sites. That would be interesting in the good, expensive way.

Until then, quantum mechanics gives consciousness plenty of metaphysical leverage, but not yet a clean laboratory lever. Quantum weirdness may still matter profoundly for theories of mind. But the evidence that thinking itself has been shown to influence quantum states is, at present, weak. The control group has, once again, been controlled.

Footnotes#

Sources#

1. Alexander, Scott. “The Control Group Is Out Of Control.” LessWrong, 2014.

2. Chalmers, David J., and Kelvin J. McQueen. “Consciousness and the Collapse of the Wave Function.” In Consciousness and Quantum Mechanics. 2021. arXiv:2105.02314.

3. Okon, Elias, and Miguel Ángel Sebastián. “A Consciousness-Based Quantum Objective Collapse Model.” Synthese (accepted version on arXiv), 2018. arXiv:1801.05487.

4. Kent, Adrian. “Collapse and Measures of Consciousness.” Foundations of Physics 51 (2021). arXiv:2009.13224.

5. Proietti, Massimiliano, et al. “Experimental test of local observer independence.” Science Advances 5 (2019): eaaw9832. doi:10.1126/sciadv.aaw9832

6. Sassoli de Bianchi, Massimiliano. “Quantum measurements are physical processes. Comment on ‘Consciousness and the double-slit interference pattern: Six experiments.’” Physics Essays 26 (2013). doi:10.4006/0836-1398-26.1.15

7. Radin, Dean, et al. “Consciousness and the double-slit interference pattern: Six experiments.” Physics Essays 25 (2012): 157–171. doi:10.4006/0836-1398-25.2.157

8. Tremblay, Nicolas. “Independent re-analysis of alleged mind-matter interaction in double-slit experimental data.” PLOS ONE 14 (2019): e0211511. doi:10.1371/journal.pone.0211511

9. Walleczek, Jan, and Nikolaus von Stillfried. “False-Positive Effect in the Radin Double-Slit Experiment on Observer Consciousness as Determined With the Advanced Meta-Experimental Protocol.” Frontiers in Psychology 10 (2019): 1891.

10. Radin, Dean, et al. “Commentary: False-Positive Effect in the Radin Double-Slit Experiment on Observer Consciousness as Determined With the Advanced Meta-Experimental Protocol.” Frontiers in Psychology 11 (2020).

11. Radin, Dean, and Arnaud Delorme. “Psychophysical Effects on an Interference Pattern in a Double-Slit Optical System: An Exploratory Analysis of Variance.” Journal of Anomalous Experience and Cognition 2 (2022). doi:10.31156/jaex.24054

12. Bösch, Holger, Fiona Steinkamp, and Emil Boller. “Examining psychokinesis: The interaction of human intention with random number generators—a meta-analysis.” Psychological Bulletin 132 (2006): 497–523. doi:10.1037/0033-2909.132.4.497

13. Bösch, Holger, Fiona Steinkamp, and Emil Boller. “In the Eye of the Beholder: Reply to Wilson and Shadish (2006) and Radin, Nelson, Dobyns, and Houtkooper (2006).” Psychological Bulletin 132 (2006): 533–537. doi:10.1037/0033-2909.132.4.533

14. Jahn, R. G., et al. “Mind/Machine Interaction Consortium: PortREG Replication Experiments.” Journal of Scientific Exploration 14 (2000): 499–555.

15. Maier, Markus A., Moritz C. Dechamps, and Matthias Pflitsch. “Intentional Observer Effects on Quantum Randomness: A Bayesian Analysis Reveals Evidence Against Micro-Psychokinesis.” Frontiers in Psychology 9 (2018): 379. doi:10.3389/fpsyg.2018.00379

16. Grote, Hartmut. “Commentary: Intentional Observer Effects on Quantum Randomness: A Bayesian Analysis Reveals Evidence Against Micro-Psychokinesis.” Frontiers in Psychology 9 (2018): 1350. doi:10.3389/fpsyg.2018.01350

17. Maier, Markus A., and Moritz C. Dechamps. “A Pre-Registered Test of a Correlational Micro-PK Effect: Efforts to Learn from a Failure to ‘Replicate.’” Journal of Scientific Exploration 36 (2022). doi:10.31275/20222235

18. Bierman, Dick J. “Does Consciousness Collapse the Wave Function.” Mind and Matter (accepted manuscript), 2003. arXiv:physics/0312115

19. de Barros, J. Acacio, and Gary Oas. “Can We Falsify the Consciousness-Causes-Collapse Hypothesis in Quantum Mechanics?” Foundations of Physics (arXiv version), 2017. arXiv:1609.00614

20. Lucido, Richard J. “Testing the Consciousness Causing Collapse Interpretation of Quantum Mechanics Using Subliminal Primes Derived from Random Fluctuations in Radioactive Decay.” Journal of Consciousness Exploration & Research 14 (2023): 185–194.

21. Lucido, Richard J. “Replication of Results from a Test of the Consciousness Causes Collapse Interpretation of Quantum Mechanics Using Subliminal Priming Methodology.” Journal of Consciousness Exploration & Research 15 (2024).

22. Lucido, Richard J. “Do Cats Collapse the Wave Function? Confronting the Measurement Problem with Subliminal Priming.” Journal of NeuroPhilosophy 4 (2025). doi:10.5281/zenodo.15003998

23. Essentia Foundation. “Has experimental psychology proven that consciousness causes the collapse of the wave function?” 2025.

24. Bem, Daryl J. “Feeling the future: experimental evidence for anomalous retroactive influences on cognition and affect.” Journal of Personality and Social Psychology 100 (2011): 407–425. doi:10.1037/a0021524

25. Wagenmakers, Eric-Jan, et al. “Why psychologists must change the way they analyze their data: the case of psi: comment on Bem (2011).” Journal of Personality and Social Psychology 100 (2011): 426–432.

26. Bem, Daryl J., Jessica Utts, and Wesley O. Johnson. “Must Psychologists Change the Way They Analyze Their Data?” Journal of Personality and Social Psychology 100 (2011): 716–719.

27. Ritchie, Stuart J., Richard Wiseman, and Christopher C. French. “Failing the Future: Three Unsuccessful Attempts to Replicate Bem’s ‘Retroactive Facilitation of Recall’ Effect.” PLOS ONE 7 (2012): e33423. doi:10.1371/journal.pone.0033423

28. Bem, Daryl, Patrizio Tressoldi, Thomas Rabeyron, and Michael Duggan. “Feeling the future: A meta-analysis of 90 experiments on the anomalous anticipation of random future events.” F1000Research 4 (2015/2016): 1188. doi:10.12688/f1000research.7177.2

29. Müller, Markus. “The physics of first-person perspective: An interview with physicist Dr. Markus Müller.” Essentia Foundation, 2021.

30. Catani, Lorenzo. “The double-slit experiment doesn’t reveal the essence of quantum weirdness.” Essentia Foundation, 2023.

31. Adlam, Emily. “Does science need intersubjective confirmation?” Essentia Foundation, 2023.

32. Carlesso, Matteo, et al. “Present status and future challenges of non-interferometric tests of collapse models.” Nature Physics 18 (2022): 243–250. doi:10.1038/s41567-021-01489-5

33. Bassi, Angelo, Mauro Dorato, and Hendrik Ulbricht. “Collapse Models: a theoretical, experimental and philosophical review.” Entropy 25 (2023): 645. doi:10.3390/e25040645