Research synthesis/ Ritual, evidence, and what modern life dropped/ July 2026

Wisdom hidden in plain sight

Tarot readings and healing crystals reliably produce experiences that feel more accurate and more significant than they should. The usual explanation — ritual psychology and self-belief — is right that nothing supernatural is involved and wrong about the mechanism, in a way more interesting than the folk account. What follows is that correction, then the same lens turned on the rest of human history: what was abandoned that worked, what is practised now that does not, and what is worth adopting, ranked.

I — Why the cards feel accurate

The short version: no self-deception is involved, and no belief is required for the effect to occur. That is not a charitable gloss — it is the specific finding of a body of research on honest placebo, and one of the stranger results in modern medicine.

Placebos work when the patient is told they are placebos

In 2010 Ted Kaptchuk's group at Harvard ran a trial on 80 irritable bowel syndrome patients. Half received nothing. Half received sugar pills in a bottle labelled "placebo pills," with the explicit verbal instruction that these were inert, made of an inactive substance, like sugar pills. No deception of any kind. The pills worked: d = 1.14 improvement on the IBS Global Improvement Scale at 11 days, and adequate relief in 59% of the placebo group versus 35% of controls.

That result has since scaled up. A 2025 meta-analysis in Scientific Reports pooled 60 randomised trials covering 4,648 participants and found an overall effect of SMD = 0.35 for open-label placebo across conditions.

The obvious objection is that people are simply reporting what they think they should report. In 2022 Fabrizio Benedetti's group tested this directly, using naloxone — a drug that blocks opioid receptors. If honest placebo analgesia were merely polite self-report, naloxone would do nothing to it. Instead, naloxone abolished the effect, which means non-deceptive placebo is running on the same endogenous opioid machinery as the deceptive kind. The same story holds in Parkinson's disease, where PET imaging shows placebo triggering real striatal dopamine release (de la Fuente-Fernández, Science, 2001).

The correction is precise: the mechanism is not self-belief but sincere engagement with a structured process. Those are different, and only one of them requires holding a false belief.

The theoretical account that makes this cohere is predictive processing. On this view the brain is not a passive sensor but a prediction engine, and perception — including the perception of one's own symptoms and internal states — is an inference combining top-down expectation with bottom-up signal. Ongaro & Kaptchuk argued in PAIN (2019) that symptom experience is itself constructed this way, which is exactly why an honest placebo need not deceive: the inference engine is not being lied to, it is being handed a high-precision prior.

Tarot stacks four mechanisms, only one of which is deflationary

Divination is not one phenomenon. When a reading lands, at least four distinct things are happening at once, and it is worth separating them because they have very different standing.

1. The Barnum effect — this is the debunking part

In 1949 Bertram Forer gave 39 students an identical personality sketch assembled from newspaper horoscopes, telling each it was individually prepared. Average rated accuracy: 4.26 / 5. Statements that are true of nearly everyone are experienced as uncannily specific, and confirmation bias does the rest — hits are remembered and misses forgotten. A meaningful share of "how did it know that" is this and nothing else.

2. Projective elicitation — the reader supplies the content

An ambiguous image prompts a self-generated narrative, which is the same mechanism the Rorschach was built on. The card holds no information about its reader; it provides an unstructured surface onto which they project what they already knew and had not articulated. This is well established as a general psychological principle. Applied to tarot specifically it remains an inference by analogy — I found no controlled trial of tarot as a reflective intervention, and that gap is worth stating plainly rather than papering over.

3. Ostensive detachment — the genuinely profound one

Pascal Boyer's 2020 analysis of why divination exists at all across nearly every human society identifies a feature he calls ostensive detachment: the oracle publicly signals that it is not the author of its own message. This matters more than it sounds. A difficult conclusion can enter awareness without anyone — including the person holding the cards — taking responsibility for having said it. The deck becomes a device for hearing something already suspected but not directly sayable.

4. Randomisation as a tool against fixation

Steven Levitt ran a study through a public website that flipped a coin for people facing real decisions — quit the job, end the relationship, break the habit. Over 20,000 flips. People whose coin told them to make a change were 25% more likely to actually make it, and reported greater happiness six months later. No mysticism, no interpretation, just an external randomiser breaking status-quo bias. Omar Moore made a related argument in 1957 about Naskapi hunters reading cracks in a burnt caribou shoulder blade to choose a hunting direction — functionally a randomiser preventing predictable, game-depleting patterns.

Crystals get a cleaner and less flattering verdict

Christopher French at Goldsmiths gave 80 people either real quartz or indistinguishable plastic fakes, after priming everyone with a leaflet describing the sensations they might feel. Both groups reported tingling, warmth, and heightened concentration at statistically identical rates. Nobody could tell which they were holding.

The honest reading is that the stone contributes nothing and the dedicated five minutes of instructed attention contributes everything. This is not a dismissal — five minutes of structured attention is a real intervention that most people never deliberately schedule. It is simply available for free, and the mineral is a delivery mechanism carrying a price.

Sourcing caveat

The French crystal study was presented at the British Psychological Society Centenary Conference in 2001 and exists as a conference abstract. It does not appear to have been published as a peer-reviewed journal article. It is widely cited as though it were. Treat it as presented research — suggestive and well-designed, but not through peer review.

Where this stops

Every effect described above is confined to subjective states and their physiological correlates. The same 2025 meta-analysis that found SMD = 0.35 for self-reported outcomes found SMD = 0.09 for objective ones — not statistically significant. Pooled objective tumour response in the placebo arms of oncology trials is about 1%. Hróbjartsson & Gøtzsche's Cochrane analyses, covering 202 trials across 60 conditions, concluded that placebo has no important clinical effect in general, with modest effects surviving mainly for patient-reported pain and nausea.

So: ritual reliably changes how a person feels, sometimes through measurable neurochemistry. It does not change facts about the world, it does not reach information the cards have no access to, and it does not move disease biology. That line is where the psychology of ritual ends and the claim that the cards know something begins.

Sources for this section
  • Kaptchuk et al., "Placebos without Deception," PLOS ONE, 2010.
  • "Effects of open-label placebos across populations and outcomes," Scientific Reports, 2025 — 60 RCTs, n=4,648.
  • Benedetti, Shaibani, Arduino & Thoen, naloxone-blocked open-label placebo analgesia, 2022.
  • de la Fuente-Fernández et al., "Expectation and Dopamine Release," Science, 2001.
  • Ongaro & Kaptchuk, "Symptom perception, placebo effects, and the Bayesian brain," PAIN, 2019.
  • Forer, "The Fallacy of Personal Validation," J. Abnormal & Social Psychology, 1949.
  • Boyer, "Why Divination? Evolved Psychology and Strategic Interaction in the Production of Truth," 2020.
  • Levitt, "Heads or Tails: The Impact of a Coin Toss on Major Life Decisions," NBER WP 22487, 2016.
  • Moore, "Divination — A New Perspective," American Anthropologist, 1957.
  • Hróbjartsson & Gøtzsche, Cochrane, "Placebo interventions for all clinical conditions," 2010.

II — How the ranking works

The ordering is by potential positive experiential impact on a life. That is deliberately not the same as evidence strength, and not the same as mortality benefit — a hazard ratio thirty years out is not an experience. The ranking is a composite of five axes, set out openly here so the weighting can be disputed rather than reverse-engineered.

Every ranked entry carries a five-segment strip showing its score on each axis. Segments run left to right; taller is stronger.

What the axes mean

Evidence quality gates entry and caps the ceiling — nothing speculative ranks high regardless of how appealing it is. Effect magnitude is documented size, not mere statistical significance. Experiential reach asks whether it changes the daily felt texture of being alive or only a statistic. Modern deficit asks how severely contemporary life removes it — a practice still available by default scores low here even when excellent. Accessibility is whether it is free and immediate or requires money, equipment, or a community that must first be found.

III — The list, ordered by impact

Twelve practices that clear the evidence bar and would meaningfully change a life. The first four are a tier apart, and they are all variations on a single theme that modern secular life has dismantled almost completely.

01

Recurring embodied membership in a group

Intended historically: religious congregation, guild, lodge, village. What it actually delivers: repeated, non-optional, face-to-face contact with the same people over years.

Strong

Holt-Lunstad's meta-analyses — 148 studies, 308,849 participants — put the mortality risk of weak social connection at roughly that of smoking 15 cigarettes a day, and a separate 2015 analysis found social isolation at OR 1.29, loneliness at 1.26, living alone at 1.32. VanderWeele's analysis of 74,534 women in the Nurses' Health Study found weekly-or-more religious service attendance associated with HR 0.67 for all-cause mortality over twenty years.

VanderWeele is explicit that the mediators are social integration, reduced depression, healthier behaviour, and a sense of purpose — not anything supernatural. The active ingredient is therefore available outside religion. What is not easily available is the structure: an obligation that recurs regardless of inclination, among people not selected for similarity. That is the part secular life has no replacement for, and it is why this ranks first despite scoring poorly on accessibility.

Honest caveat: this is observational epidemiology, and healthy-attender bias is real — sick people cannot easily attend. Repeated-measures designs mitigate it but cannot eliminate it.

02

Moving or singing in time with other people

Intended historically: worship, work song, war dance, processional. What it delivers: rapid, endorphin-mediated bonding that ordinary conversation cannot produce.

Strong

This is the delivery mechanism for the first entry, and it has better experimental evidence than the thing it delivers. Wiltermuth & Heath (Psychological Science, 2009) showed across three experiments that people who walked, sang, or moved in synchrony subsequently cooperated more in economic games requiring real personal sacrifice — and did so without any accompanying mood boost, meaning synchrony is doing the work directly rather than via feeling good.

Tarr, Launay, Cohen & Dunbar (Biology Letters, 2015) randomised 264 Brazilian adolescents across exertion and synchrony conditions and found both independently raised pain threshold — a standard proxy for endorphin release — with the combination largest. Pearce's seven-month study found singing groups reached the social closeness of craft groups in about a third of the time, the so-called ice-breaker effect.

Practically: choir, social dance, rowing, martial arts forms, congregational singing. The specific content is close to irrelevant. Synchrony and shared exertion are the variables.

03

Ritual at the hinge points — grief, transition, loss

Intended historically: wake, shiva, keening, funeral games, mourning dress. What it delivers: restored sense of control when control has been taken away.

Moderate–Strong

Norton & Gino (Journal of Experimental Psychology: General, 2014) ran three actual experiments across three kinds of loss — bereavement, relationship breakup, and lab-induced financial loss. Performing a ritual reduced grief, and the effect was mediated by restored perceived control. The finding that matters most: it worked on people who did not believe rituals work. This is not a belief effect.

Modern secular life is close to scriptless here. It has no equivalent of the shiva week, the year of mourning dress, or the wake — whose structural genius was to tell everyone else how to behave toward the bereaved, and to tell the bereaved when it was permitted to stop. The practice is nearly free and needs no institution, which is why it scores maximum on accessibility despite ranking third.

Important — read the honesty section

Norton & Gino 2014 is not retracted and appears to hold. But a cluster of adjacent, more famous ritual papers by overlapping authors has been retracted for data fabrication. See section VII.

04

Breathing at roughly six breaths a minute

Intended historically: rosary recitation, yogic pranayama, mantra, Gregorian chant. What it delivers: maximal baroreflex coupling between heart rate and blood pressure.

Strong

This is the single most striking convergence in the whole survey. Bernardi et al. published in the BMJ in 2001 a study recording people reciting the Latin Ave Maria rosary and, separately, a traditional yoga mantra. Both practices spontaneously slowed breathing to almost exactly six breaths per minute — which is the resonance frequency at which cardiovascular oscillations couple most powerfully. Two traditions, separated by continents and millennia, independently converged on the same physiological optimum without any theory of what they were doing.

The autonomic effect is well replicated: reliable increases in heart-rate variability measures. Be sceptical of the popular extension, though — a more recent meta-analysis found the effect on clinical blood pressure specifically is not significant, so "breathe slowly to lower your blood pressure" outruns the data.

It is free, portable, requires no equipment, and works within minutes. Best accessibility score on the list.

05

Eating with other people, regularly

Intended historically: the feast, the symposium, the potlatch, the family table. What it delivers: sustained low-intensity social contact bundled with something that must be done anyway.

Moderate

Dunbar's 2017 national survey of 2,000 UK adults found frequency of social eating predicted happiness, life satisfaction, trust, and network size — people who ate socially reported roughly double the number of people they could count on. Evening meals produced significantly more felt closeness than midday ones, especially with more people, laughter, and reminiscing.

The Korean KNHANES data adds the negative case: eating dinner alone was associated with OR 1.42 for depressive symptoms and 1.49 for suicidal ideation across 14,093 adults, fully adjusted. Interestingly, the suicidal ideation association disappeared among people who exercised regularly.

Two honest deflations. All of this is cross-sectional — happier, more trusted people get invited to more dinners, and that reverse path is not ruled out. And the famous family-dinner literature largely dissolves under proper controls: Musick & Meier's fixed-effects analysis of 17,977 adolescents found effects on adult wellbeing were entirely mediated through earlier adolescent wellbeing. Cohesive families eat together; the meal is more symptom than cause.

06

Anchoring the day to light

Intended historically: canonical hours, dawn prayer, sunset ritual, firelight evenings. What it delivers: circadian entrainment via retinal photoreceptors to the suprachiasmatic nucleus.

Strong (mechanism)

Morning light advances circadian phase by anywhere from thirty minutes to two hours depending on timing and intensity, and the phase-response mechanism is among the best-characterised in human physiology. In the other direction, cool white LED light suppresses melatonin around 12% versus 1.5% for incandescent or firelight-spectrum sources — meaning the shift from fire to screens is a real, measurable perturbation, not a wellness talking point.

But note what does not follow. The 2023 Cochrane review of blue-light filtering lenses across 17 RCTs concluded they "probably make no difference" to eye strain or sleep. The mechanism is solid; the product built on it is not. Morning outdoor light and dimmed evenings follow from the evidence. The eyewear does not.

07

Structured contemplation of one's own death

Intended historically: memento mori, Stoic premeditatio malorum, Buddhist maranasati, ars moriendi. What it delivers: recalibrated gratitude and prosociality via bounded confrontation with finitude.

Moderate (mechanism)

Here the evidence and the practice have to be carefully separated. I could find no controlled trial of the named branded practices — premeditatio malorum and maranasati are essentially untested as manualised interventions, and nearly everything written about them by those names is practitioner content.

What does exist is experimental work on the underlying mechanism. Frias and colleagues found that induced "death reflection" — bounded, deliberate contemplation, distinct from the threat-driven mortality salience of terror management research — measurably increased gratitude relative to controls. A related study found mortality salience increased prosocial behaviour at d = 0.70. The Scrooge effect literature shows similar results for generosity in economic games.

This ranks seventh rather than lower because the impact-per-unit-effort is extraordinary — it reframes everything downstream of it — and because modern life is uniquely structured to prevent it. The medieval ars moriendi tradition existed because dying was public, communal, and prepared for. Ours is none of those things.

08

Deliberate heat

Intended historically: Finnish sauna, Japanese onsen, Turkish hammam, Roman bath, sweat lodge. Historically framed as hygiene, purification, and sociality — not cardiovascular medicine.

Moderate–Strong

The Kuopio Ischaemic Heart Disease cohort followed 2,315 Finnish men for a median of 20.7 years. Against a once-weekly baseline, sauna 4–7 times per week showed HR 0.37 for sudden cardiac death, 0.50 for fatal cardiovascular disease, and 0.60 for all-cause mortality, with a clean dose-response and no threshold. A companion paper on the same cohort found HR 0.34 for dementia.

Those numbers are large enough to be suspicious, and the caveat is significant: nearly the entire flagship sauna literature comes from the same Finnish cohort, reanalysed repeatedly by the same group. There is one independent replication in a Japanese hot-bathing cohort of about 30,000. It is observational, and in a country where nearly everyone saunas, the people who sauna most may simply be the most robust. A 2026 UK study found sauna wellbeing effects were partly explained by increased social belonging — suggesting the communal framing carries weight independent of the heat.

09

Slow, skilled, weight-shifting movement

Intended historically: tai chi and qigong as martial art and moving meditation. What it delivers, measurably: balance, and a large reduction in falls with age.

Strong

A 2023 meta-analysis of 24 randomised trials found tai chi reduced fall risk in older adults at RR 0.76, with consistent improvement across timed-up-and-go, functional reach, Berg balance scale, and gait speed, and a dose-response with practice duration. This is among the most rigorously supported findings in the entire survey, and it is a millennia-old practice that was never designed with falls in mind.

The experiential case is blunt: a hip fracture at 75 is frequently the beginning of the end, and this is one of the few interventions with real trial evidence against it. Impact rises steeply with age, which is why it sits mid-list rather than top for a general reader.

10

Writing down what has not been said out loud

Intended historically: confession, teshuvah, Naikan self-accounting, examen. What it delivers: modest but real health and mood benefit from structured disclosure — and this is where tarot legitimately belongs.

Moderate

Pennebaker's original 1986 finding — students who wrote about trauma for fifteen minutes on four nights had roughly half the health-centre visits over the next six months — launched an enormous literature. The honest summary of that literature is Frattaroli's 2006 meta-analysis in Psychological Bulletin: across 146 randomised studies, the overall effect is r = 0.075. Real, replicated, and considerably smaller than the popular retelling.

This is the honest home for tarot. As prediction it is unsupported. As a structured prompt that forces articulation — with the added feature of ostensive detachment carrying an uncomfortable conclusion — it is a delivery vehicle for a mechanism that does have evidence. The cards do what a good journaling prompt does, with better theatre and a randomiser attached. That is not nothing, and it is also not divination.

Two adjacent practices with better effect sizes: forgiveness interventions (Wade et al., 53 studies, Δ+ = 0.56 versus no treatment) and self-compassion training (21 RCTs, medium-to-large effects). Both are structured repair, which is what confession traditions were.

11

Time outdoors, around two hours a week

Intended historically: not a practice at all — simply the default condition of human life until very recently.

Moderate

White et al. (Scientific Reports, 2019) analysed 19,806 people and found a threshold at 120 minutes per week of nature contact, above which odds of reporting good health rose to OR 1.59 and high wellbeing to 1.23, peaking around 200–300 minutes. No benefit accrued below the threshold, and it did not matter whether the time came in one visit or several.

The authors are refreshingly honest that this is cross-sectional and that nature contact explained about 1% of variance. Bratman's 2015 PNAS walking study is more mechanistically interesting — a 90-minute nature walk reduced subgenual prefrontal cortex activation at d = 1.01 versus an urban walk — but it had 38 participants and the behavioural rumination effect was only marginal at p = 0.07, a nuance the popular retelling always drops.

The forest-bathing phytoncide and natural-killer-cell story should be set aside entirely. That specific mechanistic claim rests on small, weakly controlled studies and is the weakest part of this literature.

12

Meditation, including the long silent kind

Intended historically: contemplative traditions across Buddhism, Christianity, Islam, and Hinduism. What it delivers: moderate reductions in depression and anxiety.

Moderate–Strong

Included partly for symmetry: meditation is frequently lumped in with the woo, and it is one of the better-evidenced entries here. A meta-analysis of 42 randomised trials found moderate-to-large effects on depression (g = 0.59) and anxiety (g = 0.56), with smaller significant effects on PTSD.

Silent retreat specifically is thinner than its reputation: a 2025 systematic review of Vipassana research since 2010 found only 11 qualifying studies, of which just three were randomised, and rated the overall evidence "moderate" with high risk of bias and no blinding. The practice is well supported. The intensive retreat format is less so.

IV — The full ledger

Every practice surfaced by the research, what it was intended to do, and how well it holds up. Rows expand on click, each carrying the named studies, the figures, and the caveat that qualifies them. Column headings sort; the selectors filter by tier and category. The speculative and the debunked are included, so the whole field is visible in one place.

Practice ▾ Category ▾ Intended to do ▾ Evidence ▾

V — What does not work

Rated three ways, because "bogus" is too blunt an instrument. A practice can have a false mechanism and still produce a real effect through an honest route — and saying so is more useful than a flat dismissal, since it separates what is worth keeping from what is worth ceasing to pay for.

Mechanism false, effect real

CrystalsMechanism false

Real quartz and plastic fakes produced statistically identical reported sensations in 80 primed participants (French, Goldsmiths, 2001 — conference abstract, not peer-reviewed).

What it still doesCreates a scheduled, bounded interval of focused attention with a plausible frame attached. That is a genuine intervention. The stone is an inert prop, removable without loss.

Tarot as divinationMechanism false

No evidence cards carry predictive information. Apparent accuracy is fully accounted for by the Barnum effect, confirmation bias, and cold reading.

What it still doesFunctions as a projective prompt, a randomiser against decision paralysis, and — via ostensive detachment — a way to hear an unwelcome conclusion without owning its authorship. Legitimate as structured reflection. See entry 10.

HomeopathyUnsupported

Australia's NHMRC reviewed 176 studies across 61 conditions in 2015: "there are no health conditions for which there is reliable evidence that homeopathy is effective." The Lancet's 2005 Shang analysis concluded effects are compatible with placebo; its accompanying editorial was titled "The End of Homoeopathy." NHS England ended routine funding in 2017, upheld by the High Court in 2018.

What it still doesHomeopathic consultations routinely run 45–60 minutes against a rushed conventional appointment. That is a real therapeutic ritual with a real context effect — attributable to the consultation, not the remedy.

Reiki & therapeutic touchUnsupported

Emily Rosa's experiment — conducted at age nine, published in JAMA in 1998 — asked 21 experienced practitioners to detect an unseen human energy field across 280 trials. They scored 44%, below chance, with no correlation to experience.

What it still doesSustained calm attention and human touch have their own well-documented anxiolytic effects. Those are real. The biofield that was specifically tested is not.

Chakra balancing & energy healingUnsupported

No validated instrument detects chakras as discrete energy centres. Claimed electromagnetic-signature studies are not independently replicated.

What it still doesThe practices bundled with it — breathwork, body scanning, meditation — have genuine evidence. That is evidence for those practices, not for the model wrapped around them.

Partly supported, but not for the stated reason

AcupunctureMixed

The most careful reading available. Vickers' individual-patient-data meta-analysis of roughly 20,000 patients found acupuncture beat both usual care and sham — but the margin over sham was small: 0.23 SD for back and neck pain, 0.16 for osteoarthritis, 0.15 for chronic headache. Meanwhile sham needling at deliberately wrong points performs nearly as well as correct points, which falsifies meridian theory specifically.

What it still doesProduces a real, modest, reproducible analgesic effect from needling plus an extended attentive session. Vickers himself concludes factors beyond needle placement drive most of it. Reasonable for chronic pain; the qi framework is decoration.

MultivitaminsGenuinely unresolved

The USPSTF issued an "insufficient evidence" statement in 2022 for cardiovascular and cancer prevention, and recommends against beta-carotene and vitamin E for that purpose. But COSMOS — three cognitive substudies, over 5,000 adults aged 65+ — found a daily multivitamin improved cognition, roughly equivalent to slowing cognitive ageing by about two years over three years.

The catchCOSMOS was funded in part by Pfizer Consumer Healthcare/Haleon, which makes the Centrum Silver product used in the trial, and by Mars Edge. The companies state they had no role in design or analysis. It remains a single trial family awaiting independent replication. Don't flatten this to a verdict in either direction.

Cold plunges & the Wim Hof methodOverhyped

The famous Kox PNAS 2014 endotoxin study is real and interesting, but it had 24 subjects, wasn't blinded, tested a bundled package of breathing plus cold plus meditation, and cannot isolate cold from the hyperventilation that independently raises adrenaline. A depression RCT found the method produced improvements — and so did the active control, comparably. A 2024–25 meta-analysis of 11 cold-water-immersion studies found inflammatory markers increased post-immersion, stress reduction appeared only at 12 hours, and mood did not improve at all.

What it still doesBrown adipose tissue activation is genuine physiology. An acute cold shock reliably produces a sharp, short-lived lift in alertness and mood, and those effects are real. The immune claims are not established.

Time-restricted eatingBest trial was negative

The TREAT trial (Weiss, UCSF, JAMA Internal Medicine, 2020) randomised 116 adults to 16:8 eating versus three structured meals. No significant weight difference — about 2 lb versus 1.5 lb over 12 weeks — and the fasting group lost more lean mass. Weiss, himself a long-time practitioner, said he was surprised. Broader meta-analyses show modest losses explained by unintentional calorie reduction, and adding time restriction to calorie restriction confers no extra benefit.

What it still doesWhere a narrow eating window reduces intake without deliberate counting, it functions as an adherence tool. The metabolic "eating window" mechanism is not supported.

Blue ZonesContested

Saul Newman won the 2024 Ig Nobel in Demography for work arguing that extreme-age clusters track missing birth certificates and pension fraud rather than longevity: only 18% of "validated" supercentenarians have a birth certificate, birthdates cluster suspiciously on multiples of five, and introducing birth certificates to a region drops recorded supercentenarians by 69–82%. Dan Buettner's team disputes this vigorously and published a rebuttal in The Gerontologist. Newman's core paper remains a preprint.

Honest positionThis is a live unresolved dispute, not a settled debunk. What survives either way: the individual habits Blue Zones popularised — plant-heavy diet, walkability, strong social ties, sense of purpose — are each independently supported by unrelated evidence. It is the demographic claim about those five regions that is shaky.

Unsupported outright

AstrologyUnsupported

Shawn Carlson's double-blind test in Nature (1985) had astrologers match natal charts to personality inventories at chance level, using a protocol they agreed in advance was fair. Dean's "time twins" work found people born minutes apart no more similar across 100+ characteristics than random pairs. Dean & Kelly's 2003 review of 40+ controlled studies found astrologers could not beat chance even on extraversion.

CaveatsCarlson is genuinely contested — Ertel's 2009 reanalysis argues a subsample beat chance. The time-twins study I could only verify through secondary journalism. The skeptical case rests on convergence across several lines, not any single study.

Antioxidant megadosesActively harmful

The most important entry in this section. Bjelakovic's Cochrane review pooled 78 randomised trials and 296,707 participants. In low-bias trials, beta-carotene showed RR 1.05 and vitamin E RR 1.03 for increased all-cause mortality. Their conclusion: no evidence supports antioxidant supplements for prevention, and beta-carotene and vitamin E appear to increase mortality.

What it still doesNothing good. This is the one place on the list where the popular practice is not merely useless but associated with harm.

Vitamin D for general preventionUnsupported

VITAL randomised 25,871 adults to 2,000 IU daily with a median 5.3 years of follow-up. No significant reduction in major cardiovascular events (HR 0.97), cancer incidence, or all-cause mortality (HR 0.99). Essentially null on the two claims that drove the boom.

What it still doesTreats actual diagnosed deficiency, which is a real condition. Supplementing an adequate person to prevent disease is not supported.

"Adrenal fatigue"Not a diagnosis

Cadegiani & Kater's 2016 systematic review in BMC Endocrine Disorders is titled, with unusual directness, "Adrenal fatigue does not exist." No endocrinology society recognises it.

What it still doesNames a real cluster of felt experience — exhaustion, poor sleep, low drive — which usually has an identifiable cause worth actually investigating. The label diverts from that.

Detoxes, cleanses, foot pads, colonicsUnsupported

The liver, kidneys, lungs, gut and skin run continuous detoxification; the kidneys alone filter around 200 quarts of blood daily. Detox foot pads darken through oxidation and moisture whether or not they touch skin. Colonic irrigation carries genuine risk of electrolyte imbalance and bowel perforation.

What it still doesCutting alcohol and processed food for a week produces real short-term effects. That is ordinary dietary change, and the juice-heavy versions — high sugar, low protein and fibre — can be counterproductive.

Ear candlingUnsupported and dangerous

The FDA finds no valid scientific evidence of medical benefit. Burning an identical candle without inserting it in an ear produces the same waxy residue — it is candle byproduct. Documented harms include burns, canal blockage, and perforated eardrums.

What it still doesNothing. This one has no redeeming mechanism.

Grounding mats, EMF shields, alkaline water, raw water, sea moss, IV drips, jade eggsUnsupported

Grouped because the pattern is identical. No plausible mechanism, no independent replication, and in several cases regulatory action: the FTC settled with sellers of radiation-reducing phone stickers in 2011 and warned about 5G protection products in 2020; Goop paid $145,000 over jade egg claims in 2018. Stomach acid neutralises alkaline water on contact. Raw water carries real pathogen risk. Sea moss products have tested above California limits for cadmium and lead, with unregulated iodine risking thyroid dysfunction. IV vitamin drips deliver what the kidneys promptly excrete, with infusion-site infection risk attached.

What they still doMostly they cost money. Several carry non-trivial physical risk. An EMF sticker over an antenna would, if anything, make the phone transmit harder.

Sleep-tracker perfectionismIatrogenic

Worth naming because it is the most modern failure mode here. Baron et al. coined "orthosomnia" in the Journal of Clinical Sleep Medicine in 2017 to describe patients who develop insomnia-like anxiety chasing perfect wearable sleep scores, sometimes trusting an inaccurate device over both their own felt rest and a clinician's reassurance.

What it still doesUsed non-obsessively, tracking retains real diagnostic value. The harm is the pursuit of the number, which is a wellness-era ritual with a genuinely bad mechanism.

The symmetry check — dismissed, but they hold up

A debunking list with no entries running the other way would be a bias, not an analysis. These are routinely lumped in with the woo and shouldn't be:

VI — Instruments pointed at the old machinery

Where current research is deliberately engineering what ritual produced by accident. Included partly because it is the strongest evidence that the traditional practices were doing something real: it can now be watched on instruments.

Wearables on people walking through fire

Dimitris Xygalatas's field studies are the cleanest example in the literature. At the San Pedro Manrique fire-walking festival in Spain, his team put heart-rate monitors on 12 fire-walkers and 26 spectators. Because walkers cross the coals one at a time rather than in unison, movement synchrony could be ruled out. The finding, published in PNAS in 2011: walkers' heart rates rose and fell almost in unison with those of their friends and relatives in the crowd, but not with unrelated spectators. Durkheim's "collective effervescence," previously a piece of sociological theory, turned out to be measurable on a chest strap and specific to existing social bonds.

His Mauritius work went further. Comparing a low-ordeal ritual (collective prayer) against the high-ordeal Thaipusam Kavadi ritual — body piercing with hooks and skewers, hours of carrying heavy structures — and measuring real anonymous charitable donations afterward, both performers and mere observers of the painful ritual gave significantly more. The more pain reported, the more was donated (Psychological Science, 2013).

Psychedelic-assisted therapy, with the critique attached

Johns Hopkins' 2016 trial of psilocybin in 51 patients with life-threatening cancer diagnoses found substantial sustained reductions in depression and anxiety, with about 80% showing clinically significant improvement at six months. The detail that matters here: the intensity of the "mystical-type experience" during the session statistically mediated the size of the benefit — measured on an instrument descended from William James's phenomenology of religious experience. Science is now using religious-experience scales as dependent variables.

The critique is serious and belongs in the same breath. The FDA rejected Lykos Therapeutics' MDMA-assisted therapy for PTSD in August 2024 after an advisory committee voted 10–1 against. The central problem is functional unblinding: participants can obviously tell whether they received a psychedelic, which collapses the double blind and lets expectancy inflate results. Muthukumaraswamy's 2021 analysis found psilocybin trials show the highest unblinding rates of all, meaning this critique applies to the Hopkins and Imperial work too, not only to the rejected MDMA application. Treat effect sizes in this field as likely overestimated.

The rest, briefly

VII — Where this evidence is shaky

A document like this is only worth as much as the things it admits. Four problems surfaced during the research that a less careful version would quietly omit.

Much of the famous ritual research was fabricated

This is the big one. "Don't Stop Believing: Rituals Improve Performance by Decreasing Anxiety" (Norton, Gino, Brooks et al., 2016) is the single most-cited demonstration that arbitrary, invented rituals reduce anxiety — the salt-sprinkling, paper-crinkling study that appears in virtually every popular treatment of this subject. It was retracted in November 2024. The authors requested the retraction themselves after an audit found implausible patterns in the heart-rate data — streaks of repeated and round-numbered values — and stated they no longer had confidence in the data's integrity.

It is not isolated. "Enacting Rituals to Improve Self-Control" (2018) was retracted after an audit found 15 participants' condition codes had been altered between the original dataset and the reported results. "Rituals Enhance Consumption" (2013) has not been retracted but remains under unresolved data-provenance review. All sit within the broader data-fabrication findings concerning Francesca Gino at Harvard Business School, surfaced by the Data Colada investigators in 2023.

The widely repeated claim that arbitrary rituals work just as well as meaningful ones — and that merely labelling an action a "ritual" is what produces the benefit — has lost its best evidence.

The broader thesis survives, but in a weaker and more hedged form than any popular account admits. What holds up comes from uninvolved teams: Hobson, Bonk & Inzlicht (PeerJ, 2017) found a week of practising a novel ritual sequence reduced the brain's error-related negativity signal at d = 0.51 without harming performance — ritual buffers the sting of mistakes. And Lang, Krátský & Xygalatas (2022) pre-registered a study of 268 people stripping ritual of symbolic content, and found anxiety reduction that was real but noisy, with confidence intervals crossing zero, concentrated in people already high in cognitive anxiety. That hedged version is the honest current state.

Three citations that do not survive checking

Nearly every mortality finding here is observational

The sauna, religious attendance, volunteering, and social connection results are cohort studies, not trials. Each carries a version of the same confound: the people who sauna four times a week, attend services, volunteer, and maintain thick social networks are systematically different from those who don't, in ways no statistical adjustment fully captures. Volunteering illustrates it cleanly — the meta-analysis found RR 0.78 for mortality across cohort studies, and the same authors state plainly that the randomised trials do not confirm the wellbeing benefits.

Replication has been unkind to some favourites

Dunn's 2008 Science finding that spending money on others increases happiness is a staple of popular psychology. The 2020 registered replication ran three high-powered preregistered experiments totalling 7,861 participants: the first replicated, the second found nothing, and the third replicated with a small effect. The effect is probably real on average and considerably more fragile than its reputation.

VIII — What the pattern suggests

One observation stands out across everything above, and it is not the one the wellness industry would draw.

The practices with the strongest evidence and the largest experiential payoff are almost all social, synchronous, and obligatory — and those are precisely the ones that cannot be bought, downloaded, or done alone at 6am. Meanwhile the enormous commercial apparatus of contemporary wellness is concentrated almost entirely on the individual, purchasable, and optional: the supplement, the device, the tracked metric, the solo morning routine. That inversion is close to exact, and it is not a coincidence — the effective practices are hard to monetise because they require other people and recurring commitment.

The second pattern is that traditions repeatedly found real physiological optima without any correct theory of what they were doing. Rosary and mantra both landed on six breaths a minute. Fire-walking festivals produced measurable cardiac synchrony between people who love each other. Costly initiation produced verifiable group commitment. Divination functioned as a randomiser against cognitive fixation. None of these traditions knew about baroreflexes, endorphins, costly signalling, or status-quo bias. Cultural evolution found the answers without understanding them, which is worth respecting without becoming credulous about the explanatory stories attached.

This returns to the opening case. The honest position on tarot is neither that it works nor that it is nonsense, but that it is a real intervention carrying a false label — scheduled reflective attention, an ambiguity that invites projection, a randomiser that breaks fixation, and a device for hearing a hard thing without having to say it. All four survive the removal of the metaphysics. Nothing is lost except the part that was never doing the work.