The evidence

What a made object can honestly be built on, what the studies actually found, and which of them did not survive a second look.

Everything this product does rests on four ordinary things — attention, intention, ritual and meaning — and on nothing else. None of them require you to believe in magic, and the strongest finding on this page is that being told plainly how something works does not stop it working. What follows is the case for that, the case against the parts of it that collapsed, and the numbers, so you can weigh it yourself rather than take our word.

What does the work

Each of these has survived replication or was built from many studies at once.

  1. Naming the moment

    An if-then plan — "when this happens, I will do that" — ties an action to a cue you will actually meet. Across 94 independent tests and more than 8,000 people, it produced a medium-to-large improvement in follow-through (d ≈ .65). This is the cleanest, best-evidenced and least magical route by which an object changes what somebody does. We do not write the plan for you: the forge asks you to name your intention in one line, and the emblem is cut from that line and put where you will meet it.

    Gollwitzer & Sheeran, 2006 — Advances in Experimental Social Psychology

  2. Keeping it in view

    Across 138 studies, monitoring progress toward a goal made reaching it measurably more likely — most strongly where the progress was something you could see. An emblem you meet on a lock screen is an intention kept in view, forty to a hundred times a day, without ever going looking for it.

    Harkin et al., 2016 — Psychological Bulletin

  3. Ritual, measured in the body

    Seventy-five people in Mauritius were put through a public-speaking stress test, then either performed a habitual temple ritual or sat and relaxed. The ritual group reported less anxiety and showed it physiologically, in higher heart-rate variability. This strand is independent of the retracted work below, and is corroborated across field, lab and physiological methods — though the samples are small.

    Lang, Krátký & Xygalatas, 2020 — Phil. Trans. R. Soc. B

  4. Honesty does not break it

    Eighty people with irritable bowel syndrome were given a placebo and told, in plain words, that it was an inert sugar pill. They improved anyway, significantly more than the untreated group. A later meta-analysis of non-deceptive placebos across five conditions found a positive pooled effect. This is the single most important finding on this page: it is the reason this product can tell you exactly how it works instead of selling you a spell.

    Kaptchuk et al., 2010 — PLOS ONECharlesworth et al., 2017 — J. Evidence-Based Medicine

  5. Standing for something you value

    Affirming a value you hold buffers stress in a completely unrelated domain. It has been measured in the body — people who wrote about a value they held before a stress test produced less cortisol — and again in the head, where the same exercise improved problem-solving specifically in people already under chronic stress. An object inscribed with a value you chose is an affirmation cue you carry.

    Creswell et al., 2005 — Psychological ScienceCreswell et al., 2013 — PLOS ONE

  6. Things become part of who you are

    The "extended self" is why a ring, a jacket or a father's watch is not interchangeable with an identical copy: possessions get folded into identity. It is also why an object singled out, named and handled is worth more to its owner than one off a shelf.

    Belk, 1988 — Journal of Consumer Research

What did not hold

This field was hit hard by the replication crisis, and the most famous findings in it are among the casualties. Leaving them out would be quoting the literature the way the literature is criticised for being quoted.

  1. The lucky ball that improved putting

    The most-cited finding in this whole area: golfers told their ball was lucky sank 65% of putts against a control group's 48%. Two high-powered pre-registered replications found the effect "elusive" — the lucky-ball participants performed almost identically to controls, and no effect at all survived in the strongest designs. A luck-improves-performance claim is not available to anyone, and we do not make one.

    Damisch, Stoberock & Mussweiler, 2010 — Psychological ScienceCalin-Jageman & Caldwell, 2014 — the replication

  2. The ritual paper that was retracted

    "Don't Stop Believing: Rituals Improve Performance by Decreasing Anxiety" reported that pre-performance rituals cut anxiety and improved singing and maths. It was retracted in November 2024 at the authors' own request, after an audit left them without confidence in the integrity of the data. It is the reason the ritual claim above is sourced to fieldwork and heart-rate variability instead.

    Brooks et al., 2016 — the paper, marked retractedThe retraction notice, November 2024

  3. And its neighbour, flagged

    The 2014 grief-and-ritual paper — rituals after a loss reduced grief by restoring a sense of control, even in people who did not believe in rituals — has not been retracted, and its finding is one we would have liked to use. One of its authors is at the centre of a data-fraud investigation. It is cited here so you can weigh it yourself, and it is load-bearing nowhere in this product.

    Norton & Gino, 2014 — J. Experimental Psychology: General

  4. Death, and defending your worldview

    Terror management theory — that awareness of mortality drives us toward meaning, symbols and belonging — is the most sweeping explanation on offer for why people keep sacred objects, and a 2010 meta-analysis of 277 experiments put it at a moderate effect. Its signature paradigm then failed to replicate in Many Labs 4, with the original authors involved, and a 2022 pre-registered meta-analysis shrank it sharply — to non-significance in places — once publication bias was corrected. An influential framework; not a foundation.

    Burke, Martens & Faucher, 2010 — the meta-analysisKlein et al., 2022 — Many Labs 4Schindler et al., 2022 — Personality & Social Psychology Review

  5. The lab coat

    Wearing a coat described as a doctor's coat improved sustained attention, but only when it carried that meaning and was physically worn — the study that named "enclothed cognition". Replication has been mixed, and virtual versions failed. Suggestive about symbol and embodiment; too thin to build on.

    Adam & Galinsky, 2012 — J. Experimental Social Psychology

Why anyone reaches for an object

Not a failure of reasoning. A set of ordinary, universal habits of mind.

  1. Losing control is the switch

    People whose sense of control was experimentally reduced saw images in noise, found correlations that were not in the data, and adopted superstitions. Malinowski watched the same thing in the Trobriand Islands a century ago: no magic around safe lagoon fishing, extensive ritual around the dangerous open sea. Charms cluster where competence runs out — exams, surgery, weather, the first night of something.

    Whitson & Galinsky, 2008 — ScienceMalinowski, 1948 — Magic, Science and Religion

  2. Knowing better does not stop it

    The standard model of thinking assumes that when the slow mind catches the fast mind in an error, it corrects it. It often does not — it detects the error and goes along with it anyway. This is why superstition tracks neither intelligence nor education, and why "I know it does nothing, but I will do it anyway" is a coherent sentence rather than a confession.

    Risen, 2016 — Psychological Review

  3. Contact and likeness

    Educated modern adults reliably refuse juice a sterilised cockroach has touched, and are reluctant to wear a laundered sweater once worn by someone they consider evil. The same two intuitions — what touches carries, what resembles is — run positively through relics, saints' medals and a grandmother's ring.

    Rozin, Millman & Nemeroff, 1986 — J. Personality and Social Psychology

  4. And the world must look orderly

    When personal control is threatened, people shore it up by affirming some external structure that keeps things in order — fate, government, a watching God, a charm. The direction of this is well supported; a registered replication across two countries found one of its best-known versions weaker and less reliable than the original, so treat the size with care.

    Landau, Kay & Whitson, 2015 — Psychological BulletinHoogeveen et al., 2019 — a registered replication across two countries

How common this is

India
49% say lives can be influenced through the evil eye; Hindus and Muslims alike at 51%.

Pew Research Center, 2021 — Religion in India

Central & Eastern Europe
A median of 48%, highest in Greece and Latvia (66% each), then Ukraine, Armenia, Moldova, Russia and Bulgaria.

Pew Research Center, 2017 — Central and Eastern Europe

Latin America
At least a third in every country surveyed; more than half in Panama, Venezuela, the Dominican Republic, Honduras, Chile and Brazil.

Pew Research Center, 2014 — Religion in Latin America

Sub-Saharan Africa
A quarter or more in eleven countries believe in the protective power of charms, shrines and sacred objects — 75% in Senegal, 5% in Rwanda.

Pew Research Center, 2010 — Islam and Christianity in Sub-Saharan Africa

Turkey & Kazakhstan
Clear majorities say the evil eye is real (69% and 66%); wearing a talisman for protection is most common in Pakistan (41%) and Albania (39%).

Pew Research Center, 2012 — The World's Muslims

The United States
About a quarter say objects such as crystals or stones can carry spiritual energies — 42% among the spiritual but not religious.

Pew Research Center, 2023 — Spirituality Among Americans

And carrying one
24% of Americans carry a lucky charm at least occasionally — and it is essentially flat across income, from 20% above $80k to 23% below $39k. This is not a poverty or an education story.

YouGov Omnibus poll, 2018 — no stable public link

The traditions

Every one of these is older than the research, and most of them do the same four things.

Omamori
Japan — a sealed brocade pouch, never opened, renewed after about a year and returned to the shrine to be burnt.
Nazar boncuğu
Turkey, Greece, the Levant — blue glass cut as an eye, to take a gaze meant for you.
Hamsa and mezuzah
A hand at the throat, an inscription at the doorpost.
Saints’ medals, scapulars, relics
Catholic Europe — the logic of contact, worn.
Milagros and azabache
Latin America — a small silver limb for the thing that needs mending; jet pinned to an infant.
Juju and Adinkra
West Africa — charms, and a vocabulary of stamped symbols each carrying a proverb.
Yantra and rudraksha
India — geometry drawn as a seat for attention; seeds strung and counted.
Oberegs
The Slavic world — woven, stitched and hung at the threshold.
Thai amulets
A serious market, priced in part by which monk blessed the piece.

The omamori is the one we learnt a rule from. It expires: it is carried for about a year, then returned and burnt, and a new one taken. An expiry is not a limitation dressed up — it is what stops a charm becoming furniture, and it puts the bearer back in front of the question of what they want now. That is why an emblem here has a term.

A picture in a phone

There is one finding that argues against this product as built, and it belongs here rather than in a drawer. Digital possessions can become part of the extended self — but probably less strongly than material ones, because something infinitely copyable is hard to hold as singular, and the sense of an object having absorbed something from being carried is weaker. We think that is right, and it is why nothing here is a picture you download. The emblem is cut from answers only you gave, under a sky that does not come round again; it is named, awakened by a spell you say once and that then freezes as you said it, and it is met where you actually look. Whether that closes the gap is an open question, and an honest page says so.

Belk, 2013 — Journal of Consumer Research

What we will never claim

No talisman here brings luck, changes an outcome, protects you from harm, improves your health, or makes you perform better. No claim about energy, vibration or influence at a distance. Those things are either unsupported or, in the case of health, capable of real harm through delayed care and false confidence. The honest description is the one at the top of this page, and it is the only one this product is allowed to make — of itself, and in every word it writes for you.

Where this page is weak

Begin a forging →