The Costume Decision Is Mostly Logistics
Most of the decision about what to be for Halloween has nothing to do with your personality. It is settled by four things you already know: how many days you have, what you are prepared to spend, where you will be standing at nine o’clock, and who will be standing next to you. A quiz that asks for your favourite sweet and hands you a costume is answering a question nobody asked. The quiz above asks the four real ones first, lets two of them rule things out, and only then asks what you would actually like to wear.
You are also, statistically, right on time. In the National Retail Federation’s 2025 survey of 8,045 consumers, 73% planned to celebrate, 49% had started shopping in September or earlier, and the most common place people said they looked for costume ideas was an online search, at 37%, ahead of shops at 27% and friends at 21%. Total spending was forecast at a record $13.1 billion, with $4.3 billion of it on costumes. That is the scale of the decision you are making with a cup of tea.
Myth 1: The Best Costume Is the Most Original One
The adult top five in NRF’s survey has been the same five costumes for two years running, with only the order moving. In 2025: witch, 5.6 million adults; vampire, 2.7 million; pirate, cat and Batman, 1.6 million each. In 2024 it was witch at 5.8 million, vampire at 3 million, then cat, Batman and pirate. NRF’s own write-up of the 2025 list says it plainly: the most popular costumes stay pretty stable from year to year.
It is worth asking why, because the answer is not that millions of people lack imagination. The witch is recognisable from ten metres in a dark garden. It is made from clothes most wardrobes already contain. It goes under a coat. It leaves the face free to talk, eat and see. It costs nothing and survives a whole evening. In other words, it solves every constraint in the first half of the quiz at once, and it has been doing so for long enough that it has become the definition of the holiday rather than a costume worn at it. Originality is a taste, and the quiz asks for it directly: whether three other witches at the party would spoil your night or make it. If the honest answer is that you would not care, the classic is not a failure of nerve. It is the correct answer. And if the witch is the one, the witch type quiz will tell you which of ten kinds you would be, which is a better costume brief than “witch” on its own.
Myth 2: Going Topical Makes You Stand Out
Google has published a Halloween search tracker called Frightgeist for more than a decade. Its 2025 list, published in October, was dominated by a single film: the five most-searched costumes in the United States were Rumi, Zoey, Mira, Jinu and Baby Saja, all from Netflix’s KPop Demon Hunters. Sixth was Chicken Jockey from the Minecraft film, seventh Labubu, then Derpy the tiger, Elphaba and the Lorax. Labubu was also the top dog costume and the second most-searched children’s costume.
A trending costume is, by definition, the one the most people are searching for at the same time. It is the maximum-duplication choice, and the list that tells you it is trending is published in October, weeks before the night, when everyone else can read it too. That is not an argument against it. Being one of eleven Labubus at a party is a specific kind of fun, and the quiz has a type for people who want exactly that. It is an argument against choosing it by accident while believing it will make you distinctive. The duplication badge on each idea exists for that reason, and its rule is simple enough to check: high if the costume is on NRF’s adult top five or Frightgeist’s top ten for 2025, medium if it is on another 2025 list or is a long-running perennial, low if it is on none of them. The 2026 lists will not exist until late September and October; the perennials will not have moved.
Myth 3: More Effort Means a Better Night
Effort is rewarded only where it can be seen and only for as long as it can be worn. The failure modes of the ambitious costume are always the same ones: a mask that cannot be seen out of, a build that cannot sit down, paint that cannot survive a drink, a silhouette that does not fit through a door. The Consumer Product Safety Commission’s costume guidance is written for children but it is the right checklist for anyone: eye holes large enough for full vision and a secure fit if there is a mask, cosmetics rather than a loose mask where possible, a Flame Resistant label on the costume, the wig and the beard, nothing flimsy or billowing near a candle, and a hem you cannot trip on.
The visibility point has numbers behind it. A study in JAMA Pediatrics took 42 years of United States crash data, 1975 to 2016, and compared the evening of October 31 with the same evenings one week before and one week after. There were 608 pedestrian deaths on the 42 Halloween evenings against 851 on twice as many control evenings: a 43% higher risk, 2.07 deaths an hour against 1.45, peaking around six in the evening as the light goes. For children aged four to eight the risk was ten times higher. The authors’ suggestions include reflective patches on clothing. If your night involves a street, that single fact should outrank any question of taste, and it is why the plan on your result puts visibility before anything else when you answer “outdoors on foot”.
There is a smaller, stranger piece of evidence about masks that costume quizzes never mention. In 1976 a team of psychologists watched 1,352 trick-or-treating children at 27 homes in Seattle, where a bowl of sweets sat by the door with an instruction to take one. Some children were asked their names; some were not. Some arrived alone; some in groups. The children who were anonymous and in a group took extra sweets far more often than the identified children on their own, and the effect has been a standard example of deindividuation ever since. A costume that hides your face changes how you behave, not only how you look. That is either the best reason to wear one or the best reason not to, depending on the night, which is why the quiz asks how you feel about covering your face before it recommends anything with a mask.
Myth 4: Costumes Are Mostly for Children
The money says otherwise and has for some time. NRF’s 2025 forecast put spending on adult costumes at $2 billion against $1.4 billion for children’s and $0.86 billion for pets, and its 2024 figures were $1.8 billion and $1.3 billion. Costuming at American Halloween parties was, according to the historical record, as often for adults as for children in the early twentieth century; the idea that it is a children’s holiday that adults are gatecrashing is the later invention.
The industry knew this early. The largest American costume maker for much of the twentieth century, Ben Cooper, Inc. of Brooklyn, founded in 1937, built its business on licensed characters in a cheap plastic mask and a vinyl smock. It licensed Spider-Man in 1963, when the character was virtually unknown, and Batman in 1964. In 1979 it produced a costume of the creature from Alien, the first based on an R-rated film, a product with no child in mind at all. By 1991 the company claimed 70 to 80 percent of the licensed character costume business; it filed for bankruptcy twice, in 1988 and 1991, and was bought by Rubie’s in 1992. The children’s costume of the mid-century, a mask and a smock, was also the form the adult costume took for decades, which is part of why so many adults still assume a costume has to cover the face. It does not, and the survey data says the most-worn adult costumes mostly do not.
Myth 5: Costume Contact Lenses Are Just Make-Up
They are medical devices, and in the United States they are the one costume accessory it is illegal to sell without a prescription. The Food and Drug Administration’s page on decorative lenses states that all contact lenses must be prescribed by a doctor, including decorative ones, that any lens sold without a prescription is illegal, and that the risks of wearing them include corneal abrasion, allergic reaction, decreased vision, infection and blindness. It names the places not to buy them from: a street vendor, a beauty supply store, a flea market, a novelty store or a Halloween store. That covers most of the places people buy them.
The reason this belongs in a costume quiz rather than a footnote is that the lenses are the finishing touch on exactly the costumes the quiz recommends to people who answered “indoors, mild, an hour or two, paint not a mask”: the vampire, the icon, the zombie. A lens fitted by an optometrist is a fine accessory. One from the shop that sold you the cape is a gamble with a cornea, for an effect that is invisible from more than a metre away in party lighting. The other label worth reading is the one the CPSC asks for: Flame Resistant, on the costume, the wig and the beard, because the costumes that catch on candles are the flimsy, billowing ones, and those are the costumes that look best in photographs.
Why the Quiz Gives You Two Answers
The first eight questions are logistics: lead time, budget, venue, weather, company, how long you will wear it, how you feel about covering your face, and whether you make things or buy them. Every answer carries a fit score for each of the eight costume types, and two answers are hard constraints. Saying you need the costume within a day rules out The Build, because the quiz will not recommend a costume you cannot finish, and saying you are going alone rules out The Set, because a matching costume needs someone to match with. The interlude names anything ruled out and why, and the Back button undoes it.
The last eight questions are taste, and they never mention a constraint. They are laid out so that each type is the strongest answer in exactly four questions and the second answer in exactly four others, and no two types are the strongest answer together more than twice, so no type can win by appearing more often. Every type’s total is then measured against what random answering would give that type, because the logistics scores are deliberately uneven and would otherwise favour whichever type collected the most positive numbers. The result is the best fit on both halves together, among the types not ruled out. When two finish level, the tie goes to the type your answers picked outright more often, and the result says so.
The panel also shows what each half says on its own, because the disagreement is often the useful part. If your taste says The Build and your night says The Classic, the honest reading is not that you are a classic person. It is that you want to make something and have not given yourself the weeks to do it, which is a scheduling problem with a known fix. What the page will not do is tell you how rare your type is. A costume strategy is not rare or common. What is rare is being the only Labubu in the room, and that is what the duplication badge estimates, by a stated rule, from published lists rather than from anything measured at a party. If you want the research on why the costume changes the wearer and not just the room, the aesthetic quiz covers the enclothed cognition studies, including the mixed record on replicating them.
The Eight Costume Types, With the Ideas Behind Each
The Classic is for a night with little time, little money or a lot of walking, and a taste that wants recognition without a speech: witch, vampire, black cat and pirate, all in NRF’s 2025 adult top five, plus the ghost from the children’s list. Its one watch-out is that black disappears on a dark street, so it takes reflective tape or a light-coloured prop outdoors.
The Pun is for more wit than lead time and a taste for the click: cereal killer, Freudian slip, identity theft, ceiling fan, smarty pants. None of them is on any 2025 list, which is the point. It dies in a loud bar where nobody can hear the explanation, and it needs one test on a stranger before the night.
The Build is for two weeks or more, a real budget and a glue gun: the plague doctor, a foam-armour knight, an astronaut, and the two 2025 costumes that are genuinely a build, Chicken Jockey and a full-paint Elphaba. It is ruled out below two days and it needs a full test wear, sitting down and drinking included, a week before.
The Scare is for house parties and bar crawls with a face you are willing to cover: Ghostface, Michael Myers, a zombie, Pennywise, Art the Clown, all franchise perennials rather than 2025 trends. Full masks cut peripheral vision and heat up, and come off to cross a road. If the monster matters more to you than the mask, the mythical creature quiz sorts the folklore rather than the film.
The Icon is for indoors, mild weather and the right kind of attention: Barbie, Morticia Addams, Marilyn Monroe, Freddie Mercury at Live Aid, Audrey Hepburn. It is the type most likely to reach for coloured lenses, which is why the FDA’s rule is on its card. Cold doorsteps are where it suffers.
The Set is for anyone with someone to match: Elphaba and Glinda and Astrid and Hiccup, the top duos of 2025; the HUNTR/X trio, the top three costumes of 2025 in one go; and the perennial Ghostbusters and Inside Out emotions. It is ruled out when you are going alone, and it needs a plan for the person who drops out.
The Comfort is for long, cold nights on foot: the inflatable T. rex, an animal onesie, The Dude, Where’s Wally, and a pumpkin borrowed from the top of the pet list. Inflatables cannot sit and need spare batteries; a light-coloured onesie is visible in headlights without any help.
The Trend is for people who want to be part of the year’s moment on purpose: Rumi, Labubu, Chicken Jockey, Elphaba and the Lorax from Frightgeist’s 2025 top ten, every one of them high duplication by definition. The 2026 list arrives in October, which is also when the popular ones sell out. Princess was the second most-planned children’s costume of 2025; if that is the brief, the Disney princess quiz narrows it to one. And for the rest of the year, when the question is what to wear rather than who to be, the style quiz is the version of this exercise with no hat in it.
The First Costumes Were Made to Hide You
Guising, going from house to house in disguise for food or money, is recorded in Scotland and Ireland from at least the sixteenth century. A description from Scotland in 1895 has masqueraders carrying lanterns made from scooped-out turnips, visiting homes to be rewarded with cakes, fruit and money; an account in Scots has the boys running about with their fause-faces on and their turnip lanterns in hand. Folklorists’ best guess at the original purpose is that the disguised visitors were impersonating the aos sí, or the souls of the dead, and collecting offerings on their behalf, and that the disguise also protected the wearer from the things it imitated. The custom crossed the Atlantic with Scottish and Irish emigrants: children were reported “guising” in Kingston, Ontario in 1911, disguised with all kinds of masks and costumes; the phrase “tricks or treats” was first printed in Sault Ste. Marie in 1917; and by the 1930s the practice was common enough in the United States for A. S. Fishbach, Ben Cooper and others to start mass-producing costumes for it.
The point of the original costume was to be unrecognisable. The false face was the costume; the whole transaction depended on the householder not knowing who was at the door. The modern costume has exactly the opposite job. Nearly every question in the second half of the quiz is, one way or another, about being recognised: whether you want the click, the flinch, the compliment, the “are you all a set?”. The most-worn adult costume in America is a witch precisely because nobody has to ask. Somewhere between the turnip lantern and the 5.6 million witches, the costume stopped being a way of hiding and became a way of being seen, and the only costume type on this page that still does the original job is the one behind a full mask. If the quiz sent you there, you are, in the oldest sense of the word, guising. Everyone else is dressing up, which is a newer and, on the evidence above, a far more popular thing to do.
