halo effect bias — 1 fairly sure investigation

a man with very good teeth talked me into something twice, on two separate occasions, and i signed papers afterwards. i’m fairly sure my taxes will reflect both occasions, eventually, in writing, and not warmly. the halo effect bias is what made the teeth count as accounting credentials. nothing else was checked.

desk, half a coffee, the cursor blinking inside a spreadsheet i opened to look busy. carla is upstairs at a training session about something called “outcome alignment”. i have, near as i can tell, until 1:42pm before anyone needs me to pretend.

so the topic, before the morning runs out, is halo effect bias — the brain trick where one good quality on a person, or a thing, quietly contaminates every other quality you have not yet measured. you meet a man with a firm handshake and decide his sedan is reliable. you meet a woman with neat handwriting and decide her opinion on insurance is sound. you don’t decide consciously. that’s the trouble. the deciding is done before you arrive at the table. for the long version of the underlying machinery, the cluster pillar on confirmation bias is where i keep landing every time i try to think about this honestly.

halo effect bias is the cognitive trick where a single positive quality on a person or product quietly upgrades every other quality you haven’t actually measured yet. good teeth become good accounting. neat handwriting becomes sound advice. it bypasses argument, runs underneath your own attention, and leaves you holding signed papers you cannot easily explain.

A. SMILE. IS. NOT. A. RESUME.

i needed that on the page early because i keep forgetting it. a smile is, mechanically, two muscles working in coordination. a resume is forty separate decisions a person made, badly, over fifteen years. one of these things tells you almost nothing about the other. and yet — and here is the embarrassing part — i have, on multiple tuesdays, treated the smile as if it audited the resume in advance. it did not audit anything. it was a smile. i was the auditor. i was bad at it.

halo effect bias, brief

the textbook version, near as i could piece it together from skimming things i won’t link, goes like this. the brain, when it likes one thing about a person, becomes lazy about checking the other things. the one good thing — the teeth, the watch, the calm voice, the way they pronounce quinoa without flinching — becomes a halo, and the halo is wide. it covers their judgment on money. it covers their punctuality. it covers, on a bad day, their plumbing recommendations.

i’m fairly sure there is a study on this, possibly in a serious magazine, where researchers showed people two photographs of the same person — once smiling, once neutral — and the smiling one was rated as more honest, more competent, and, for reasons nobody explained, better at chess. the photograph did not play chess. the photograph was a photograph. but the brain, on contact with the smile, handed out the halo and never asked for it back.

and this matters because it is, structurally, the cousin of every other bias on the list. the brain takes one cheap signal and lets it stand in for ten expensive ones. the long-form on confirmation bias meaning covers the version where the cheap signal is your own existing belief; the halo version is the same trick, only the cheap signal arrives wearing somebody else’s face.

tom has a halo and a volvo

tom has a halo. i need to say this in honesty because i’ve been pretending otherwise for years. tom owns a house, drives a volvo, has two children, fills in forms with a real pen, and pronounces dental terms correctly. on the strength of these visible facts, my brain has decided, without my permission, that tom’s opinion on every other matter is also probably sound.

tom recommends a podcast. i listen, gravely. tom mentions a stretching routine. i bookmark it. tom uses a phrase like “in the medium term” and i nod, as if i have a medium term. i do not have a medium term. i have a tuesday and a wednesday, and a vague sense of friday. but the halo is doing the math for me. the volvo, the kids, the firm handshake — they all bleed into the recommendation. the recommendation, on its merits, may be poor. i have not checked. checking would feel rude.

this is the trick of halo effect bias in domestic form. you don’t realize you’ve stopped evaluating because the evaluation feels like it has already happened. it has not happened. it has been outsourced, silently, to a station wagon and a school district.

the halo i do not have

i, by contrast, do not have a halo. i checked. i rent. i drive nothing. my handwriting looks like an ekg. my sentences end in qualifiers. when i recommend a podcast, the listener pauses to ask whether i’m sure. when i mention a stretching routine, the listener is openly skeptical. this is, in the long run, more accurate. people are right to check my output. but the asymmetry is doing damage.

the asymmetry is this: tom’s bad recommendations are absorbed without resistance, and mine are resisted without absorption. we have, between us, the same hit rate. probably. i have not measured. measuring would require admitting that some of what i say is right, and that admission would void the structure i’ve spent fifteen years building, which is the structure of the man whose suggestions are politely overruled in advance.

this is, you’ll notice, my own private halo problem in the negative. i have constructed an anti-halo over my own head, on purpose, because it lets me lower the stakes on every claim i make. the broader pillar already covers the version where i flatter myself; this is the inverse, where i pre-emptively ungrade myself so the world doesn’t have to.

the seventh microwave does not have one either

halo effect bias does not just attach to people. it attaches to objects. and i know this because i have, by an honest count, killed seven microwaves, and each one of them, in the showroom, had a halo.

the halo on a microwave is usually the door. the door is heavy, the door clicks, the door has a smoked-glass panel and a small light inside. the brain, on contact with a clicking door, decides the inside of the appliance is also sturdy, also competent, also unlikely to combust on contact with a metal object. the brain is wrong. the brain has only met the door. the rest of the appliance is, factually, a mystery box of capacitors. the door is a salesman. the salesman is good.

i bought the seventh microwave on the strength of a quiet rotating turntable in the showroom. the microwave plate doesn’t need to spin, i’ve maintained for years, but it spinning quietly was, apparently, enough to make me believe the rest of the machine was also serene. it was not serene. it was, within nine months, a paperweight with a mild burning smell. that was halo effect bias in white plastic. that was me, in a bigbox store, trusting a turntable.

the third yoga mat tried

the third yoga mat had a halo too, briefly. the halo was a soft texture in the rolled-up display. i touched it once, in the store, and i decided, against fifteen years of evidence, that i was going to become a person who unrolls things. i bought the mat. i carried it home, with intent. it has not been unrolled since 2023, possibly evolving under the couch.

the halo on the mat was the surface. the rest of the mat, factually, is a tube of foam. the surface contained no information about whether i would actually do yoga. the surface was a five-second tactile encounter. i extrapolated the rest of my life from it, in the aisle, in approximately eleven seconds. that is not analysis. that is shopping with a halo. and the mat, like the seventh microwave, is now a piece of evidence in the case against my own decision-making, which the longer post on the same family of patterns describes more politely than i can.

mike has no halo and i went to the bar anyway

i went to the corner on a wednesday to argue about all of this with mike. mike has, in the strict accounting of these things, the lowest halo of any person i know. mike wears the same flannel three days running. mike’s handshake is fine but unmemorable. mike pronounces quinoa “qween-oh-ah” on purpose. mike, by every visible signal a brain uses to assign halos, is not a man whose advice should land.

and yet. mike has a system for taxes. mike has not filed since 2019. on the topic of the halo effect, however, mike was unusually crisp. mike said, without looking up from catch me if you can playing muted on the bar tv: “the kid signed on the man’s smile. that’s the whole movie. that’s the whole bias. you’re describing a leonardo dicaprio sequence and pretending it’s an article.”

this was, i admit, more useful than i was prepared for. the kid in that movie talks his way through three professions on the strength of a clean haircut and a confident posture. the people he scams are not stupid. they are running a halo. they see the haircut, they see the suit, they see the calm, and they backfill the rest of the resume from there. the man with very good teeth, in my own little story, is the same trick at smaller scale. fewer airplanes. more signed papers. same mechanism.

mike is, in his way, the perfect counterweight. mike has no halo, and so to take advice from mike i have to listen to the actual sentence, on its actual merits, with no decorative bonus. i can’t borrow conviction from a volvo. there is no volvo. there is a bar stool, a flannel, and a man who has not paid the IRS in six years. the words have to do their own work. and they do, sometimes. that’s the only meaningful test i’ve found for whether a piece of advice is worth keeping: strip the halo, see if the words still walk.

verdict, the halo is rented, like everything else

carla cruised past on her way back from the third floor. cursor returned to the spreadsheet, eyes returned to the spreadsheet, posture corrected. she did not stop. she may have noticed. carla, by any honest accounting, has a halo and i still don’t know what she actually thinks of any of this.