attraction.lab

What Is the Ick? One Study, Three Different Sample Sizes

Exactly one empirical study of the dating ick exists. Its abstract says 138 people, every report of it says 125, and it missed its own power target.

Research-backed writing on attraction, dating and relationships. Drafted by AI against a named bibliography, published only when every claim links to its source.
33 min read
Editorial illustration for: What Is the Ick? One Study, Three Different Sample Sizes

Before the article. If the aversion you are trying to understand arrived after someone pushed past a no, pressured you, went through your phone, or made you justify your time, that is not the thing this research is about, and the first section says so in more detail. Find A Helpline, in the sources at the bottom, lists free and confidential services by country. In the United States the National Domestic Violence Hotline is 1-800-799-7233, or text START to 88788, 24 hours a day.

With that said. The ick — a sudden collapse of attraction set off by something small and specific a person did — has a dictionary entry, a Cleveland Clinic explainer, and roughly a decade of coverage. It has exactly one empirical study behind it, and this page is mostly an account of what happened when that study was checked.

Three things came out of the checking. The study's own abstract gives a different sample size, 138, from the 125 that every report of it uses. It fell nine participants short of the number its own preregistered power analysis called for, which is a detail no page on the first page of results mentions. And although three separate databases label it open access, it could not be read: the publisher returns 403 to a plain request and an error page to a real browser. So no effect size from it appears here, and every figure is attributed to whichever source carries it.

What else is on this page: where the word came from, including the one-year error in the date given by the top-ranking result; a count of the literature in three databases, run on the day of writing; a dissection of the 51-person experiment that the popular disgust explanation rests on; and the one distinction worth taking away, which is between an ick that carries compressed information about a person and an ick that reports the height of the standard being applied. The evidence cannot tell you which yours is, and this page does not pretend to.

It contains no list of behaviours to avoid, because such a list is both unfalsifiable and useless. Our editorial policy explains why every empirical claim here carries a linked source.

#1

Before the rest of this page: the version of this that is not about taste

Most people looking this up are looking it up in good faith and usually about themselves — whether the feeling means something, whether it makes them shallow, whether they have just thrown away something that was fine. Almost all of what follows is about that. One thing has to go first, though, because it gets filed under the same word and it is not the same thing.

If what you feel is revulsion that arrived after someone pushed past a no, pressured you about sex, went through your phone, turned up somewhere you had not invited them, or made you account for your time, that is not a quirk-triggered aversion and no amount of reframing should talk you out of it. The research below is about sudden distaste at trivial cues. It has nothing to say about aversion that follows a boundary being crossed, and the honest reading of that feeling is usually the literal one.

The same goes for a reaction that does not look like distaste from the inside so much as dread, or numbness, or a wish to be anywhere else. If the relationship involves being monitored, isolated from other people, or frightened, that is the subject of a different page and a different kind of help. Find A Helpline, in the sources at the bottom, lists free and confidential services by country. In the United States the National Domestic Violence Hotline is 1-800-799-7233, or text START to 88788, 24 hours a day. And if the feeling is distressing enough that it is driving the decision rather than informing it — if every relationship ends this way and you cannot tell why — that is a reasonable thing to take to a therapist, who can do something with it that an article cannot.

With that said. What follows is what the evidence on the ordinary version actually consists of, which turns out to be one study, and what happened when this site tried to check it.

#2

What the word means, and what it is not

The ick is a sudden drop from attraction into something closer to distaste, set off by a specific and usually trivial thing a person did — not a gradual cooling, and not a conclusion reached by thinking about it. The two features people consistently report are the speed of it and the disproportion: the trigger is small, the reaction is not, and afterwards the attraction is difficult to recover.

Two things it is not. It is not a clinical term. It does not appear in any diagnostic manual, there is no instrument that measures it, and nobody has published a validated scale for it, which matters for everything below: when a study measures the ick, it measures whatever it asked people on a questionnaire, and different questions would give different answers.

And it is not a verdict on the other person. The single study on the subject is explicit that its authors cannot tell the two possibilities apart — an ick that flags a real incompatibility, and an ick that flags nothing except the standard being applied. That distinction runs through the rest of this page and it is the one the genre usually skips, because skipping it is what allows a page to be written as a list of things not to do.

This page is not that list. A catalogue of behaviours that have repelled somebody is both unfalsifiable and useless: the pilot data below found triggers as specific as the way a person ran for a bus, and there is no version of that list that amounts to advice. What can be said is what was measured, by whom, in how many people, and how well — and that is a shorter story than the volume of writing on this subject suggests.

#3

Where the word came from, and the date the top result gets wrong

The origin is usually told in one line, and the two most-read versions of that line disagree by a year. It is worth getting right, because it is the one claim on this subject that is cheap to verify and still gets published wrong.

Dictionary.com, the first result for this query, attributes the dating sense to "Once in a Lifetime," which it dates to 1999, an episode of Ally McBeal in which the title character uses "the ick" for an attraction that has inexplicably died on her. Wikipedia's article on the ick says the first documented use was on an Ally McBeal episode that aired in 1998.

The episode list settles it. "Once in a Lifetime" is season 1, episode 15, directed by Elodie Keene, and it aired on 23 February 1998. Dictionary.com has the episode right and the year wrong by one.

The rest of the timeline is better attested. The term sat mostly unused in this sense for over a decade; Dictionary.com puts its revival in the later 2010s, and Wikipedia credits its popularisation to Olivia Attwood, a contestant on the 2017 series of Love Island, with the TikTok wave following from around 2020. The institutional marker is sharper than either: the Oxford English Dictionary published an entry for ick as a noun and interjection on 2 March 2023, which is roughly when a piece of slang stops being slang.

So the word is about twenty-eight years old in this sense, went mainstream around nine years ago, and acquired a dictionary entry three and a half years ago. The first study of it is younger than the dictionary entry, which is the next section.

#4

The first page of results, opened on 5 November 2026

This site audits the pages it competes with rather than guessing at them, so here is the measurement, with the method stated so it can be re-run.

Google returned its unusual-traffic challenge to the headless browser again today, so the result set comes from this session's search tool rather than from a scrape of the results page, and the "People also ask" block could not be collected. Ten results for what is the ick. Nine downloaded cleanly; phys.org returned HTTP 403 and was not read. Each of the nine was then searched, as text, for the authors of the only study on the subject, for its journal, and for a sample size.

The result:

  • Seven of the nine do not mention the study at all. Not the authors, not the journal, not a number. That includes Dictionary.com, PureWow, Cleveland Clinic — 2,149 words of explanation from a named psychologist, with no citation in it — Wikipedia, the Institute for Family Studies, a psychotherapy practice blog, and a grammar site.
  • Two name it. Psychology Today writes it up at 2,109 words, names Collisson and colleagues, names the journal and links the DOI. The BPS Research Digest does the same at 1,191 words.
  • One of the nine gives the sample size. The BPS Digest. Psychology Today describes the study in detail, reports percentages from it, and never says how many people were in it.
  • None of the nine — zero — mentions preregistration or the study's power analysis. That is where this page found its number, and it is in the next section but one.

The pattern is familiar from other topics measured here: the strongest page cites correctly and the rest paraphrase. What is unusual is how thin the citing layer is. On a term with a dictionary entry, a Cleveland Clinic explainer and a decade of coverage, the number of first-page pages that tell you how many people have ever been studied is one.

#5

The literature, counted on 5 November 2026

Before any argument about what the evidence shows, here is how much evidence there is. The queries are written out so the count can be re-run rather than taken on trust.

Europe PMC, title search for `ick`: 47 records, and not one of them is about the dating sense. They are the ICK/CILK1 kinase gene, inhibitor-cystine-knot peptides from spider and sea-anemone venom, the "ick factor" in organ donation and in chemotherapy avoidance, and a study on whether a robot speaking Berlin dialect is judged more trustworthy. Searching the same database for the one relevant paper by its DOI returns zero records: the only empirical study of the ick is not indexed in Europe PMC, and so not in PubMed either.

PubMed's own interface could not be queried today at all — every request to its search API returned HTTP 500, from two different servers, which is noted here rather than quietly dropped.

OpenAlex, title search for `ick`: 357 works, dominated by the same gene, by Pick's disease, and by the "ick factor" in the literature on eating insects. The dating study is in there, and as far as these three databases are concerned it is alone.

Then the citation count, which is itself three different numbers. Twenty-one months after publication, Crossref reports 3 works citing it, Semantic Scholar reports 1, and OpenAlex's own citation filter returns 0. Crossref counts 20 references in the paper; Semantic Scholar counts 17.

The honest summary: one study, no replication, and a literature so small that the databases disagree about whether anyone has cited it. That is not a reason to dismiss the phenomenon — people plainly experience it — but it is the reason every confident mechanism you will read about it is an extrapolation. The site's page on the five love languages describes the opposite situation, a popular idea with thirty years of tests behind it, and the contrast is instructive: there, the evidence was large enough to convict. Here there is barely enough to indict.

#6

The one study, and what it actually did

Brian Collisson, Eliana Saunders and Chloe Yin, all at Azusa Pacific University, published "The ick: Disgust sensitivity, narcissism, and perfectionism in mate choice thresholds" in Personality and Individual Differences, volume 238, article 113086. It went online on 5 February 2025 and into the May 2025 issue.

It has two parts.

The pilot was a content analysis of TikTok videos tagged #theick, coded to find out what people actually name when they name an ick. The categories that came out of it, as reported by both of the pages that cite the study, were things like physical appearance, being publicly embarrassing, speech and verbal habits, and conspicuously trend-chasing behaviour. The detail worth keeping from the pilot is that physical appearance was a less frequent trigger than behaviour or personality — which cuts against the usual framing of the ick as a looks-based reflex. Those categories then supplied the items for the main study.

The main study was a one-sitting online survey of single adults. Participants were asked whether they had met the term, then — given a definition — whether they had experienced it, how many times, and which of the piloted triggers applied to them; the trigger list shown was matched to the participant's gender, so women saw icks reported by women and men saw icks reported by men. They then completed measures of disgust sensitivity, other-oriented perfectionism (holding a partner to exacting standards) and grandiose narcissism.

That design is worth holding in mind before the findings. It is cross-sectional and retrospective: everything is one questionnaire, at one moment, with people recalling aversions from their own dating history and reporting their own traits. There is no partner in it, no behaviour observed, and no follow-up. It can find out which traits travel with reporting more icks. It cannot find out whether the trait caused the ick, whether the ick was a reasonable read, or what happened next — the same structural limit this site ran into on what the first few dates can tell you.

#7

How many people were in it depends on which source you read

This is where checking the study stopped being routine. Three different sample sizes are in circulation for the same piece of research, and the discrepancy is not in the tabloid coverage — it is in the paper's own metadata.

The abstract distributed with the article's record states, in as many words, that "a sample of single adults (N = 138)" reported their familiarity with and experiences of the ick.

Every outlet that appears to have read the paper reports 125. The BPS Research Digest: "the researchers recruited 125 single adults aged between 24 and 72." tutor2u: "125 single adults aged 24–72." Newsweek: "125 single participants." StudyFinds: "125 single adults." One outlet, UniLad, breaks it down as 74 men and 51 women, which sums to exactly 125 — so the figure is coming from a table, not from a press release.

And the paper has a third identity. The accepted version was posted to SSRN in 2024 under a different title: "The Ick Effect: Disgust, Narcissism, and Perfectionism in Romantic Aversion." Same authors, same DOI prefix for a preprint, different name.

The most likely explanation is dull and worth stating plainly: 138 recruited, 125 analysed after exclusions, with the abstract carrying the pre-exclusion figure. That would be ordinary. What is not ordinary is that the figure travelling with the article in every database that indexes it is the one no report of the study uses, and that this page cannot close the gap by looking — for the reason in the section after next.

So: treat 125 as the analysed sample, because that is what the people who read the tables say and the gender split corroborates it, and treat the 138 in the abstract as unexplained. Neither number is large.

#8

The number nobody on the first page gives: it missed its own target

Of the nine first-page pages read today, one reports the sample size and none reports this, which is the figure that determines what any of it means.

StudyFinds, covering the study in February 2025, notes that the authors ran a preregistered power analysis before collecting data, that the analysis called for 134 participants, and that the final sample of 125 fell short of it — adding, correctly, that the results should therefore be read with caution.

That single sentence reframes the whole topic. A power analysis is the calculation of how many people you need to have a decent chance of detecting an effect of the size you expect. Filing it in advance is good practice and the authors did it, which is to their credit and more than most of this genre manages. Falling short of it means the study was, by its own stated standard, underpowered. In an underpowered study the effects that do reach significance tend to be overestimates, and the effects that do not reach significance cannot be read as absences. Both halves of that matter here, because the findings in the next section are a mix of hits and misses, and in a sample like this the pattern of which was which is not stable information.

This is the layer under the number, and it is where this site has repeatedly found what the first page is missing: not the absence of a citation but the absence of the detail that tells you how much the citation can carry. The same question — how was this measured, and was it measured well enough to mean anything — is what took apart the thirty-item quiz behind the love languages and the famous bridge experiment discussed in first-date mistakes.

To be clear about proportion: a preregistered study that misses its target by nine people is a long way from a bad study. It is a small study that said so. The problem is downstream, where it becomes "psychologists reveal the top cause of the ick."

#9

What it found, and where the write-ups disagree

The descriptive findings are the most solid part, and they are genuinely interesting.

  • Around half of participants had already met the term, and the familiarity gap by gender is wide: 63% of the women against 39% of the men, per the breakdown UniLad reports.
  • Given a definition, 64% said they had experienced it. So roughly a third of single adults aged 24 to 72, handed the concept, said it had never happened to them.
  • Frequency is where it gets awkward. The BPS Digest reports an average of around ten times across a dating history, then adds that this drops to six once you exclude one person who reported 300. Psychology Today reports the average as approximately six, mentions the person who reported 300, and does not say that the six is the outlier-excluded figure. One study, one statistic, two incompatible presentations — and a mean that moves by 40% on the removal of a single respondent is telling you the distribution is long-tailed rather than giving you a typical experience.
  • Only 28% told the partner who caused it, per Psychology Today. Which means the ick is overwhelmingly resolved by the person who has it, silently, and usually by withdrawing — a mechanism this site has described from the other side in breadcrumbing and in what pocketing means.

Then the trait findings, and here the three available accounts do not agree with each other:

  • The abstract says disgust sensitivity correlated with cumulative ick experiences, and that narcissism and perfectionism were "positively associated with ick frequency."
  • Psychology Today says grandiose narcissism predicted whether someone got the ick but not how often, and that perfectionism predicted both.
  • The BPS Digest says higher disgust proneness made people more likely to experience it "though this didn't impact how often they felt it" — the reverse of what the abstract says about disgust.

All three were written from the same paper. The disagreement is about the distinction between likelihood and frequency, which is exactly the distinction an underpowered study is least able to resolve, and it cannot be adjudicated here without the tables. What survives all three readings: people higher in disgust sensitivity, grandiose narcissism and other-oriented perfectionism report more of this. Which of those predicts the having versus the counting is unsettled.

One finding is consistent and worth noting because it cuts against the stereotype: although women were far more familiar with the word, the frequency of the experience was comparable across genders. The ick is not a thing women do.

Further illustration for: What Is the Ick? One Study, Three Different Sample Sizes
#10

The gendered triggers, and why they are the weakest part

Both citing pages report the same secondary result, and it is the one most likely to end up as a headline, so it deserves its own warning.

Among the women, perfectionism and narcissism correlated most strongly with icks based on physical appearance and public embarrassment. Among the men, disgust sensitivity was linked to icks about speech habits and about conspicuously trend-chasing behaviour — a partner visibly trying to fit in.

Take that lightly, for three reasons that compound.

First, these are correlations between traits and trigger categories within gender subgroups, which means the 125 has already been cut to roughly 74 and 51 before the analysis starts. A study that missed its power target for the whole sample has considerably less to work with inside each half.

Second, the trigger lists were not the same for the two groups. Women were shown icks that women had reported in the TikTok pilot; men were shown icks that men had reported. That is a defensible design choice — showing people irrelevant items wastes the questionnaire — but it means a gender difference in which triggers correlate with which traits is partly a difference in what each group was asked about. The comparison is not like-for-like.

Third, this is the exact result that produced the coverage. One of the pages on the first page of results leads on the misogyny example from the pilot; a tabloid headline from the same week reads, in effect, that getting the ick may mean you are a narcissist. The study does not support that sentence about any individual. A modest correlation across 125 people says nothing diagnostic about one person, and the site's standing position on this — that the tests here are educational tools and not diagnoses — applies with more force to a trait correlation from an underpowered survey than it does to a validated questionnaire.

#11

Why "it's just disgust" is doing less work than it sounds

The mechanism offered everywhere is that the ick is disgust, and disgust is an evolved pathogen-avoidance system misfiring on a romantic prospect. It is a tidy story. It is also where the citation chain gets thin enough to see through.

Psychology Today supports the step from disgust to lost attraction with one reference: Phelan and Edlund, "How Disgust Affects Romantic Attraction," in Evolutionary Psychological Science. Unlike the ick paper, this one is free to read, so this site read it in full. Here is what it is.

Fifty-one participants, randomly assigned to three groups: a disgust induction (n = 18), a hygiene induction (n = 16), and a neutral condition (n = 17). Disgust was induced by showing people a series of disgusting photographs. They then rated nine target faces, each viewed twice, half of which had been digitally given a facial disfigurement.

The results. Disfigured faces were rated less attractive overall, F(1,45) = 20.876, p < .01, ηp² = .32 — a large and unsurprising effect. The induction interacted with that, F(2,45) = 4.173, p < .05, ηp² = .16, and people in the disgust condition rated the photographs lower than people in the neutral condition, which is the finding that gets cited.

Three things about it that the citation does not carry:

The control condition did the same thing. The authors write that "surprisingly, the ratings given by the participants in the cleanliness condition were also lower," and explain it after the fact by noting that their hygiene images included hand soap and people washing, which may itself have induced a feeling of needing to wash. When a cleanliness prime depresses attractiveness ratings as much as a disgust prime does, the design is not cleanly isolating disgust.

The paper discloses a stimulus error. A footnote records that, due to a coding error, participants in the disgust condition saw a different mix of the scarred face stimuli than participants in the neutral condition did.

And it is not about a partner. The study induced disgust with gross-out photographs and measured ratings of strangers' faces, half of them digitally scarred. The ick is a reaction to a specific person you already liked, triggered by something they did. Using the first to explain the second is a long reach across a gap of 51 people in three cells.

None of which says disgust is irrelevant to the ick — the trait correlation in the ick study itself is the better evidence for that, and it is real. It says that the confident evolutionary account circulating on this topic rests on much less than its confidence implies. This site has found the same shape before, in the 72-hour rule and in the honeymoon phase: a strong mechanism story bolted onto a weak measurement.

#12

Open access that could not be opened

A note on method, because it determines what this page is allowed to claim and because it is a problem readers of research keep hitting.

The ick paper is labelled open access. Unpaywall records it as open at the publisher under a CC BY-NC-ND licence, dated 5 February 2025. OpenAlex agrees and calls it hybrid open access. Semantic Scholar lists an open-access PDF for it.

None of those routes produces a readable copy. Unpaywall's record has `url_for_pdf: null` and offers the DOI landing page as the only location. The ScienceDirect page returns HTTP 403 to a plain request, and returns Elsevier's "there was a problem providing the content you requested" interstitial to a real headless Chrome with a real user agent. The SSRN copy of the accepted version also returns 403. There is no PubMed Central deposit, because the paper is not indexed there at all. Three databases assert the article is free to read; the article is not reachable.

So every figure on this page from the ick study comes from its abstract or from the two pages that quote its tables, each attributed where it appears, and no effect size from it is reported here — because effect sizes are exactly what gets garbled in secondary reporting, as the contradiction over likelihood-versus-frequency above demonstrates. That restraint is the same one this site applied when two of the key papers on staying friends with an ex turned out to be unreadable: where the source cannot be checked, the number does not get published. Our editorial policy explains why, and the corrections page lists what has had to be fixed.

#13

So is the ick information, or is it a standard?

This is the question people are actually asking, and the evidence supports a frame rather than an answer. The frame is still worth more than the confident versions.

There are two things the feeling could be, and they look identical from inside.

It could be compressed information. You noticed something — a flicker of contempt towards a waiter, a joke that landed on someone who could not answer back, a reflex of entitlement — and the aversion arrived before the reasoning did. The trivial surface detail is a hook the memory hung itself on, not the content. People are not reliably able to report why they stopped wanting something, so a real observation showing up as an inexplicable feeling about jorts is a perfectly ordinary thing for a mind to do.

Or it could be a standard revealing itself. This is the reading the study's own trait findings point at: the people reporting more icks score higher on holding partners to exacting standards and on grandiose narcissism. In that version the trivial detail is the content, and what the feeling tells you is about the threshold being applied rather than about the person it was applied to.

The study cannot tell you which, in general or in your case, and nobody should pretend otherwise. What tilts the odds a little:

  • Does it survive a description? If you can say what the thing actually was and it still sounds like something — a disregard, a cruelty, a dishonesty — that is likelier to be the first kind. If describing it out loud makes it evaporate, that is information too.
  • Is it the pattern or the person? The frequency data are the useful part here. Six icks across a dating history is a different thing from an ick in every relationship. The second is a fact about the threshold, and it is the version this site wrote about in why smart people keep dating the wrong people.
  • What else was happening? The disgust-sensitivity correlation suggests the system is state-dependent. People are harsher when tired, stressed, unwell or already ambivalent, and an aversion that arrives in one of those states and is gone a week later was probably about the state.

What this page will not do is tell you to override it. The feeling is yours, the cost of a false positive is one relationship that might have been fine, and the cost of a false negative is considerably higher. Nobody owes anybody continued attraction, and "I was told my standards are too high" is a worse reason to stay than any ick is to leave.

#14

Does it go away? There is no study that can answer that

The most-asked practical question about the ick is whether it passes, and it has no evidential answer at all. This is worth saying flatly, because every page that answers it is answering from clinical impression or from nothing.

Answering it would need a design nobody has run: find people at the moment the ick arrives, follow them for months, and record what happened to the relationships. That is a prospective longitudinal study, and the literature on the ick consists of one cross-sectional survey. Even that survey, which asked people to recall icks from their dating history, did not ask what happened next — or if it did, neither of the two pages that read it reports the result.

What exists nearby is suggestive and no more. The 28% disclosure figure implies most icks are never discussed, which means most of them are resolved by withdrawal rather than by repair, which in turn means that even if the feeling could pass, the usual course of events does not give it the chance. And the ick study's authors said at the time that they planned follow-up work on long-term relationship outcomes. Twenty-one months later, the citation counts in the section above — 3, 1 and 0 depending on the database — suggest that work has not yet landed.

So the honest answer to "does the ick go away" is: nobody has measured it. Anecdotally it clearly does for some people, since 64% of a sample of adults reported having had one and most of those people went on to have relationships. But the thing you want to know — the odds for you, now, about this person — is not something any published evidence can supply, and a page that gives you a percentage for it has invented the percentage.

The nearest well-measured neighbour is a different question: what predicts whether early attraction grows into something. That one has real data behind it, and it is covered in what builds attraction over time and in how many dates it takes to know.

#15

If you are the one who got icked

Almost everything written on this subject addresses the person who has the ick. The arithmetic says the other role is just as common and gets no pages, so here is the short version, and it is mostly about what not to conclude.

You will usually not be told. 28% disclosure means that in roughly three cases out of four the person withdraws, the messages thin out, and no reason is given. If you have been on the receiving end of that and have spent weeks auditing a specific moment, the honest news is that you are unlikely to ever learn whether there was one, and that the reconstruction you build in the meantime is probably wrong — the pilot data are full of triggers nobody could anticipate or avoid.

It is not a judgement you can act on. The study's findings run in an uncomfortable direction for the standard interpretation: the trait that most consistently travels with reporting icks is holding partners to exacting standards. That does not mean the person who withdrew from you was a narcissist — a correlation across 125 people says nothing about an individual, which is the point made two sections up and it applies in this direction too. It does mean that treating someone else's sudden aversion as information about your worth is reading a very noisy signal as though it were a clean one.

And the behaviour, not the feeling, is the part worth judging. Someone deciding they are no longer attracted to you is not a wrong they have done. Vanishing, or staying while going cold, or keeping you available without intending anything, is a choice about how to treat a person and a different matter entirely — the distinction this site drew in micro-cheating and in the signs of a situationship.

If the pattern is that this keeps happening and you cannot see why, that is a reasonable thing to take somewhere better than an article — and the page on why the same thing keeps happening is the closest this site has to a researched answer.

#16

Why a TikTok word is worth studying at all

There is a reflex among researchers to dismiss terms like this as internet noise, and a reflex in the opposite direction to treat a single study as proof that the thing is real and understood. A paper published in April 2026 makes the better argument.

Emily Impett, writing in Personality and Social Psychology Review, argues that popular lay theories about relationships — her case study is the love languages — "shape how millions of people understand and communicate about romantic relationships, yet they are rarely examined scientifically," and that the right response is to study them rather than dismiss them as misconceptions. Her framework runs in both directions: lay theories tell researchers what people actually need and struggle with, and research gets translated back. Examining why an idea resonates, she argues, reveals unmet needs and failures of scientific communication.

That is a good description of what the ick is. It names a real experience that the literature on attraction had no word for and no instrument to measure: not the gradual decline that the honeymoon-phase research documents, and not a considered judgement, but a fast, disproportionate reversal. The reason millions of people adopted the word is that it described something they had and could not otherwise name, and that is a finding about relationships even before anyone measures it.

It also explains the risk, which Impett names too: oversimplified frameworks. The ick arrived with a built-in explanation — evolved disgust, mate-choice thresholds — and that explanation is now repeated with far more confidence than its one underpowered study and its 51-person supporting experiment can support. A word that describes an experience accurately is not the same as a theory that explains it, and the gap between those two is where most of the writing on this subject lives. This site made the same distinction about the five love languages yesterday, for the same reason: the metaphor caught on because it pointed at something true, and that is independent of whether the mechanism it proposed holds up.

#17

What none of this can tell you

The limits go in the body rather than in a footnote, because the most common use of a page like this one is to settle something it cannot settle.

There is one study. Not one good study among several — one, cross-sectional, retrospective, self-report, in 125 single adults, which is fewer than the 134 its own preregistered power analysis called for. Every statement on this page about traits and icks rests on it alone. No replication exists. No second research group has looked.

The study could not be read. It is labelled open access by three separate databases and is reachable through none of them. Everything quoted from it here comes from its abstract or from the two first-page pages that report its tables, attributed at each point, and no effect size from it appears on this page for that reason.

Its own abstract contradicts its coverage about the sample size, 138 against 125, and this page could not resolve that — only state it, and give the reason for preferring 125.

The three available summaries of its trait findings contradict one another about which traits predict having an ick and which predict having more of them. That disagreement is in the secondary sources, not in this page's reading of them, and it is unresolvable from outside the paper.

The supporting experiment on disgust and attraction has 51 people in three cells, found the same effect in its cleanliness control as in its disgust condition, discloses a stimulus coding error, and is about digitally scarred photographs of strangers rather than about partners.

Nothing longitudinal exists. Whether an ick passes, whether it predicts anything about the relationship, whether acting on it leads anywhere better or worse — unmeasured, all of it.

And there is no instrument. No validated scale for the ick means no way to compare across studies even if more arrive. What the one study measured is what it asked.

What survives is small and still useful. The experience is common but not universal: given a definition, 64% of single adults aged 24 to 72 said they had had one, and about a third said they had not. Women know the word far better than men do; they do not have the experience more often. Most icks — 72% — are never mentioned to the person who caused them. Appearance is a less common trigger than behaviour. And the traits that travel with reporting more of them are about the height of the standard rather than about the partner, which is the single most useful thing in the literature and also the thing one underpowered survey is least equipped to prove.

Beyond that, the feeling is yours to read, and nobody studying it is further along than you are. More of this under Dating, and the site's attachment-style test is the nearest thing here to a measured look at what you bring to the reaction — an educational tool, not a diagnosis.

Pulling it together

The honest summary is smaller than any headline on this subject and more useful than most of them.

The ick is a real and common experience with almost no science behind it. One cross-sectional survey of 125 single adults, underpowered by its own preregistered standard, unreplicated, and unreadable behind a publisher's bot check despite being labelled open access. Its descriptive findings are the trustworthy part: given a definition, 64% of adults aged 24 to 72 said they had had one and about a third said they had not; women knew the word far better than men but did not report the experience more often; appearance was a less frequent trigger than behaviour; and roughly 72% of icks were never mentioned to the person who caused them.

Its trait findings point somewhere genuinely interesting and cannot carry much weight: the people reporting more icks score higher on disgust sensitivity, on grandiose narcissism, and on holding partners to exacting standards. Which of those predicts having one versus having many is reported three different ways by three different summaries of the same paper.

The evolutionary mechanism everyone repeats rests on less than it sounds. The experiment usually cited for it had 51 people in three cells, found the same drop in its cleanliness control as in its disgust condition, discloses a stimulus coding error, and measured ratings of digitally scarred photographs of strangers rather than reactions to a partner.

What that leaves you with is a judgement rather than a rule, and two questions worth more than any percentage: does the thing still sound like something when you describe it out loud, and is this a fact about one person or about every person you have dated. Nobody studying the ick is further along than you are on either.

Frequently asked questions

What is the ick in dating?

The ick is a sudden drop from attraction into distaste, triggered by something specific and usually trivial that a person did. Its two consistent features are speed — it arrives rather than develops — and disproportion between the size of the trigger and the size of the reaction. It is not a clinical term: it appears in no diagnostic manual, and no validated instrument measures it, which means that when a study measures the ick it measures whatever its questionnaire asked. The Oxford English Dictionary published an entry for ick as a noun and interjection on 2 March 2023. One empirical study of the dating sense exists, by Brian Collisson, Eliana Saunders and Chloe Yin in Personality and Individual Differences in 2025.

Where did the term the ick come from?

From an episode of Ally McBeal, and the date usually given is wrong by a year. Dictionary.com, the top result for this question, attributes the dating sense to "Once in a Lifetime" and dates the episode to 1999. The episode is season 1, episode 15, directed by Elodie Keene, and it aired on 23 February 1998; Wikipedia's article on the ick also says 1998. The term then sat mostly unused in this sense for over a decade. Wikipedia credits its popularisation to Olivia Attwood on the 2017 series of Love Island, with the TikTok wave following from around 2020, and the Oxford English Dictionary added an entry in March 2023.

Is the ick backed by science?

Barely. Exactly one empirical study of it exists, and the literature is small enough that the databases disagree about whether anyone has cited it: twenty-one months after publication, Crossref reports 3 citing works, Semantic Scholar reports 1, and OpenAlex's citation filter returns 0. The paper is not indexed in PubMed or Europe PMC at all — a title search for "ick" in Europe PMC returns 47 records, all of them about the ICK kinase gene, venom peptides, or the "ick factor" in organ donation and chemotherapy. The study itself is a one-sitting, retrospective, self-report survey of 125 single adults, which is nine fewer than its own preregistered power analysis called for. So the phenomenon is real in the sense that people plainly experience it, and every confident explanation of why is an extrapolation.

How many people were in the ick study?

Three numbers are in circulation for the same piece of research. The abstract distributed with the article's record states N = 138. Every outlet that appears to have read the paper reports 125 single adults aged 24 to 72 — the BPS Research Digest, tutor2u, Newsweek and StudyFinds all give that figure, and one report breaks it down as 74 men and 51 women, which sums to exactly 125. The accepted version was also posted to SSRN in 2024 under a different title, "The Ick Effect." The likely explanation is dull — 138 recruited, 125 analysed after exclusions, with the abstract carrying the pre-exclusion figure — but it could not be confirmed, because the published paper is unreachable behind the publisher's bot check despite being labelled open access.

Does getting the ick mean you are a narcissist?

No, and the study that produced those headlines does not support the claim about any individual. What it found was a correlation, across roughly 125 people, between reporting more icks and scoring higher on grandiose narcissism, on other-oriented perfectionism, and on disgust sensitivity. A modest correlation in a sample that size says nothing diagnostic about one person. It is also worth knowing that the three available summaries of the paper disagree about which of those traits predicts whether someone gets the ick and which predicts how often — the kind of distinction an underpowered study is least able to settle. The defensible reading is that frequent icks say more about the height of the standard being applied than about the partners it was applied to.

Does the ick go away?

Nobody has measured it. Answering that question would require finding people at the moment the ick arrives and following them for months, and no such study exists — the entire literature is one cross-sectional survey that asked people to recall icks from their dating history and, as far as either page that read its tables reports, did not ask what happened next. One nearby figure is suggestive: only about 28% of people told the partner who caused it, so most icks are resolved by withdrawal rather than by anything that would give the feeling a chance to pass. Anecdotally it clearly does fade for some people, since 64% of a sample of adults reported having had one and most went on to have relationships. Any page that gives you a percentage for your own odds has invented it.

Is the ick a red flag or am I being too picky?

The evidence supports a frame rather than an answer, because the two possibilities look identical from the inside. One is that the feeling is compressed information: you noticed something real — a flicker of contempt, a joke at someone's expense who could not answer back — and the aversion arrived before the reasoning did, with the trivial detail acting as the hook memory hung it on. The other is that the standard is revealing itself, which is the reading the study's own trait findings point at. Two questions tilt the odds. Does it survive a description — if you say what the thing actually was, does it still sound like something? And is it the pattern or the person: an ick in every relationship is a fact about the threshold, not about the people. Nobody should tell you to override the feeling; the cost of a false positive is one relationship that might have been fine.

Why is disgust the explanation everyone gives for the ick?

Because it is tidy, and the citation chain behind it is thinner than its confidence implies. The usual claim is that the ick is an evolved pathogen-avoidance system misfiring on a romantic prospect. The experiment most often cited for the step from disgust to lost attraction is Phelan and Edlund (2015) in Evolutionary Psychological Science, which this site read in full because, unlike the ick paper, it is free. It had 51 participants in three cells — disgust induction n = 18, hygiene n = 16, neutral n = 17 — and three problems the citation does not carry: the cleanliness control condition lowered attractiveness ratings as much as the disgust condition did, which the authors call surprising and explain after the fact; the paper discloses a stimulus coding error in the disgust condition; and it measured ratings of strangers' faces, half of them digitally disfigured, not reactions to a partner. The better evidence that disgust is involved is the trait correlation in the ick study itself.

Sources

What this article draws on. For the broader bibliography see our sources page.

Keep reading