Real World Appeal

Methodology

The research behind the threshold model

Real World Appeal reads a man's first impression the way strangers form it: in a fraction of a second, as a whole, and with a threshold. The photos set the band on a 70–155 perception axis, nine answers about how you carry yourself move the read inside that band, and a natural ceiling adds up only the levers that published research shows are changeable.

This page says what each part rests on. The studies below do not validate the product — we have run no clinical trial and say so — but they are the reason the model is shaped the way it is, and every one of them can be checked.

Key numbers

  • 100 ms — the exposure after which trait judgments of a face already match unlimited-time judgments (Willis & Todorov 2006).
  • 11 — the meta-analyses in which Langlois and colleagues found raters agree about who is attractive, within and across cultures (Langlois et al. 2000).
  • 37 cultures, n≈10,047 — the scale of Buss's cross-cultural survey of mate preferences (Buss 1989).
  • 76% / 27% — the higher odds of a 'yes' per step in expansive posture at a speed-date, and for an expansive profile photo on a dating app (Vacharkulksemsuk et al. 2016).
  • About 25% — how much more desirable than themselves online daters' targets are on average, with reply rates falling as the gap widens (Bruch & Newman 2018).
  • r = 0.13–0.32 — how weakly self-rated attractiveness tracks strangers' ratings across six studies, N = 1,180 (Greitemeyer 2020).
  • > 70% — the share of men's bodily attractiveness explained by perceived upper-body strength (Sell et al. 2017).
  • 11,570 people, 41 countries — the registered-report test of whether face impressions share one structure across world regions (Jones et al. 2021).

Why a threshold, not a ladder?

Because the people doing the judging behave like threshold-setters, and the field that measures perception has modelled yes/no decisions as an S-shaped curve for a century.

Two literatures point the same way. In decision theory, Simon showed that people set an aspiration level and accept the first option that clears it rather than optimising (Simon 1956); mate-choice theory formalised the same idea as a fixed-threshold rule — each candidate is accepted or passed as he is met (Janetos 1980). In psychophysics, the probability that a stimulus gets a 'yes' rises as a sigmoid, the psychometric function (Wichmann & Hill 2001): flat below, steep around the threshold, flat again above.

Dating data looks like that curve. In a speed-dating event, women did not pick by their stated preference lists; they set an aspiration level from their own self-perceived attractiveness and chose men who cleared it (Todd et al. 2007). In four U.S. cities' messaging data, people aimed about 25% above themselves and the chance of a reply dropped markedly as the desirability gap grew (Bruch & Newman 2018). Returns to being 'a bit more attractive' are not constant: they are small below the line, large at it, and small again above it.

How the model uses it: the 70–155 axis is divided into six bands with a threshold around 115. The report tells you how far you sit from the next inflection and which lever crosses it, instead of handing you a number to rank yourself with. Above the line, the modules shift from 'more handsome' to type, presence and market fit — because that is where the curve flattens.

The sigmoid is a modelling choice informed by these results, not a curve measured on you personally. Nobody has plotted one man's exact response curve, and we don't claim to.

Simon 1956Janetos 1980Wichmann & Hill 2001Todd et al. 2007Bruch & Newman 2018

How fast is the read, and what is in it?

Trait judgments of a face are already formed after 100 milliseconds, they are made on the whole face at once, and they land on a few dimensions that look similar around the world.

Judgments made after a 100-ms glance agree with judgments made with no time limit (Willis & Todorov 2006), and brief 'thin slices' of behaviour predict interpersonal outcomes about as well as long observation (Ambady & Rosenthal 1992). The face is judged holistically: pairing a top half with a better-looking bottom half raises the rating, and the effect vanishes when the halves are misaligned (Abbas & Duchaine 2008).

Impressions from 1,000 everyday photos fall on three dimensions — approachability, dominance and youthful-attractiveness (Sutherland et al. 2013) — and the underlying valence–dominance structure largely replicated across 11,570 participants in 41 countries, with some dependence on how the data are analysed (Jones et al. 2021).

How the model uses it: one read is produced first, as a whole, and only then separated into face, physique and outfit. The seven signals the report says 'she reads in 1.5 seconds' are dimensions of that read — approachable, confident, mature, stable, warm, high-quality, playful — not a checklist of features.

The 100-ms result is about photographs of faces. We never stretch it to cover height, build, clothes or voice; those have their own evidence below.

Willis & Todorov 2006Ambady & Rosenthal 1992Abbas & Duchaine 2008Sutherland et al. 2013Jones et al. 2021

Why face, physique and outfit are read separately

Because they carry different weights and different levers: the face predicts the overall judgment best, the body is read mainly as strength, and clothing changes the read within seconds.

When 127 men and 133 women rated faces, bodies and combined photos, face ratings predicted the combined judgment best, and men weighted women's bodies more under a short-term framing (Currie & Little 2009). For men's bodies, perceived upper-body strength explains more than 70% of the variance in attractiveness, with height and leanness adding to it (Sell et al. 2017), and the body composition raters find most attractive coincides with the one they find healthiest (Brierley et al. 2016).

Clothing works fast: with up to five seconds per image, a made-to-measure suit raised impressions of confidence and success over an off-the-peg one that differed only in small details (Howlett et al. 2013).

How the model uses it: face and physique get their own reads on the same axis, each with aesthetic sub-axes. Today's outfit is read but not scored — it changes every time you get dressed — and the ceiling treats fit, grooming and body composition as separate levers because the research treats them separately.

These weights come from group averages. A single viewer, in a single setting, can weight the channels differently — which is why the scene and market modules exist.

Currie & Little 2009Sell et al. 2017Brierley et al. 2016Howlett et al. 2013

Why your habits move the read inside the band

Because body language measurably changes who gets picked at zero acquaintance — and because the effect is real but bounded, it moves the number within the band the photos set, never across it.

Across 144 speed-dates, each standard-deviation step in expansive, open posture raised the odds of a 'yes' by 76%; on a dating app, the same person's expansive photo was 27% more likely to be chosen than his contracted one, and the effect ran through perceived dominance and openness (Vacharkulksemsuk et al. 2016). Gaze that turns toward the viewer is rated more likeable and more attractive (Mason et al. 2005), direct gaze is preferred most when the face is smiling (Conway et al. 2008), and a broader smile can offset a less attractive face (Golle et al. 2014).

How the model uses it: the nine presence answers — eye contact, how long you hold it, whether she looks back, making way, grooming effort, fidgets, self-monitoring — move the read by roughly two to five points inside the band. Where the photos contradict the self-report (you say you keep your space; your shoulders are rolled in), the photos win and the report says so. The presence module shows you exactly what your habits did to the number.

These are zero-acquaintance effects measured in brief encounters. They say nothing about a second date, and a self-report is an honest guess, not a measurement — which is why its influence is capped.

Vacharkulksemsuk et al. 2016Mason et al. 2005Conway et al. 2008Golle et al. 2014

Why 'noticed' is a count, not a score

Because attention is the first thing attractiveness buys, and attention is unevenly distributed in a way eye-tracking can see.

Across five experiments, attention was selectively drawn to physically attractive targets (Maner et al. 2003). In eye-tracking of real-world scenes, more attractive faces were looked at longer and more often, and more so for the opposite sex (Leder et al. 2016).

How the model uses it: the notice-rate module projects, for 100 women passing you in public, how many register you at all, how many look twice and how many feel a pull — anchored to your band, and projected again at your ceiling. It is a frequency, which is how attention actually works, not a grade.

It is a projection from perception research and your photos, not a measurement of any particular street. Treat the three numbers as a shape, not a census.

Maner et al. 2003Leder et al. 2016

Why we hold your own guess against the read

Because self-ratings of attractiveness track strangers' ratings only weakly, and people act on their self-rating whether or not it is right.

In six studies with 1,180 participants, self-rated and stranger-rated attractiveness correlated between 0.13 and 0.32; the less attractive overestimated themselves considerably, while the attractive were accurate or slightly underestimated (Greitemeyer 2020). The self-rating is not idle: in speed-dating, women's self-perceived attractiveness set how high they aimed, correlating 0.58 with the mate value of the men they chose (Todd et al. 2007).

How the model uses it: before the analysis you tell us how you think strangers see you. That guess never moves the number. The calibration module puts it on the same six-band ladder as the read, explains the gap with your own answers and photos, and names what the miscalibration makes you do wrong in dating — aim too low, or too high.

The direction of error in Greitemeyer's samples was on average; an individual can be off in either direction, which is exactly what the module checks.

Greitemeyer 2020Todd et al. 2007

Why the ceiling counts only changeable levers

Because the research is specific about what moves a first impression without surgery: body composition, strength cues, facial hair, fit, posture and gaze — and equally specific about what does not move.

Body composition and strength carry most of the body read (Brierley et al. 2016; Sell et al. 2017). Among 351 women, heavy stubble was rated the most attractive of four facial-hair levels (Dixson & Brooks 2013). Tailoring alone shifted five-second impressions (Howlett et al. 2013), and posture and gaze changed who was chosen (Vacharkulksemsuk et al. 2016; Mason et al. 2005). Height, by contrast, is fixed and matters: among 650 students, women preferred a larger height gap than men did (Stulp et al. 2013).

How the model uses it: the natural ceiling is the read you could reach with every changeable lever tuned and no cosmetic procedure — the roadmap's gains sum exactly to the gap, the gap is capped, and height and bone structure are listed as hard limits rather than hidden in the number.

The ceiling is an estimate of the best realistic version of you, not a promise. Lever sizes are the model's judgment informed by these studies, not effect sizes copied from them.

Brierley et al. 2016Sell et al. 2017Dixson & Brooks 2013Howlett et al. 2013Vacharkulksemsuk et al. 2016Mason et al. 2005Stulp et al. 2013

Why seven cultural markets and six scenes

Because agreement about who is attractive is real but partial: it holds across cultures at the core, varies at the edges, and the same man is read differently depending on where and how he is encountered.

Raters agree about who is attractive within and across cultures (Langlois et al. 2000), and the structure of face impressions largely holds across world regions (Jones et al. 2021). But what a face signals shifts with the market: a preference for exaggerated sex-typical faces appeared mainly in highly developed, urban populations across 12 groups (Scott et al. 2014), men and women weight partner traits differently across 37 cultures (Buss 1989), and private taste accounts for nearly as much of a facial judgment as shared taste (Hönekopp 2006).

Setting matters too: the short-term versus long-term framing changed how much bodies counted (Currie & Little 2009), and the same expansive posture was worth 76% at a speed-date but 27% on an app (Vacharkulksemsuk et al. 2016).

How the model uses it: the cultural-lens module reads your observable traits against the mainstream aesthetic of seven markets, and the scene module against six discovery mechanisms — apps, nightlife, gym, social circle, work, street. Both are relative reads of you; neither ranks groups of people, and the prefaces say so.

Markets are described as mainstream tendencies with large individual variance. No module claims that any group 'prefers' anything universally.

Langlois et al. 2000Jones et al. 2021Scott et al. 2014Buss 1989Hönekopp 2006Currie & Little 2009Vacharkulksemsuk et al. 2016

What the model cannot do

It cannot predict whether one specific person will want you, it cannot see past two photos and nine answers, and it is not a clinical instrument.

Machine learning on more than 100 traits predicted 4–18% of how much people desire others in general and 7–27% of how much they are desired — and nothing of the unique pull between two particular people (Joel et al. 2017). Nearly half of a facial-attractiveness judgment is private taste (Hönekopp 2006). The read describes how a mainstream market tends to respond in the first glance; it is not a verdict on you, and it is not a validated psychological test.

How the model uses it: every module carries this disclaimer, the ceiling is capped, the scores never become a rank of women, and the plan is written as levers you control — because that is the part the evidence supports.

We calibrate existing peer-reviewed work into an engine that answers in about a minute. We have not run an original large-sample study, and we don't claim to.

Joel et al. 2017Hönekopp 2006

Frequently asked questions

Is the 70–155 axis a score out of 100?

No. It is a perception axis shaped like an IQ distribution: 105–114 is the median everyday man, around 115 is the threshold where one lever has crossed, 130–144 is a head-turner. It describes how strongly a mainstream market reacts in the first glance, not a mark out of 100 and not a rank.

Is this a validated psychological test?

No. The model is calibrated on the peer-reviewed research listed on this page, and the analysis runs on your photos and nine answers. It is a self-improvement and calibration tool, not a clinical instrument, and we have not run an original large-sample validation study.

Do my questionnaire answers change my number?

Within the band only. The photos set the band; the nine presence answers move the read by roughly two to five points inside it, in line with the zero-acquaintance body-language evidence. Your own guess at how strangers see you never moves the number — it is held up against the read in the calibration module.

Why does the ceiling exclude surgery?

Because the ceiling is the best realistic version of you with the levers research shows are changeable: body composition, strength, grooming, fit, posture, gaze. Height and bone structure are listed as hard limits, and the gap between read and ceiling is capped.

Why do you say 'a mainstream market' instead of 'women'?

Because the research says agreement about attractiveness is real but partial — nearly half of a facial judgment is private taste, and preferences shift across cultures and settings. The read describes a tendency, not any one person's reaction.

Studies referenced

Every entry was verified against the Crossref record for its DOI on 16 October 2026. The one-line notes say what each paper contributes to the model.

  1. Willis, J., & Todorov, A. (2006). First impressions: Making up your mind after a 100-ms exposure to a face. Psychological Science, 17(7), 592–598. doi:10.1111/j.1467-9280.2006.01750.x

    Trait judgments made after a 100-millisecond look at a face agree with judgments made without a time limit.

  2. Ambady, N., & Rosenthal, R. (1992). Thin slices of expressive behavior as predictors of interpersonal consequences: A meta-analysis. Psychological Bulletin, 111(2), 256–274. doi:10.1037/0033-2909.111.2.256

    Short slices of behaviour predict interpersonal outcomes about as well as longer exposure.

  3. Abbas, Z.-A., & Duchaine, B. (2008). The role of holistic processing in judgments of facial attractiveness. Perception, 37(8), 1187–1196. doi:10.1068/p5984

    Attractiveness is judged on the whole face, not feature by feature: a top half paired with a better bottom half is rated higher, and the effect disappears when halves are misaligned.

  4. Sutherland, C. A. M., Oldmeadow, J. A., Santos, I. M., Towler, J., Burt, D. M., & Young, A. W. (2013). Social inferences from faces: Ambient images generate a three-dimensional model. Cognition, 127(1), 105–118. doi:10.1016/j.cognition.2012.12.001

    First impressions from 1,000 everyday photos fall on three dimensions: approachability, dominance and youthful-attractiveness.

  5. Jones, B. C., DeBruine, L. M., Flake, J. K., et al. (2021). To which world regions does the valence–dominance model of social perception apply?. Nature Human Behaviour, 5(1), 159–169. doi:10.1038/s41562-020-01007-2

    A registered report with 11,570 participants in 41 countries: the valence–dominance structure of face impressions largely replicates across world regions, with some dependence on analytic choices.

  6. Langlois, J. H., Kalakanis, L., Rubenstein, A. J., Larson, A., Hallam, M., & Smoot, M. (2000). Maxims or myths of beauty? A meta-analytic and theoretical review. Psychological Bulletin, 126(3), 390–423. doi:10.1037/0033-2909.126.3.390

    In 11 meta-analyses, raters agree about who is attractive, within and across cultures, and attractive people are judged and treated more favourably.

  7. Buss, D. M. (1989). Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioral and Brain Sciences, 12(1), 1–14. doi:10.1017/S0140525X00023992

    Across 37 cultures (n≈10,047), what men and women weight in a partner differs in consistent ways.

  8. Scott, I. M., Clark, A. P., Josephson, S. C., et al. (2014). Human preferences for sexually dimorphic faces may be evolutionarily novel. PNAS, 111(40), 14388–14393. doi:10.1073/pnas.1409643111

    Across 12 populations, a preference for exaggerated sex-typical faces appears mainly in highly developed, urban settings — what a face signals shifts with the market it is read in.

  9. Hönekopp, J. (2006). Once more: Is beauty in the eye of the beholder? Relative contributions of private and shared taste to judgments of facial attractiveness. Journal of Experimental Psychology: Human Perception and Performance, 32(2), 199–209. doi:10.1037/0096-1523.32.2.199

    Private taste accounts for nearly as much of a facial-attractiveness judgment as shared taste — the consensus is real but it is only about half the story.

  10. Currie, T. E., & Little, A. C. (2009). The relative importance of the face and body in judgments of human physical attractiveness. Evolution and Human Behavior, 30(6), 409–416. doi:10.1016/j.evolhumbehav.2009.06.005

    With 127 men and 133 women rating face, body and combined photos, the face predicts the overall judgment best; men weight women's bodies more in a short-term framing.

  11. Sell, A., Lukaszewski, A. W., & Townsley, M. (2017). Cues of upper body strength account for most of the variance in men's bodily attractiveness. Proceedings of the Royal Society B, 284(1869), 20171819. doi:10.1098/rspb.2017.1819

    Perceived upper-body strength explains more than 70% of the variance in men's bodily attractiveness; height and leanness add to it.

  12. Brierley, M.-E., Brooks, K. R., Mond, J., Stevenson, R. J., & Stephen, I. D. (2016). The body and the beautiful: Health, attractiveness and body composition in men's and women's bodies. PLoS ONE, 11(6), e0156722. doi:10.1371/journal.pone.0156722

    For male bodies, the fat and muscle levels raters found most attractive coincide with the levels they found healthiest — the lever is body composition, not thinness.

  13. Dixson, B. J., & Brooks, R. C. (2013). The role of facial hair in women's perceptions of men's attractiveness, health, masculinity and parenting abilities. Evolution and Human Behavior, 34(3), 236–241. doi:10.1016/j.evolhumbehav.2013.02.003

    351 women rated heavy stubble the most attractive of four facial-hair levels; full beards scored highest on masculinity and parenting.

  14. Howlett, N., Pine, K., Orakçıoğlu, I., & Fletcher, B. (2013). The influence of clothing on first impressions: Rapid and positive responses to minor changes in male attire. Journal of Fashion Marketing and Management, 17(1), 38–48. doi:10.1108/13612021311305128

    274 raters, up to five seconds per image: a made-to-measure suit raised impressions of confidence and success over an off-the-peg one that differed only in minor details.

  15. Stulp, G., Buunk, A. P., & Pollet, T. V. (2013). Women want taller men more than men want shorter women. Personality and Individual Differences, 54(8), 877–883. doi:10.1016/j.paid.2012.12.019

    Among 650 students, women preferred a larger height gap than men did; satisfaction with a real partner peaked when he stood about 21 cm taller.

  16. Vacharkulksemsuk, T., Reit, E., Khambatta, P., Eastwick, P. W., Finkel, E. J., & Carney, D. R. (2016). Dominant, open nonverbal displays are attractive at zero-acquaintance. PNAS, 113(15), 4009–4014. doi:10.1073/pnas.1508932113

    Across 144 speed-dates, each standard-deviation step in expansive posture raised the odds of a 'yes' by 76%; on a dating app, profiles with expansive photos were 27% more likely to get a 'yes'.

  17. Mason, M. F., Tatkow, E. P., & Macrae, C. N. (2005). The look of love: Gaze shifts and person perception. Psychological Science, 16(3), 236–239. doi:10.1111/j.0956-7976.2005.00809.x

    Faces whose gaze turned toward the viewer were rated more likeable, and in a second experiment more attractive, than faces that looked away.

  18. Conway, C. A., Jones, B. C., DeBruine, L. M., & Little, A. C. (2008). Evidence for adaptive design in human gaze preference. Proceedings of the Royal Society B, 275(1630), 63–69. doi:10.1098/rspb.2007.1073

    Direct gaze is preferred most when the face is smiling, and that preference is strongest for opposite-sex faces judged on attractiveness.

  19. Golle, J., Mast, F. W., & Lobmaier, J. S. (2014). Something to smile about: The interrelationship between attractiveness and emotional expression. Cognition and Emotion, 28(2), 298–310. doi:10.1080/02699931.2013.817383

    Smile intensity strongly shapes which of two faces looks more attractive — a broader smile can offset a less attractive face.

  20. Maner, J. K., Kenrick, D. T., Becker, D. V., et al. (2003). Sexually selective cognition: Beauty captures the mind of the beholder. Journal of Personality and Social Psychology, 85(6), 1107–1120. doi:10.1037/0022-3514.85.6.1107

    Across five experiments, attention is selectively drawn to physically attractive targets — noticing is not evenly distributed.

  21. Leder, H., Mitrovic, A., & Goller, J. (2016). How beauty determines gaze! Facial attractiveness and gaze duration in images of real world scenes. i-Perception, 7(4). doi:10.1177/2041669516664355

    Eye-tracking in real-world scenes: more attractive faces are looked at longer and more often, and more so for the opposite sex.

  22. Greitemeyer, T. (2020). Unattractive people are unaware of their (un)attractiveness. Scandinavian Journal of Psychology, 61(4), 471–483. doi:10.1111/sjop.12631

    Six studies, N = 1,180: self-ratings and strangers' ratings correlate only modestly (r = 0.13–0.32); the less attractive overestimate themselves, the attractive are accurate or slightly underestimate.

  23. Todd, P. M., Penke, L., Fasolo, B., & Lenton, A. P. (2007). Different cognitive processes underlie human mate choices and mate preferences. PNAS, 104(38), 15011–15016. doi:10.1073/pnas.0705290104

    In a 46-person speed-dating event, stated preferences did not predict choices; men chose on physical attractiveness, and women set their aspiration level from their own self-perceived attractiveness (r = 0.58 with the mate value of the men they chose).

  24. Bruch, E. E., & Newman, M. E. J. (2018). Aspirational pursuit of mates in online dating markets. Science Advances, 4(8), eaap9815. doi:10.1126/sciadv.aap9815

    Messaging data from four U.S. cities: people pursue partners about 25% more desirable than themselves, and the probability of a reply drops markedly as the desirability gap widens.

  25. Joel, S., Eastwick, P. W., & Finkel, E. J. (2017). Is romantic desire predictable? Machine learning applied to initial romantic attraction. Psychological Science, 28(10), 1478–1489. doi:10.1177/0956797617714580

    Machine learning on 100+ traits predicts 4–18% of how much people desire others in general and 7–27% of how much they are desired, but nothing of the pull between two specific people.

  26. Simon, H. A. (1956). Rational choice and the structure of the environment. Psychological Review, 63(2), 129–138. doi:10.1037/h0042769

    Decision-makers set an aspiration level and accept the first option that clears it — the origin of 'satisficing'.

  27. Janetos, A. C. (1980). Strategies of female mate choice: A theoretical analysis. Behavioral Ecology and Sociobiology, 7(2), 107–112. doi:10.1007/BF00299515

    Formalised the fixed-threshold rule of mate choice: each candidate is accepted or passed as he is met, against an internal standard.

  28. Wichmann, F. A., & Hill, N. J. (2001). The psychometric function: I. Fitting, sampling, and goodness of fit. Perception & Psychophysics, 63(8), 1293–1313. doi:10.3758/BF03194544

    The standard treatment of the psychometric function — the S-shaped curve psychophysics uses to describe how the probability of a 'yes' rises with stimulus strength.

See the model read you

About a minute. The analysis runs first; the full report is $5 one-time, no subscription.

Or look at a sample report first