By the Numbers

Before splitting anything by gender, look at the whole population: a long-term partner is the top major reason, and casual sex sits near the bottom. These don't sum to 100% because they aren't either/or — the same person can want a relationship and be open to casual. That overlap is exactly why the "82% want love / 18% want sex" framing (see the last group) is a category error.

Casual sex is the one reason where men and women diverge — men out-cite it by about two and a half to one. On the other major reasons (long-term partner, casual dating, friends) Pew found no significant gender gap. So the honest story isn't "men want sex, women want love" — both rank a relationship first; men simply layer more casual openness on top. It's an overlay, not a different species.

Both sexes want the man taller — but women's pull is the stronger one, and it shows up in real couples: the man is taller in ~92% of couples — a few points above the ~90% that random pairing of the height distributions alone would produce (Stulp), while the sharper sign is how rare the reverse is: female-taller couples turn up roughly 14× less often than chance (Gillis & Avis found 1 in 720, vs ~1 in 50 expected). The catch the height-pill misses is that the curve bends. Success with height is curvilinear, not linear: the very shortest men are penalized, but the optimum is only moderately above average, and extremely tall men get screened out by shorter women too. A window, not a ladder — which is why "it's over under six feet" is fatalism, not data. See the Height Pill →

In first-time hookups about one in nine women orgasm, versus two-thirds in committed sex. This helps explain why lower casual openness can be rational rather than prudish: when researchers held expected pleasure and anticipated stigma constant, the gender gap in accepting casual-sex offers largely vanished. Women aren't less interested in pleasure — they're less interested in the version of it that reliably underdelivers.

The broader gender gap persists beyond the hookup context: ~95% of heterosexual men usually orgasm versus ~65% of heterosexual women when sexually intimate. So the Armstrong chart can show that relationships improve women's odds without implying commitment fully solves the gap. Casual isn't the only place the deal is lopsided; it's just the most lopsided.

This is the demand asymmetry made personal, and it cuts both ways. Women are far more likely to feel buried in attention; men are far more likely to feel invisible for the lack of it. Same market, opposite complaints — which is also why "she feels like a body, he feels like a wallet" has a quieter cousin: she feels swarmed, he feels ignored. The market's response is one chart down: pay-to-play →

Read this as a baseline-vs-subgroup comparison, not a clean gender split: the all-user bars already include these women, and Pew doesn't publish a tidy men-vs-women breakdown here — but women under 50 run well above the overall rate, and Pew reports men experience all three at lower rates. This is the "danger" half of why casual openness differs: the downside — harassment, threats, unsolicited explicit content — concentrates on younger women. Naming the male attention deficit above doesn't cancel this; both are true at once. The scam layer of the same danger picture cuts the other way: men under 50 are the group most likely to meet a suspected scammer on the apps (63%). Ruling M-TBD-65 →

The market's answer to the attention asymmetry one chart up: men are the buying side — 41% of male users have paid versus 29% of women — and the gradient runs exactly where you'd predict: with income (45% upper / 36% middle / 28% lower) and with age (41% of users 30+ vs 22% under 30). Paid users also report better experiences, 58% vs 50%. Read that last pair as a selection effect until proven otherwise: people willing to pay skew richer, older, and more intent on a result, and Pew can't say whether the boost bought the satisfaction or the satisfied were the sort who'd pay. What the chart can say is simpler: when one side of a market feels invisible, someone will sell visibility — Exposure is the lever men reach for first, wallet open.

Across 27 samples of real couples, partners' rated looks correlate at about r = .39 — real matching, far from random, and equally far from lockstep. This is the statistical spine under the SMV-Matching framework: on looks alone the market sorts assortatively — the average couple matches rather than one side relentlessly trading up (the "trading up" people notice enters through status instead). Two honesty flags carried over from the framework page: that matching exists is Tier-1, replicated for decades; the exact working coefficient is Tier 2, because the 2024 dyadic meta re-analyzes Feingold's 1980s-era samples rather than adding new data. And r ≈ .4 leaves plenty of scatter — individual couples mismatch constantly; the curve only binds on average. See the framework: SMV Matching →

No-sex-in-a-year among young adults roughly 2.4×ed in a decade — and the rise skews male. Ueda et al. (2020, JAMA Network Open) put men aged 18–24 reporting no sex in the past year at 18.9% → 30.9% between 2000–02 and 2016–18 — a far steeper climb than women the same age. (Single-wave young-men cells are small, so read the exact figure as directional — Tier 2.) It's the attention deficit made literal: more young men aren't just feeling ignored on the market, they're falling out of it.

Women wanted roughly 69% of the marriages that ended — though that figure rests on just 92 marital breakups in the panel (95% CI 61–78%). The surprise: non-marital breakups ran about 56% women-wanted among cohabiters (~53% among never-cohabiting couples) — neither significantly different from 50/50. So "women always leave" is really "women leave marriages more" — the gap is specific to the institution, not to women, which points at marriage carrying costs wives disproportionately feel rather than a simple branch-swing instinct.

Two things break the "every partner chips away at bonding" story. First, the biggest jump is from 0 to 1 — having any premarital sex, not the marginal extra partner — which points at who stays at zero (highly religious, traditional, already low-divorce) rather than a dose effect. Second, it's non-monotonic: women with 2 partners divorce more than those across the 3–9 band (which sits flat around 24–26%, not a clean step). The raw data simply don't show the monotonic dose-response a per-partner "bonding depletion" would require. The association is real but confounded — and the "oxytocin runs out" mechanism underneath the claim is pseudoscience. See the full duality →

Actor and partner personality are not the same channel. A separate nine-year cohort study of 972 people in 486 different-sex German couples found longitudinal associations between a person’s own lower neuroticism and higher conscientiousness and that person’s relationship satisfaction, but no detectable partner-personality effects. It sharpens this card’s own-versus-partner split; it does not prove that a partner’s character never matters. The study used three survivor cohorts, a short personality inventory and a single-item satisfaction measure, and remains observational. Lab find · 2026-07-31 · Bach et al. 2025 Primary study →

Once a couple exists, a person’s own experience of that relationship — perceived commitment, appreciation, sexual satisfaction, perceived partner satisfaction, conflict — carries far more current predictive information than either person’s trait inventory. That does not mean partner qualities are irrelevant: they may operate through the relationship experience being measured. It also does not make 45% a causal share, a success probability, or a compatibility formula. These are contemporaneous self-reports among formed couples, and some predictors sit conceptually close to satisfaction itself. The harder result is prospective: no model explained more than 5% of change over the typical one-to-two-year follow-up. Baseline quality was predictable; who would improve or decay was not reliably predicted from these baseline self-reports. See one bounded Kept-rung response pattern →

Across three intensive-measurement studies of older couples, moments when both partners were above their own usual positive-emotion level were associated with lower concurrent cortisol even after accounting for either person’s positive affect and a long list of person, medication, behavior and diurnal controls. Shared positivity also predicted lower cortisol at the next assessment; the reverse temporal path was not detected. That makes the shared state more than a duplicate count of two individual moods. It does not show that arranging a pleasant moment causes better health, relationship quality or longevity, and cortisol is a proximal biomarker rather than any of those outcomes.

Initiation is male-default, and not narrowly: across four U.S. cities, ~81% of first messages came from men, and women's average reply rate ran under 20%. It's the demand asymmetry from the attention chart seen from the other side — men do the reaching, women do the selecting. The pattern repeats off the apps: marriage proposals run ~95%+ male, and the man still pays the first date about three-quarters of the time. What this chart deliberately can't show is the popular claim that men are initiating less than they used to — no one has measured initiation as a time series, so we don't chart it. The withdrawal shows up only in downstream outcomes — the two charts below. See the model: The Men's Strike →

The permanent-single class is real Tier-1 data, not a meme: never-married-at-40 more than quadrupled in four decades, and Pew projects ~1 in 4 of today's young adults will never marry. The gender twist matters. Under 30, men are the more-single sex — Pew puts it at 63% vs 34%, though the GSS gap is far narrower (~56/54), so trust the direction, not the exact spread. But it flips with age: by 65+ women are the more-single sex by more than 2:1, because men remarry and women outlive them. The young surplus of single men becomes an old surplus of single women.

Most checked-out singles say they simply have other priorities or like it that way — but read these as self-reported: the flattering answers run high, the stigmatized one runs low, so the involuntary share is almost certainly understated. The one reason that splits sharply by sex is exactly that stigmatized one — "no one would be interested in me": men 26% vs women 12%, the involuntary-singlehood tell. And not-looking is itself gendered by age: among singles 40+, 71% of women aren't looking vs 42% of men — the female exit the apps never see, and the reason a "men's strike" lands on a market many women have already left.

This one we deliberately don't chart with hard numbers, because the evidence is genuinely mixed. Belief in the double standard is widespread, but controlled studies often find it weaker than expected — sometimes men and women both get judged for high partner counts. Still, it bites where it counts: anticipated stigma measurably lowers women's stated willingness to accept casual sex (Conley et al., 2013), which means the openness gap above is probably understated. The asymmetry is real as a social perception that shapes behavior — but anyone who tells you it's been cleanly measured at some exact ratio is selling you a number that doesn't exist.

The 1950s marriage age is the anomaly on this chart. Men bottomed at 22.5 in 1956 and women at 20.1 — the lowest in the entire 1890-to-present series — before climbing nearly eight years to 30.2 and 28.6 by 2024. The mid-century floor wasn't the deep historical baseline; it was a brief postwar compression that the rest of the century slowly unwound. That reframes the familiar “people used to marry young”: the young-marriage era people picture was one unusual generation, and the long rise since is partly a return toward an older normal. Read the Deep Dive: Relationships Throughout History →

The economic architecture under the Status Trade, measured across fifty years. The husband-provides marriage fell from 85% to 55% — still the majority, but its dominance has halved, and ~45% of marriages are now equal- or female-earning. This is the erosion the frameworks page calls “fading, not hardening”: hypergamy's material base thins as women's earnings rise, and educational hypergamy — wives marrying up in schooling — has already reversed worldwide (Esteve et al., 120 countries). One asterisk from the same report: even in equal-earning marriages wives still do more housework and caregiving, so the norm is halving, not gone. See the model: The Status Trade →

By 2017 the single most common way heterosexual couples met was online — 39%, ahead of every traditional venue. That share was near zero before 1995 and 22% as recently as 2009; meeting through friends, which had held around a third through the mid-'90s, fell to 20% and was overtaken around 2013. Family, church, and neighborhood had been sliding since World War II, and the graphical web and then the smartphone finished the job, cutting the human intermediary out of the introduction. One caveat the researchers flag themselves: the 27% bar/restaurant figure is inflated because many of those meetings were online-arranged first dates in a bar, not chance encounters. And the verdict on the winning channel splits by generation: U.S. adults overall call online dating a net help in the search for a long-term partner (42% easier vs. 22% harder), but the under-30s — the cohort that grew up inside the channel — split dead even at 35% easier vs. 33% harder (Pew 2023, n = 6,034).

In-person time with friends fell from roughly 60 minutes a day in 2003 to about 20 by 2020, and among 15-to-24-year-olds the drop was steeper still — around 70% over the same span. Most of that decline had already happened before 2020, so the pandemic accelerated a slide that was long underway; the contaminated 2020 endpoint marks the floor of the trend rather than its cause. The places where casual friendship used to accumulate — the “third places” the Deep Dive is about — were emptying out for two decades before anyone had heard of lockdowns. Read the Deep Dive: Third Spaces →

Seen from the third-places angle, the settings that once did the introducing are now rounding errors: by 2017 family accounted for 7% of how couples met, church 4%, and the neighborhood just 3%. These are the same figures as the ranking above, pulled out because they are precisely the venues a third place used to be — when the room empties, so does its role as somewhere people pair off. See the full how-couples-meet ranking →

Global fertility has been falling along one long curve: about five births per woman in the 1960s, 3.3 by 1990, and 2.2 in 2024, with the UN projecting the 2.1 replacement level around 2050. The regional spread the Deep Dive lays out — Sub-Saharan Africa at 4.3, Europe and North America near 1.6, East Asia and the Pacific down at 1.3 in 2023 — is best read as one snapshot of that descent, each region sitting at a different point on the same slope rather than following different rules. Read the Deep Dive: Relationships by Country →

South Korea, home to the world's lowest recorded fertility rate in 2023, shows the attitudes shifting underneath the number. The share calling marriage “a necessity” slipped from 56.8% in 2014 to 52.5% in 2024 — and split sharply by sex, 58.3% of men versus 46.8% of women — while acceptance of cohabitation without marriage jumped from 46.6% to 67.4%. That attitude change runs about a decade ahead of the East Asia fertility figures in the Deep Dive. The 2024 uptick — TFR 0.72 to 0.75, with marriages up 14.9% in the steepest annual rise since records began in 1981 — is a single year, not yet a trend; one point does not bend a curve this deep.

At the high-fertility end of the Deep Dive's ladder sits its harshest floor: child marriage. Among women now aged 20–24, 31% in Sub-Saharan Africa were married before 18 — the highest of any region, though down from about 38% a decade earlier — and West and Central Africa still run near one in three, home to six of the world's ten highest-prevalence countries. The direction of travel is downward everywhere, and fastest in South Asia, where the rate fell from 39% to 26% over the past decade; globally it is now about one in five. This is the coercion underneath the numbers that a bare fertility ranking never shows.

The one part of “the wall” that comes from the body, measured where the logging is best. ACOG marks the real turns at 32 (gradual) and 37 (faster) — not 30 — and IVF outcomes trace the same slope. Egg donation is the tell: when the eggs are young, success barely tracks the recipient's age, which points at egg quality rather than the womb or her general health. The gentler half of the story is natural conception: about 82% of 35-to-39-year-olds trying conceive within a year — while the infamous “1 in 3 over 35 will fail” figure traces to French birth records from 1670–1830. Real slope, wrong cliff. Read the Deep Dive: What the Wall Actually Is →

Across 130 countries and territories, no nation houses a larger share of its children with a lone adult: 23% of American kids, more than triple the global 7% — and the highest rate in the study. Two definitions do real work here. This counts solo-parent households: one parent, no other adult under the roof. And part of the U.S. lead is a housing fact as much as a family one — American children are far less likely to live in extended families (8% vs 38% globally), so the same broken pairing that makes a three-generation household in Lagos makes a “single-parent household” in Ohio. How the number was built — death, then divorce, then never-married — is the Deep Dive's whole arc. Read the Deep Dive: Single Parenthood →

Two series, one flip. Births outside marriage went from one in twenty to about four in ten — and have sat near 40% since 2009: a resettled baseline, not a spike. Underneath, the never-married share of single mothers grew elevenfold, pulling nearly level with the divorced, separated, and widowed combined — the widow who defined the category in 1950 has nearly vanished from it. One flattening caveat before anyone charts “fatherless America” off this: “unmarried” isn't “alone” — about 62% of births to never-married women arrive to cohabiting couples (Lamidi, via Child Trends). Read the Deep Dive: the route in →

The money line is brutal and undisputed: single-mother families run near six times the married-couple poverty rate — a second adult in the house is the largest anti-poverty program in American life. And the load lands asymmetrically: four custodial parents in five are women, a split that has moved only a few points in three decades of Census measurement. What this chart deliberately can't tell you is how much of the child-outcome gap is the split itself versus who ends up in each group — rigorous causal designs find real effects that are consistently smaller than the raw gaps (McLanahan), the same selection lesson as the body-count chart. Read the Deep Dive: the ledger →

Across the Investment Model literature, satisfaction is the strongest single correlate of commitment — but the exit door and the life already built matter almost as much. In relationship samples, satisfaction, alternatives, and investments together accounted for about 63% of the variance in commitment; commitment in turn correlated with stay/leave behavior at r = .47. That is a map of association, not a recipe for trapping someone: making alternatives worse can raise measured commitment without making a relationship good, safe, or worth preserving.

One partner pushes harder to discuss or change; the other evades, shuts down, or leaves; each response intensifies the other. Across 74 studies, that loop had a moderate association with worse outcomes. The useful unit is the cycle, not a villain: the meta-analysis combined who demanded and who withdrew, and the pooled result does not establish which behavior started the conflict or whether outside stress caused both the pattern and the dissatisfaction.

Married adults report a deeper center of gravity around the relationship. The gap survives statistical adjustment for age, gender, education, religion, race, and relationship duration, but that still does not prove the certificate caused it: people with stronger commitment may be more likely to marry. The cohabiting average also hides a bright line — about half of engaged or very-serious cohabiters said things were going very well, versus only 6% of cohabiters who described the relationship as neither.

The average drop is about three-tenths of a standard deviation and remarkably similar for mothers and fathers. That makes the transition a predictable load, not evidence that one sex suddenly values the relationship less. The strongest caution is the comparison group: only nine independent samples followed nonparents over a comparable span, and satisfaction can decline with relationship time even without a baby. This is good evidence for a first-year dip among parents; it is weaker evidence for how much of that dip the baby uniquely causes.

Three years in, marriage is the single largest destination, but a majority have not married that partner: roughly a third remain together without marrying and just over a quarter have split. Education changes the route sharply — 53% of first cohabitations among women with a bachelor’s degree transitioned to marriage, versus 30% among women without a high-school diploma. These are paths, not treatment effects: the survey did not randomly assign anyone to move in.

Even when paychecks are nearly even, wives average about 4.5 more hours of unpaid care and housework each week, while husbands average 3.6 more leisure hours and 3.1 more paid-work hours. Among parents in these marriages the care gap widens: mothers log 12.2 hours versus fathers’ 9.0. The diary shows allocation, not motive — job schedules, preferences, bargaining, and norms are folded together — but the accounting itself is not ambiguous.

The share rose more than fivefold after Loving v. Virginia, and 11 million married Americans were in an intermarriage by 2015. Composition matters: growth in Asian and Hispanic populations expanded the number of cross-group pairings available, while acceptance also changed sharply. This definition treats Hispanic/non-Hispanic pairings as interethnic and pairings across non-Hispanic racial groups as interracial; it is not a measure of cultural distance between two individuals.

The behavior gap and the stated-desire gap point the same way. More than half of previously married women said they did not want to remarry (54%), compared with 30% of men. But age carries much of the aggregate: among eligible adults ages 25–54, women and men were about equally likely to have remarried; the large difference remained among adults 55 and older. It is a cohort-and-life-stage pattern, not a universal rule about divorcees.

A binary “ever hit” frame hides the severity gradient. Men and women report nearly identical lifetime exposure to the broad physical-violence category, while women report substantially more severe acts such as being choked, beaten, burned, having a knife or gun used against them, or being slammed into something. These are victimization reports, not incident counts or perpetration rates; one person can experience many events. Lifetime recall, nonresponse, and willingness to disclose can all move the levels.

This is an intention shift before it is a completed-fertility statistic: younger respondents can change their minds, and “unlikely” is not “never.” Still, the same question moved ten points in five years. In a 2024 follow-up limited to under-50 nonparents who already said children were unlikely, 57% named simply not wanting children as a major reason; 44% cited other priorities, 38% the state of the world, and 36% cost. Falling births are not only delayed partnering or infertility — stated preference is part of the ledger.

Traits and preferences do predict something: who is broadly desirable, and who is broadly hard to please. What they do not predict is the thing every matching service sells — that two particular people will click. And this is not a case of a small target: unique, person-specific desire was the largest component of the outcome (about 35% of the variance in desire), which means the model had the most room to work exactly where it failed. Trained on one sample and tested on the other, the two unique-desire models scored −2.2% and −0.6% variance explained — worse than predicting the average for everyone. Yet the same algorithm recovered 16–29% of that same unique desire from what people said after four minutes of conversation. The practical reading is unglamorous: paper screening can rule out, it cannot match, so the information you actually need is on the other side of meeting. And once a couple exists, the relationship outpredicts the résumé →

Headcount says there is a surplus of never-married young men — 126 for every 100 women. Apply one filter that most women say they care about, a job, and the surplus becomes a shortfall: 91. Then the ratio flips again by education, and it flips hard. A woman with a post-graduate degree who wants a similarly credentialed man is choosing inside a pool of 67 employed men per 100 women; a woman with a high-school education or less is choosing inside one of 108. That is the same country, the same year, and two opposite markets — which is the whole point. Aggregate ratios describe nobody in particular; the ratio that governs your search is the one left standing after your own filters. Every filter you add is a market you are choosing to enter. See what the never-married share has done since →

Looks-matching is real, but its strength is not fixed — it depends on how the couple met. Strangers agree strongly about who is attractive, and where there is consensus there is a market, so couples who paired off quickly were tightly sorted by appearance. Time dissolves that consensus into personal taste: for couples who knew each other about nine months or longer before dating, attractiveness matching was no longer statistically distinguishable from zero. Read it as a structural fact rather than a trick. The stranger market prices you against everyone at once, which is exactly what a dating app is; a room you keep turning up in prices you against nobody. This is correlational and retrospective — friends-first couples may differ in other ways — but it is the cleanest available evidence that the venue sets the terms. See the baseline looks-matching coefficient →

Reconciliation is not a rare event — more than a third of cohabiting couples and nearly a quarter of married ones had already ended the relationship at least once and restarted it. But cyclers looked different afterward, in the same direction on both sides of the couple and at both stages: more uncertainty about the relationship’s future, and lower satisfaction (cohabiting women averaged 14.98 versus 16.72 on the satisfaction measure; married women 16.20 versus 17.12). Past cycling also predicted more cycling. Two honest limits: this is retrospective self-report, and it is correlational — going back does not prove the relationship got worse, because whatever caused the first breakup may be producing the lower satisfaction too. What the data will support is narrower and still useful: a breakup-and-return is information about the relationship, not a reset of it.

The people who moved in together before making any commitment to marry later reported lower marital satisfaction, less dedication, less confidence in the marriage, more negative communication, and more proneness to divorce. The people who moved in after getting engaged looked no different from the people who never cohabited at all — which is why the finding is about sequence, not about living together. Differences were small and survived controls for marriage length, age, income, education, and religiousness. The mechanism is genuinely contested and we are not going to pretend otherwise: Kuperberg (2014) shows that age at moving in absorbs much of the classic premarital-cohabitation penalty, and Rosenfeld & Roesler (2019) find premarital cohabitation raises divorce risk in the first year and lowers it later. What survives all three readings is modest and practical — the lease is not the decision, and letting it stand in for one is the part that shows up later. See where first cohabitations actually go →

Every association here runs against the advertising. Spending less than $1,000 on the wedding went with a lower divorce hazard; spending $20,000 or more went with a higher one. Filling the room, on the other hand, went the industry’s way — 101 to 200 guests carried the lowest hazard on the board, and taking a honeymoon of any cost went with a lower one. So the pattern is not “weddings are bad.” It is that the variables tracking the relationship — knowing each other well, dating three or more years before the proposal, having witnesses present — line up with survival, while the ones tracking the production do not. Three cautions. This is an online retrospective survey weighted to census benchmarks, not a probability sample. It is association, never causation: couples who throw cheap weddings differ from couples who throw expensive ones in ways the controls cannot fully absorb. And the courtship result is carried by men in this specification — among men the three-year hazard ratio was 0.58, among women 0.98 and not significant, a split the pooled bar hides.

This is the rare place in this literature with randomized trials behind it, and the answer is neither “communication fixes everything” nor “nothing works.” Taught couples ended up roughly a third to a half of a standard deviation ahead of untaught ones — a real, moderate shift, larger for observable skills than for how happy people felt. Dosage mattered: moderate-length programs beat short ones. One later randomized trial makes the magnitude concrete. Among 662 Army couples, those who had moved in together before committing to marry divorced at 14.6% over two years versus 7.5% for those who had not — but among couples randomly assigned to a skills program, that gap disappeared (7.8% versus 7.4%). Caveats worth keeping: the meta-analytic samples were thin on racial, ethnic and economic diversity, stability and aggression were rarely measured as outcomes, and the divorce comparison above is a subgroup within one military population, not the trial’s headline result. See the conflict pattern these programs target →

Pooled across the whole literature, sexual communication correlated with sexual satisfaction at r = .43 and with relationship satisfaction at r = .37 — among the larger associations anything in this field posts. The internal ranking is the usable part: quality beat both frequency and raw disclosure on both outcomes. Bringing it up more often is not the lever; being able to have the conversation well is. Married samples ran higher (r = .49) than mixed-status ones (.35). The unavoidable caveat is direction — these are cross-sectional correlations, and a good sex life plainly makes the conversation easier as much as the conversation improves the sex. One moderator result is worth publishing precisely because it is so lopsided: among people low in sociosexuality, the association with relationship satisfaction was r = .69 for men and r = .16 for women, a reminder that a pooled coefficient can hide two very different populations. See the outcome gap this sits on top of →

A straight line fits the data at first glance, and that is the version that circulates. Add a squared term and the curve wins: well-being rises with frequency up to roughly once a week and stops rising after that. Two findings sharpen it. Among single people there was no association at all, linear or curved — the effect lives inside relationships, and it runs through relationship satisfaction, which follows the same flattening shape. And the size is not trivial: the authors benchmarked it against money, and the gap between having sex less than monthly and having it weekly was larger than the gap between earning $15,000–$25,000 and earning $50,000–$75,000 a year. What it does not license is a scoreboard. These are cross-sectional self-reports, so a good relationship producing more sex fits the data as well as more sex producing a good relationship; the plateau is an average across tens of thousands of people, not a target for any particular couple; and nothing here says a couple below the average is failing.

Why there are no bars on this card. The source publishes this result as a fitted curve with confidence bands rather than a table of means, so there are no per-frequency values to chart. We would rather print the finding without a picture than build a picture out of numbers the study never reported.

Tier 1 Similarity–attraction: actual similarity predicts attraction at r = .47, perceived similarity at .39, across 313 studies (Montoya et al. 2008). Ruling M-TBD-9 →

Tier 1 Flirting detection: observers correctly recognized flirting only ~38% of the time when it actually occurred, and women’s flirting was read more accurately than men’s (Hall, Xing & Brooks 2014; summary). Rulings M-TBD-1 & M-TBD-2 →

Tier 2 Sexual timing: among 2,035 married individuals, later sexual timing predicted better communication, satisfaction, perceived stability, and sexual quality (Busby, Carroll & Willoughby 2010). Ruling M-TBD-6 →

Tier 1 Stated vs. revealed preferences: the sex differences in the stated importance of looks and earning prospects disappear in live speed-dating behavior (Eastwick & Finkel 2008). Ruling M-TBD-11 →

Tier 2 Sleep & conflict: couples short on sleep behaved more hostilely in conflict discussions and showed a larger inflammatory response to them (Wilson et al. 2017). Ruling M-TBD-10 →

Tier 1 Casual-sex offers: in the Clark–Hatfield paradigm a woman’s direct offer to a male stranger is usually accepted while the identical male offer drew zero female acceptances — but women’s acceptance rises to match men’s when the proposer is attractive, famous, or a trusted friend (Conley 2011). Ruling M-TBD-13 →

Tier 1 Conformity by sex: across a 2024 systematic review of 78 conformity studies only a minority detect any gender effect (64 of 78 didn’t report gender), and a 2023 Asch replication found no sex difference (Capuano & Chekroun 2024). Ruling M-TBD-27 →

Tier 1 Match-rate skew: male Tinder test profiles converted 0.6% of likes into matches vs 10.5% for female profiles — a real attention skew, though 61% of U.S. men aged 25–54 were married or cohabiting in 2019 (Tyson et al. 2016). Ruling M-TBD-31 →

Tier 1 Apps vs. “online”: ~39% of couples who met in 2017 met online (the top channel since ~2013), but only 10% of all partnered adults — 20% under 30 — met on a dating site or app (Rosenfeld et al. 2019; Pew 2023). Ruling M-TBD-32 →

Tier 1 Meeting in college: “met in college” was never a big share — ~9% of couples at its 1995 peak, 4% by 2017, while online rose 2%→39% (Rosenfeld et al. 2019). Ruling M-TBD-43 →

Tier 1 The permanent-single floor: a record 25% of U.S. 40-year-olds had never married by 2021 — yet of those still unmarried at 40 in 2001, about one in four married by 60 (Pew 2023). Ruling M-TBD-35 →

Tier 1 Romance-scam targeting: 52% of online daters have come across a suspected scammer — men under 50 highest at 63% — while FTC median losses climb with age, from $750 (18–29) to $9,000 (70+) (Pew 2023; FTC 2022). Ruling M-TBD-65 →

Tier tags mirror each ruling’s evidence grade on the Mythbuster (hard data → Tier 1, evidence-based → Tier 2). This block grows only when a docket entry earns a graded, sourced ruling — it last grew 2026-07-27 with the romance-scam ruling out of the LE Lab doctrine harvest.