What is GenAI?

Reviews 46 meta-analyses. About 78% of measured gender differences are small or near zero (d ≤ 0.35), including nearly all verbal and communication variables. The framing source for this project.

Men are slightly more talkative (d = −0.14); women use slightly more affiliative speech (d = 0.12); men slightly more assertive speech (d = 0.09). Setting, partner and task moderate the effects more than gender does.

165 studies, 1.4 million participants. Overall verbal-ability difference favoring women is d = 0.11 — negligible. Whatever style differences exist, they are not differences in verbal skill.

113 studies testing Gilligan’s care-vs-justice claim. Care orientation d = 0.28 favoring women, justice d = 0.19 favoring men — small and inconsistent. Gilligan’s framework is kept as vocabulary, not as a measured difference.

Women lead somewhat more democratically and participatively; men more autocratically. The gap is clearest in lab studies and shrinks in real organizations, where role expectations override gender.

Recorded 396 people’s daily speech with wearable microphones. Both women and men spoke about 16,000 words per day; the difference was not significant. Disproves the "women speak 20,000 words, men 7,000" myth.

Women use more hedges, tag questions and disclaimers (d = 0.23). The gap is larger in groups, with strangers and in lab tasks, and near zero in familiar settings — tentative language works as a politeness strategy, not a sign of uncertainty.

205 studies. Women self-disclose slightly more (d = 0.18), mostly to other women and to close partners; the difference nearly vanishes with strangers.

Men make more intrusive (floor-taking) interruptions (d = 0.33), but the overall interruption difference is small (d = 0.15) and depends on group size and setting. Replaces the widely repeated "men make 96% of interruptions" claim, which came from 31 conversations.

Women used more pronouns, emotion words and references to social and psychological processes; men used more numbers, articles, long words, object references and swear words. Effects are consistent but small (most d < 0.3).

About 66,000 users. Women’s language rated warmer, more compassionate and more polite; men’s colder, more impersonal and more hostile. Assertiveness was equal. The "assertive man / tentative woman" split does not survive at scale.

75,000 volunteers, 700 million words. Gender predictable from word choice at ~92% accuracy — proof that aggregate vocabulary differences are real — while individual overlap remains large.

The standard graduate textbook. Frames gendered speech as practice within communities rather than fixed traits; synthesizes the hedging, inclusive-language and collaborative-style literature.

Women used more tentative language when talking with men than with women. Tentative women were more persuasive to male listeners and less persuasive to female listeners — hedging is audience-adaptive.

A linguist’s review of the evidence: popular "different languages" claims are overstated; most measured differences are small, situational and heavily overlapping. The main counterweight in this catalog.

Senate floor records plus experiments: gaining power increases how much men talk but not women, because powerful women who talk more are rated less competent. Floor time tracks power and backlash risk, not a female preference for brevity.

Controlled deliberation experiments: women speak less than men in mixed groups under majority rule, and reach parity only when they are the majority or the group must decide unanimously. Institutions, not dispositions, set who talks.

Catalogs the "gender-linked language effect": women — intensifiers, hedges, questions, emotion references, longer sentences; men — quantity references, directives, judgmental adjectives, elliptical sentences, "I" references. The most concrete feature list available.

New Zealand corpus data on compliments and apologies: women give and receive more compliments and use apologies as social lubricant rather than admissions of fault. Real data, one national corpus.

Standard sociolinguistics textbook; strongest on turn-taking, same-gender talk and the difference between overlapping (supportive) and interrupting (competitive) speech.

Decades of oral-argument transcripts: female justices are interrupted roughly three times as often as male justices, by advocates and by colleagues, even after controlling for seniority.

Longitudinal interviews with about 135 boys: in early adolescence they talk openly about emotional intimacy, then learn to suppress that talk by late adolescence. Lower male emotional disclosure looks learned, not innate.

Introduced "rapport talk" (connection-oriented) vs "report talk" (status/information-oriented). Influential and widely taught, but interpretive: built on transcribed examples rather than measured samples, and criticized for treating tendencies as categories.

The founding text of the field. Proposed "women’s language": hedges, tag questions, intensifiers, super-polite forms. Based on introspection, not data; several claims were later partly confirmed (Leaper & Robnett 2011; Mulac et al. 2001) and others not.

Argued women reason about moral problems through an "ethic of care" and men through an "ethic of justice." Culturally important; empirically weak — the meta-analysis (Jaffee & Hyde 2000) finds only small, inconsistent differences.

Accessible account of the word-count research behind Newman et al. (2008) and Mehl et al. (2007). Reliable, but a popularization — the underlying papers are what the model cites.

A short story told almost entirely in dialogue, in which a couple avoids naming the decision between them. A staple of discourse-analysis teaching for indirectness, hedging and who controls the topic.

The archetype of talk as negotiation of status and relationship. Elizabeth’s indirection and irony against Darcy’s blunt declaratives are the "rapport vs report" contrast two centuries before Tannen named it.

Documentary on how boys learn to suppress emotional talk. Advocacy in form, but it interviews the researchers behind Tier 2 work (Niobe Way, Michael Kimmel). Useful for tone, not for numbers.

Two couples try, and fail, to define love across an evening of drinks. Shows men holding the floor with anecdote and abstraction while the women anchor the talk in specific people.

Provocative TV series pitting gender-studies scholars against biology-leaning psychologists. Illustrates the nature–nurture argument; one-sided in editing and contested by both sides.

A single classroom of about 23 seven-year-olds tries a gender-neutral term. Engaging and illustrative; the sample is far too small to support any claim.

A contemporary study of two people who repeatedly mis-read each other’s under-stated messages. Sets the tone for how small differences in directness compound into misunderstanding.

The same marriage narrated first from the husband’s side, then the wife’s. Dramatizes how the "same" conversation is reported differently depending on who is speaking — the app’s central premise.