Version 0.5, last updated 31 October 2026. For coders, researchers, teachers, journalists and partners.
Dutch Angle (by The Missing Measure) measures how a film divides room between women and men: who speaks, whose story it is, who decides, and how the story treats them. This handbook defines every question precisely enough that two people, or a person and a language model, working independently on the same film arrive at the same score. It describes question set v0.5 with LICHAAM: 20 questions in 10 dimensions that count equally, plus one signal on the film card (K2). E1 (POWER) and the dimension LICHAAM (BODY) with K1 were added after a test on 31 October 2026 (32-proef-nieuwe-vragen/verslag.md); the dimension LIEFDE (LOVE) with R2 and N1 on 16 November 2026 (41-liefde-archief/verslag.md). The tests behind every choice are documented in the project's question review, available on request.
Status. The story judgements (S2, S3, M1, W1) are made twice, blind, for 1,323 films: over 4,972 pairs of answers, 80% differ by at most a quarter point (Krippendorff's α across the archive: S2 0.76, S3 0.88, M1 0.91, W1 0.82; on the first 60 films 0.91 to 0.96). Every question passed a face-validity test on 20 clearly female-led and male-led films. Three outside checks agree with the measurement: bechdeltest.com (Part D), the screenplay dialogue counts of The Pudding (309 films, correlation 0.97 with our share of women's words, on average 4 percentage points apart) and subtitles with speaker names (260 films, 22,675 lines: the character agrees in 71% of the lines, the gender in 90%). Not yet done: a check by human coders, a Lauzen-style hand coding, and the Vrouwen in Beeld coding for Dutch films. BODY and E1 (31 October 2026): E1 was judged twice, blind, for all films (1,362 films with an answer, α 0.93; in 76 films E1 does not count because a lead is not named in the plot). K1 was judged once across the archive (answers in 685 of 1,464 films, from 3 judgements); a second, blind reading of a spread sample of 327 films gave α 0.98 (143 pairs). The K2 signal was read once (the minor category separately, in 232 films with a search hit); on the same sample the second reading agrees on case or no case in 313 of 327 films (α 0.87; on the number of cases 0.92). LOVE (16 November 2026): R2 and N1 were read twice, blind, for 2,337 films; R2 gives an answer in 1,246 films (α 0.91), N1 in 209 films (α 0.99). There is no check by people yet. Where a rule has not been tested, this handbook says provisional.
Part A. Ground rules
A1. Sources
| Source | What we take from it | Used for |
|---|---|---|
Subtitles of the complete film (OPUS OpenSubtitles v2024, one file per film, chosen by choose.py) |
Every spoken line with its timestamps | P2, P3, St2, L1, I2, I3, J2, A5, F1, K1, K2, R2, N1 (candidate lines) |
| Wikidata | Cast list with character names, the actors' gender and date of birth, billing order, release date, director and writers, country of origin, animation | Gender, S1, G1, G3, context |
| Wikipedia plot summary (English; Dutch when there is no English one) | The story as editors summarise it | S1, S2, S3, M1, W1, E1, K2, R2, N1 |
- No images and no sound. We read what is said and what the plot summary says. Screen time and framing (BEELD) are measured only in a pilot and do not count in the angle.
- The subtitles have no speakers. They say what is said, not who says it. Part A3 describes how lines get a speaker.
- Language. A film is measured in its original language when usable subtitles exist (English, Dutch, French, Japanese and others). Otherwise it is measured on the English subtitles, and the site says so. I2 needs a word list per language; where there is none, I2 is counted on the English subtitles.
- The Wikipedia summary is written by editors and can be one-sided itself. That is why every plot judgement shows the sentence it rests on.
- When the Wikipedia summary does not match the film (other names, other events or another ending than in the subtitles), the plot questions (S2, S3, M1, W1, and the plot part of S1) are not measured, with the reason "the Wikipedia summary does not match the film". The film keeps its dialogue questions. The list is in
15-vragen-v05/data/plot_afwijkend.json(decision 31 October 2026; first cases Justice Is Done and Black and White in Color).
A2. Gender
- Gender is never guessed from a voice, a face or a name. It comes from the cast list on Wikidata: the gender of the actor who plays the character.
- When Wikidata gives no gender, or the actor's gender is not the character's, the character's gender is set by hand from how the film itself presents the character (for example the characters played by Elliot Page in Juno and Inception). That choice is marked
gbron: handin the roles file. - Characters who identify otherwise (non-binary, pangender, 'both and neither') get their own value, anders identificerend (
g: x; for example Max in Tár, Victory in My Big Fat Greek Wedding 3, and Saifee, an intersex character, in Bol), only when the film itself presents the character that way. Before 31 October 2026 such characters were left unknown, together with speakers we could not identify; now they are known speakers. Their words count towards the coverage (A3.5), and the cast line names them separately ("and 1 character who identifies otherwise"). They are left out of every women-versus-men comparison, because each formula compares two groups. - The actor's gender never decides the character's. When Wikidata gives an actor as non-binary or trans, the character still gets the gender the film presents (for example the characters played by Elliot Page: Juno, Ariadne and Bliss as women, Sam in Close to You as a trans man). An actor's non-binary entry in Wikidata is not a reason to leave a character unknown (decision 31 October 2026, after Shirin in Lipstick Under My Burkha had been left unknown for that reason). Without a gender in the text, the plot or a gender-matching cast entry, the character stays unknown, as for any other speaker (Rose, the cashier in You've Got Mail).
- Animation: the gender of the character, as the film presents it. The voice actors' ages say nothing about the characters, so the age questions G1, G2 and G3 are not scored for animated films (since 22 October 2026; before that G1 still used voice actors' ages). A film counts as animation only if Wikidata types it as an animated film; live-action films with animated scenes (Mary Poppins) do not.
- We measure gender only. The intersection with skin colour, age, sexuality and disability matters, and we do not guess those identities either.
A3. From subtitles to speakers
- A reader (a language model, in the session that measures the film) reads the subtitles in order and assigns every line to a character from the cast list, or to unknown. It writes these as line ranges in a source file (
17-nieuwe-films/bron/<imdb>.txt, or the annotation files for the first 60 films). - Characters without a name who are clearly a woman or a man (a doctor, a cop, a narrator, an announcer) get a generic role with that gender and
naam: false. They count for presence and conversation, not as a name in I2. The gender may come from the text, or from the plot summary when the plot clearly points to that one person ("the old jailer" who narrates, "the teacher"); this holds for one identifiable person only, never for a group or a whole profession (decision 31 October 2026). - Group scenes. When the text gives everyone who speaks in a scene the same gender, those unnamed lines go to a generic group role with that gender (
Men,Women,Boys,Girls,Soldiers,Nuns;naam: false). The text has to say so: forms of address ("lads", "gentlemen", "ladies", "girls", "sir", "ma'am"), self-references ("my husband", "I'm a father of three") or a setting that the dialogue names (a convent of nuns, a men's barracks). A mixed or unclear group stays unknown. This rule applies from the first reading in every batch; it is not a later repair (decision 23 October 2026, after batch 10 had first left these scenes unknown). A group role holds per scene, and only where the text in that scene fixes the speakers' gender; a whole profession or family is never given a gender on indirect evidence ("guys", "two men went to check") (decision 29 October 2026). The same goes for a scene in which only known characters of the same gender speak but it is not clear which of them says a line: those lines go to the group role. There is no maximum number of characters, as long as the text in that scene fixes their gender (decision 31 October 2026; before that the limit was two or three, for example the four truck drivers in The Wages of Fear and the four friends in Now and Then). A form of address fixes the gender of the people addressed only; the speaker counts too only when the same sentence uses we, us or our about the group ("We have him boys"). Untranslated forms of address ("signori", "ma chérie", "Caballeros") count like their English equivalents. An SDH label with a gender ("MAN:", "WOMAN 2:") fixes the gender of that one speaker for that line only (decisions 10 November 2026). Exceptions decided per film, after batch 21 (14 November 2026): a crew or group that the text once fixes as men ("brother tankies", "OK, Boys") counts as men in all its scenes (T-34); a meeting the dialogue itself calls "men's business" counts as men for the whole meeting (Phase 7); "my girlfriend" without a name counts as evidence for the partner it clearly refers to (Bad Things). These hold for those films only until Astrid decides about the rest of the archive. Decisions 16 November 2026 (after batch 23): "man" and "bro" as a form of address count like "boys"; "Rabbi" and "best man" are gender words like "Father" for a priest; "the kind of guy" said about oneself is a self-reference; untranslated "bhai" does not count. A sleeping woman counts as present in a scene (unlike an unconscious one). - Songs are left out. Credits and on-screen text are not left out separately: in the whole archive they are about 0.03% of the unknown words (measured 23 October 2026), too little to change any film.
- Coverage is the share of lines (and seconds) with a known speaker, that is a speaker who is a woman, a man or identifies otherwise. The aim is at least 0.80 of the words; below it the verdict is provisional. Check each film's coverage right after measuring it, before the next film. Films below it are named in the measurement log (in the current archive: Nomadland 0.75, Past Lives 0.71, whose subtitles contain only the English dialogue).
- While reading, the reader also tags lines that are about work (J2) or about the other as a (possible) partner (L1), and records the main couple (G1).
- Quotes in the evidence must be literal substrings of the line and at most 25 words; a script rejects any that are not.
A4. Who "she" and "he" are
| Term | Definition |
|---|---|
| Women / men | For counts: all characters with lines of that gender, named or not. |
| Named character | A character with a personal name in the cast list or the dialogue. "Waitress" or "Man at bar" is unnamed. |
| Leads | For the story questions (S2, S3, W1, E1): at most two women and two men, chosen by the judge from the cast list on the film card, as the characters the plot summary follows, mentions most and who drive events. Gender comes from the card, never from the judge. If no woman (or no man) in the list plays a real part in the plot, none is chosen for that side. |
| Main couple | The female and male character of the main romantic storyline, recorded by the reader (A3). Used by G1 only. |
Counts cover all characters with lines. Story judgements cover the leads.
A5. Units
| Unit | Definition |
|---|---|
| Line | One subtitle segment with a speaker. |
| Word | A word in a line, split on spaces. |
| Exchange | Two consecutive lines by two different known speakers, the second starting within 30 seconds of the first. The conversation network is the count of exchanges between each pair of characters. |
| Turning point | A decision by a named character without which the plot would go differently. Things that happen to someone, and decisions by unnamed groups, are not turning points. |
A6. Sign and scale
Every question compares women with men and gives a score from −1 to 1: - 0 = equal. - Positive = the men get more room (more words, more of the story, more decisions, better treatment). - Negative = the women get more room. - ±1 = everything on one side.
On the site the sign is mirrored (since 31 October 2026): women are on the right with a plus sign, men on the left with a minus sign, so a film that gives the men more room shows as, for example, −13.9°. The formulas in this handbook keep the internal convention above; only the display is flipped.
A7. Formula families
w refers to women, m to men.
| Family | Formula | Used by |
|---|---|---|
| Share | score = 1 − 2·w/(w + m): one minus twice her share |
P2, P3, S1 |
| Ratio | score = (a − b)/(a + b), where a and b are rates, each relative to its own group, and a is the one that favours the men |
St2, L1, I2, I3, J2, K1 |
| Events | score = (n_m − n_w)/n, over n events, each by a man, by a woman, or shared |
M1, F1 |
| Lead categories | Each lead gets a category worth 1, 0.5 or 0; score = mean of the men − mean of the women |
S2, S3, W1, E1 |
| Judgement | One of five anchors: −1, −0.5, 0, 0.5, 1 (Appendix 1) | A5 |
| Cast | Age difference in years divided by 20, capped at ±1 | G1, G3 |
When one side has no leads. If a film has male leads and no female lead, S2, S3, W1 and E1 score 1 (all room to the men). The reverse gives −1.
A zero denominator is never a score of 0. Below each question's minimum (A8) it is reported as too little data and does not count.
A8. Minimum numbers
| Question | Scored only when |
|---|---|
| P2 | at least 80% of all words have a speaker with a gender |
| P3 | at least 4 characters in the network's top five have a gender |
| St2 | at least 10 same-gender exchanges, and at least 2 expected for each pair type |
| S1 | billing or plot names are available (either part counts on its own) |
| S2, S3, W1, E1 | at least one lead on one side; for E1 all leads must also be named in the plot |
| K1 | at least 50 lines on each side and at least 3 judgements |
| R2 | there is a love line between a woman and a man in which someone is won over, and both runs give an answer |
| N1 | both runs find an ignored no |
| M1 | at least 3 turning points with a woman, a man or both deciding |
| I2 | at least 5 references (names plus role words) on each side |
| I3 | at least 4 named characters of each gender |
| J2, L1 | at least 50 lines on each side |
| G1 | a main couple with both ages known |
| G3 | ages known for speakers on both sides; not for animation |
| A5, F1 | always scored when the film has dialogue (F1 needs at least one evidence quote, also when the score is 0) |
A9. From questions to the angle
- Dimension score = the mean of that dimension's scored questions.
- Angle = 45 × the mean of the dimension scores, rounded to 0.1°. So, internally, −45° is everything her way, 45° everything his way, 0° in balance (on the site mirrored, see A6).
- An angle needs at least 5 of the 10 dimensions. A verdict needs at least 7; with 5 or 6 the film is shown as provisional (voorlopig) and its ball in the archive is see-through. A verdict also needs a speaker with a known gender for at least 80% of all words (the P2 coverage); below that the film is provisional too, with the coverage shown (decision 18 Oct 2026). Before this check, characters without a gender in Wikidata get the gender of the character by hand, recorded as
gbron: hand (personage). - Margin of error = 1.96 × the standard deviation of the dimension scores ÷ √(number of dimensions) × 45, rounded to whole degrees. The site says whether the interval stays on one side of zero ("even at the low end this film leans his way") or crosses it ("the direction is not certain yet"). The margin reflects how much the dimensions disagree, not the difference between coders.
- On the seesaw every scored question is a ball that falls to the side it leans to, further out the stronger it leans. The colour and shape show the dimension.
- The cast is shown separately, as the number of named women and men who speak. It explains a tilt but does not count in it.
- Two pillars. On the site the ten dimensions are grouped in two pillars: Room (WHO TALKS, WHOSE STORY, POWER, TOGETHER) and Treatment (NAME, WORK, AGE, SACRIFICE, BODY, LOVE). Each pillar is the mean of its measured dimensions. The angle stays the mean of all dimensions, so with all ten measured the Room pillar weighs four tenths and Treatment six tenths (since LOVE, 16 November 2026); the pillars do not weigh the same, and presence-driven dimensions weigh less than half, on purpose (Astrid's decision, 31 October 2026, confirmed after BODY was measured: no other formula). When the pillars point opposite ways, the film page says so. A film where women speak less than 10% of the words is labelled a world without women (and the reverse a world without men); the label explains the angle and does not change it.
A10. Verdicts
The verdict follows from the angle, equally far on both sides:
| Angle | Dutch | English |
|---|---|---|
| −5° to 5° | In balans | Level |
| 5° to 15° | Scheef naar hem / haar | Tilted towards him / her |
| 15° to 25° | Flink scheef naar hem / haar | Strongly tilted towards him / her |
| beyond 25° | Zwaar scheef naar hem / haar | Heavily tilted towards him / her |
On the site the verdict is also summed up in three boxes: It's a man's world, It's a shared world and It's a woman's world, with each measured dimension in the box it leans to (shared: at most 5° either way).
- In balans, maar… (Level, but…): the angle is at most 5°, but only because dimensions cancel each other out. Among dimensions with at least 2 scored questions, the strongest leans at least 20° one way and the strongest the other way at least 15°.
- Voorlopig (provisional): fewer than 7 dimensions (A9).
- The explanation under a verdict speaks of room ("the story clearly gives the men more room than the women"), never of roles or quality. The site always adds: the angle says nothing about whether a film is good.
A11. Who judges, and how often
| Kind | Questions | Who | How it is checked |
|---|---|---|---|
| Counted | P2, P3, St2, S1, I2 | Code | Deterministic; depends on the speaker attribution (A3) |
| Cast data | G1, G3 | Code, from Wikidata | Dates as on Wikidata on the day of fetching |
| Plot judgement | S2, S3, M1, W1, E1 (and the K2 signal) | A judge who reads only the film card: title, year, the Wikipedia plot and the cast list. No scores, no other judgements, no memory of the film. | Two runs, the second blind to the first, averaged (1,323 films; E1: all 1,464 films; see Status). K2: one run, plus a second on the sample of 327 films. Where only one run exists, the site says so. |
| Dialogue judgement | A5, I3, F1, and the J2/L1 line tags | The reader of the subtitles (A3) | Quotes checked literally; no second run yet in v0.5 |
| Candidate-line judgement | K1 | A reader who sees only the candidate lines a word list found (02-prompts/lichaam/woorden.py), each with two lines of context, and the speaker list |
One run across the archive; a second, blind run on a spread sample of 327 films (α 0.98) |
| Love reading | R2, N1 | A reader who sees only the film card: the cast, the fixed leads, the Wikipedia plot and candidate lines from the subtitles (words about courting, kissing and marrying) | Two runs, the second blind to the first; quotes checked literally; scored only if both runs give an answer |
A12. Evidence
Every scored question shows what it rests on: for subtitle questions a quote with its timestamp, for plot questions the sentence from the plot summary (prefixed Plot:), for counts the numbers behind the score. A quote that does not literally appear in its source is rejected. A judgement without evidence is not shown.
A13. Status labels on the site
| Label | Meaning |
|---|---|
| Telling / Count | Counted by code from the subtitles or the cast list |
| Castgegevens / Cast data | From Wikidata |
| Oordeel uit de plotsamenvatting (Wikipedia) | Plot judgement; with twee blinde beoordelaars or één beoordelaar (A11) |
| Oordeel van het taalmodel / Language-model judgement | Judged from the subtitles, with quotes |
| Niet gemeten / Not measured | No score; the reason is given (too little data, not applicable) |
Part B. The 20 questions
The dimension names are the ones on the site since 31 October 2026, named after the question they ask (the earlier names in brackets, which the code still uses): WHO TALKS / WIE PRAAT (AANWEZIG), WHOSE STORY / WIENS VERHAAL (VERHAAL), POWER / MACHT, TOGETHER / ONDER ELKAAR (STEM), NAME / NAAM (IDENTITEIT), WORK / WERK, AGE / LEEFTIJD, SACRIFICE / OFFER (LOT), BODY / LICHAAM (since 31 October 2026), LOVE / LIEFDE (since 16 November 2026). The headings below use the earlier names.
AANWEZIG: who is there, and who gets the words?
P2. Who gets the words?
- Question: what share of the spoken words do women speak?
- Counts: every word in a line with a speaker of known gender. Doesn't count: lines with an unknown speaker, songs.
- Formula: Share, on words:
1 − 2·w/(w + m). - Minimum: at least 80% of all words have a gendered speaker (coverage), otherwise not scored.
- Type: count.
- Inspiration: Anderson and Daniels (2016), Film dialogue from 2,000 screenplays, broken down by gender.
- Why words, not seconds: subtitle timestamps say when a line is shown, not how long someone speaks; words are a steadier measure.
- Weak points: depends entirely on the speaker attribution. A film with many unnamed voices (crowds, interviews) loses coverage.
P3. Who is central?
- Question: of the five characters who talk with others the most, how many are women?
- Counts: the conversation network (A5): every exchange between two different known speakers within 30 seconds. A character's weight is the number of exchanges they take part in.
- Formula: Share over the top five by weight:
1 − 2·(women in top five)/(women + men in top five). - Minimum: at least 4 of the top five have a gender.
- Type: count.
- Inspiration: Kagan, Chesney and Fire (2020), on gender in film character networks.
- Weak points: five is a small number, so the score moves in steps of 0.4. Consecutive lines are a proxy for "talks with": a line can answer someone other than the previous speaker.
VERHAAL: whose story is it?
S1. Who comes first, in the credits and in the plot?
- Question: are the first-billed characters, and the characters the plot summary is about, women or men?
- Counts: (a) billing: the first four cast members with a known gender in Wikidata's order; (b) plot: every mention in the plot summary of a character's name, using only name parts that belong to characters of one gender (so a shared family name such as "Corleone" is not counted).
- Formula: the mean of the two Share scores,
1 − 2·w/(w + m)for billing and for plot mentions. If only one part is available, that part. - Type: count.
- Note: S1 leaves out the share of dialogue, because P2 measures that.
- Weak points: Wikidata's order is not always the official billing. Characters referred to without a name in the plot ("his wife") are not counted.
S2. Who has a goal of their own?
- Question: for each lead, does the plot give the character an identifiable goal, and what kind?
- Categories:
own= 1: a goal that is not about a romantic partner (a mission, work, survival, revenge, freedom, an ambition, protecting someone who is not a partner).partner= 0.5: the goal is mainly about winning, keeping or pleasing a romantic partner.none= 0: no identifiable goal; the character mainly reacts or supports others. - Formula: Lead categories: mean of the men − mean of the women.
- Type: plot judgement, with a quote from the plot per lead.
- Inspiration: Lauzen's coding of protagonists' goals; the Mako Mori test (2013).
- Weak points: goals tangled up with a love story are a judgement call (Casablanca, Legally Blonde); the judge writes down the reasoning in one sentence.
S3. Who is needed?
- Question: if this lead were removed from the plot summary, would the story still work?
- Categories:
essential= 1: the story falls apart; the character drives it or is its centre.important= 0.5: the story changes noticeably, but the main line survives.replaceable= 0: someone or something else could take the place (a prize, a motive, a lamp). - Formula: Lead categories: mean of the men − mean of the women.
- Type: plot judgement, with a quote per lead.
- Inspiration: Kelly Sue DeConnick's Sexy Lamp test; asked here for both sides.
- Weak points: "would the story still work" depends on how detailed the plot summary is.
MACHT: who decides, and who has the money?
M1. Who makes the turning-point decisions?
- Question: list the turning points (A5) in the plot; who makes each decision?
- Counts: choices a named character makes: to leave, fight, lie, accept, refuse. Each is by a woman, a man, or shared when a woman and a man decide together. Doesn't count: things that happen to someone (an accident, an attack), and decisions by groups without a named character.
- Formula: Events: (decisions by men − decisions by women) ÷ all decisions, shared ones counting in the total only.
- Minimum: at least 3 turning points.
- Type: plot judgement, with a quote per turning point. When the judge's gender label is missing, it comes from the speaker list of the subtitle pipeline, never from the judge.
- Weak points: a summary that lists many small decisions weighs differently from one that lists a few big ones.
A5. Who has the money?
- Question: who owns, earns and controls the money, and who depends on it?
- Counts: what the dialogue says about income, wealth, debts, paying and buying between the main characters.
- Formula: Judgement: 1 = he has and controls the money and she depends on it; 0.5 = he has more and it matters; 0 = equal, or money plays no role between them; −0.5 and −1 the other way round.
- Type: dialogue judgement, with a quote for the value and one for the strongest argument the other way.
- Example: Sunset Boulevard: −1, Norma is rich and keeps Joe entirely.
- Weak points: class and gender are mixed up in this question.
E1. Who is the expert?
- Question: for each lead, does the plot present the character as someone with knowledge or authority that others rely on?
- Categories:
expert= 1: others ask for, follow or need this character's knowledge, advice or authority (a skill, a plan that works, a leader whose judgement counts, the only one who knows how).partly= 0.5: shows knowledge or skill, but the story does not rely on it, or others do not take it seriously.no= 0: the character follows, learns or is helped. - Formula: Lead categories: mean of the men − mean of the women.
- Type: plot judgement, two blind runs over all films, with a quote per lead (
02-prompts/lichaam/lezer_plot.md). Across the archive: answers in 1,362 of 1,464 films, α 0.93 (0.95 on the first 60 films). - Not named: when a lead is not named in the plot, E1 does not count for that film (missing rather than 0; Astrid's decision, 31 October 2026). Named means: by name, a short form, nickname or other spelling, or by the role in the cast list ("his mother" for Mother). Only a relation while the cast list gives a name ("his wife" for Anika), or a general description, counts as not named. Code searches the names (
33-lichaam/scripts/genoemd.py); a reader decides doubtful cases (02-prompts/lichaam/lezer_e1_genoemd.md,33-lichaam/uit/e1_genoemd.json). This removes 76 films. - Not the same as W1: E1 is about knowledge and authority, not about having a job. In the test (60 films) E1 and W1 correlated 0.65 per film and 0.50 per lead.
- Test (31 October 2026): α 0.91, answers in 58 of 60 films, falls her way in 6 of the 10 female-led films, separation +0.64.
- Weak points: for leads with little plot text it is an estimate.
STEM: who talks with whom, and about what?
St2. Do women talk with each other as often as men do?
- Question: how often do two women talk with each other, and two men, compared with what you would expect from how much each group takes part in conversations?
- Counts: exchanges in the network (A5) between two women (ww) and between two men (mm). Each gender's share of all exchange-ends, s, gives the expected numbers if people mixed at random:
T·s²for woman-woman andT·(1−s)²for man-man, where T is the number of exchanges. - Formula: Ratio of observed over expected:
a = mm/expected_mm,b = ww/expected_ww,score = (a − b)/(a + b). - Minimum:
ww + mm ≥ 10, and both expected numbers at least 2. - Type: count.
- Inspiration: the Bechdel-Wallace test (1985), made continuous and two-sided; correcting for the cast means a film with few women is not punished twice.
- Weak points: like P3, depends on consecutive lines as a proxy for conversation.
L1. Who talks about the other as a partner?
- Question: what share of women's lines is about a man as a (possible) partner, compared with the share of men's lines about a woman as a partner?
- Counts: lines tagged by the reader (A3) whose main topic is a person of the other gender in a romantic, sexual or marital role. Doesn't count: a man mentioned only in a work context, or lines addressed to someone but not about them.
- Formula: Ratio:
a= share of women's lines tagged,b= share of men's lines tagged,score = (a − b)/(a + b). - Minimum: at least 50 lines on each side.
- Type: dialogue judgement (line tags), counted by code.
- Inspiration: the third step of the Bechdel-Wallace test, asked both ways.
- Weak points: in a romance both leads talk about each other a lot, so the score stays near 0 while the story can still revolve around him; S2 covers that.
IDENTITEIT: who is someone in their own right?
I2. Called by name or by a role?
- Question: how often are women referred to by a role, a relationship or an endearment ("sweetheart", "the girl", "his wife"), compared with men ("buddy", "son", "her husband")?
- Counts: in all lines, mentions of the names and aliases of named characters (from the roles file), and words from the role lists (Appendix 2), each list belonging to one gender. Doesn't count: pronouns.
- Formula: Ratio:
a= role words for women ÷ (role words + names for women),b= the same for men,score = (a − b)/(a + b). - Minimum: at least 5 references (names plus role words) on each side.
- Type: count, with word lists for English, Dutch and French; other languages are counted on the English subtitles.
- Weak points: endearments can be affectionate both ways; "man" and "boy" are also ordinary words. Shared family names are removed from the aliases so they do not count for both.
I3. Whose relationship status is mentioned?
- Question: for which named characters is it said whether they are married, engaged, single, divorced, widowed, "taken" or "available"?
- Counts: per named character: the status is mentioned at least once, by themselves or by others. Doesn't count: a romance without the status being named.
- Formula: Ratio:
a= share of named women whose status is mentioned,b= the same share for men,score = (a − b)/(a + b). - Minimum: at least 4 named characters of each gender.
- Type: dialogue judgement, per character, with a quote.
- Inspiration: Lauzen's It's a Man's (Celluloid) World, which tracks marital status by gender.
- Weak points: like G3, it measures treatment rather than who leads, so even in films about women it rarely falls her way. Small casts make the shares jump.
WERK: whose work matters?
W1. Whose work matters to the story?
- Question: for each lead, does the character have work, a profession or a calling, and does it matter to the story?
- Categories:
central= 1: the work drives the plot or is the arena where the story happens.present= 0.5: the work is named or shown, but the story does not depend on it.none= 0: no work, or work is never mentioned. - Formula: Lead categories: mean of the men − mean of the women.
- Type: plot judgement, with a quote per lead.
- Inspiration: Lauzen's coding of occupations.
- Weak points: often +1 in male-led films, because the plot names the men's work and not the women's; J2 gives the second view from the dialogue.
J2. Who talks about work?
- Question: what share of men's lines is about work, a career, study or a craft, compared with women's?
- Counts: lines tagged by the reader (A3) whose main topic is the speaker's own or someone's work, study or professional skill. Doesn't count: money in general, housework.
- Formula: Ratio:
a= share of men's lines tagged,b= share of women's lines tagged,score = (a − b)/(a + b). - Minimum: at least 50 lines on each side.
- Type: dialogue judgement (line tags), counted by code.
- Weak points: in workplace films the line between work and life blurs.
LEEFTIJD: who may grow older?
G1. Who is older in the main couple?
- Question: how much older is the man in the main couple than the woman (the actors' ages at release)?
- Formula: Cast:
(age_him − age_her)/20, capped at ±1. Red flag: a gap of more than 10 years either way is flagged on the site (full years, age on the day of release; Astrid's decision, 18 November 2026; was 5); the flag is a signal, not a separate score. - Minimum: a main couple (A4) with both dates of birth on Wikidata; otherwise not applicable. If the first couple is of the same gender, the next couple of a woman and a man counts; if there is none, not applicable (since 18 November 2026). If the part has a different name on Wikidata, the actor is linked through the English Wikipedia cast list, only when that is unambiguous (since 18 November 2026). The main couple is the most important romantic storyline. A marriage that is only background does not count (as in R2), and desire for a minor is not a romance: that belongs to K2, not G1 (Astrid's decision, 19 November 2026; Lost in Translation: Bob and Charlotte, Quantum of Solace: Bond and Strawberry Fields).
- Type: cast data. Provisional: the divisor 20 has not been calibrated.
- Example: Pretty Woman: Richard Gere 40.6, Julia Roberts 22.4 at release: 18.2 years, score 0.91.
- Weak points: the actors' ages, not the characters'. Only heterosexual couples can be compared this way; a same-gender couple is not applicable.
G3. How old are the women and men who speak the most?
- Question: the mean age at release of the three women and the three men with the most lines, compared.
- Formula: Cast:
(mean age of the men − mean age of the women)/20, capped at ±1. - Minimum: ages known on both sides; not for animated films (voice actors' ages say nothing about the characters).
- Type: cast data.
- Inspiration: Lauzen's findings on age: on screen, women are younger than men.
- Note: G3 almost never falls her way, even in films about women. That is the finding, not a fault: it measures treatment, not who leads. It covers films without a romance, where G1 is empty.
- Weak points: actors' ages, not characters'; a character played at two ages counts once.
LOT: how does the story treat them?
F1. Whose suffering is used for the other's story?
- Question: is a character hurt, killed or traumatised mainly to motivate or change a character of the other gender? Look both ways.
- Counts: violence, death or loss suffered by one character, after which the plot follows the reaction of a character of the other gender. Each case records to (the gender of who suffers) and by (the gender of the character whose story the suffering serves, whose reaction the plot then follows); a case the plot depends on counts double. By is never the perpetrator: who caused the suffering does not matter. Doesn't count: suffering that drives the sufferer's own story; suffering that serves the story of someone of the same gender; both genders suffering together.
- Formula: Events: (weight of women's suffering used for men's stories − weight of men's suffering used for women's stories) ÷ total weight; 0 when there are no cases.
- Type: dialogue judgement, with at least one quote, also when the score is 0.
- Inspiration: "women in refrigerators", Gail Simone (1999), asked both ways.
- Example: Gladiator: the murder of Maximus's wife and son drives his revenge, a central case on her side.
- Weak points: "mainly to motivate" requires reading the story's structure. LOT has only one question. A second was tested on 31 October 2026 (F3 harmed or harming, F4 who dies or disappears) and dropped: F3 answers too rarely and F4 mainly measures genre.
- Correction (31 October 2026): in part of the archive the reader had filled in the perpetrator as by, so that a woman's suffering in her own story (or in another woman's story) counted as serving a man's story; this mostly hit films about women who undergo violence. All F1 cases in the archive were reread with the rule above and corrected (removed, or by set right); the prompt now says explicitly that by is not the perpetrator.
LICHAAM: who is judged on their body?
K1. Who is judged as an object of desire?
- Question: how often are women judged on their looks or sexual availability, compared with men, relative to how often each is addressed?
- Counts: a remark that rates someone on looks or sexual availability, positive or negative ("would you do her", "she's a 10", "nice ass", "he's hot", "she's easy", "fat", "ugly", "you look beautiful"). Code finds candidate lines with a word list (English, Dutch, French); the reader decides per line whether it is such a judgement, about whom, and of which kind (looks or sexual). Fixed choices for the borderline cases are in
02-prompts/lichaam/lezer_ondertitel.md. Doesn't count: song lyrics, children and animals, clothes alone, terms of endearment used as a form of address, a fixed courtesy ("the lovely Katniss"), general statements that judge nobody, "hooker" as a plain job description. - Formula: Ratio:
a= judgements of women per 100 lines addressed to women,b= the same for men,score = (a − b)/(a + b). A line is addressed to the previous speaker (within 30 seconds) when that is a different character, otherwise to the next. A judgement in a line without a known speaker counts only when it concerns a named character (so songs and voices drop out). - Minimum: at least 50 lines on each side and at least 3 judgements (
k1_min_oordelenindrempels.json; decision 31 October 2026). With a minimum of 1, a film with a single remark jumped to ±1 and pulled the whole angle by up to 9°. - Type: candidate-line judgement, one run across the archive; a second, blind run on a spread sample of 327 films tests the agreement (decision 31 October 2026). Result: answers in 685 of 1,464 films; α 0.98 on 143 pairs from the sample. Another 387 films have one or two judgements and get no K1.
- Inspiration: USC Annenberg (sexualisation); Mulvey (1975), the male gaze, here measured in what characters say.
- Test (31 October 2026): α 0.86, answers in 46 of 60 films; all nine broad comedies in the test (American Pie, Porky's, There's Something About Mary and others) lean clearly his way (+0.55 to +1.00). Like G3 and I3 it measures treatment: even in films about women, women are judged on their looks more often (2 of 10 female-led films her way, 9 of 60 films overall).
- Weak points: a judgement without a signal word ("look at her in that dress") is missed; with few judgements the score stays coarse (3 judgements on one side give ±1). LICHAAM has one question.
LIEFDE (LOVE): who has to be won over?
LOVE was added on 16 November 2026 (Astrid's decision), after three tests on the test films (40-proef-liefde/) and a measurement across the whole archive (41-liefde-archief/verslag.md). The definition has not yet been checked by people; a manual rating pack is ready.
R2. Who is won over?
- Question: in a love line between a woman and a man, who is the object of the conquest? Who is wanted, pursued, courted, chosen, won or given away, and who wants, chases and wins?
- Counts: a romantic or sexual storyline between two named characters that the plot follows, with at least one lead. Per line: she is won over (+1), he is won over (−1), mutual (0). A couple that already exists and stays together does not count. Four fixed choices: one declaration of love does not make someone the conqueror; whoever chooses between suitors is the object; letting the other go is not conquering; in a seduction as a trick or an assignment, the object is the one seduced. Does not count: a passing flirt, a marriage that is only background, sexual violence or coercion (K2), contact between a minor and an adult at least 3 years older or in authority (abuse, K2), someone held as property, and lines between two characters of the same gender or of unknown gender.
- Formula: mean over the lines, the main line counting double. The score is the mean of the two runs, and only if both give an answer.
- Kind: love reading (A11), two runs, the second blind to the first (
41-liefde-archief/WERKWIJZE.md). Result: answers in 1,246 films, α 0.91; R2 leans his way in 839 films, her way in 243, and is 0 in 164. - Inspiration: the woman as the hero's prize.
- Test (15 and 16 November 2026): three runs on the test films. Two earlier questions (is love her highest goal, does the story end with the couple) were dropped; this definition gave α 0.96.
- Weak points: in 781 of the 1,246 films R2 rests on one line, so a single side line moves the angle by ±4.5° (a tenth of 45°), sometimes more. The readers often hesitate, especially with short plots. A same-gender couple does not count, its side lines with men do (Carol).
N1. Whose no is ignored?
- Question: in a love line between a woman and a man, does one clearly reject the other, after which the other keeps courting or pressing anyway?
- Counts: her no +1, his no −1. If the persistence ends in a relationship, a kiss or a marriage between the two, it counts double. Does not count: a no that is respected; coercion, violence and assault (K2).
- Formula: weighted mean over the cases. Without a case N1 does not apply (not 0).
- Minimum: a case counts only if both runs find it.
- Kind: love reading, the same two runs as R2. Result: answers in 209 films, α 0.99; N1 leans his way in 176 films.
- Inspiration: persistence presented as romance (Lippman 2015).
- Weak points: it is rare, and a no that is only shown on screen is missed.
Part C. Outside the angle
Shown as context, not counted
- Who speaks: the number of named women and men with lines, and the share of women.
- Who made it: the gender of the director and the writers (Wikidata).
- Striking figures in the archive: for the selected films, the gender of the lead (the character who speaks the most words; groups such as "Men" or "Crowd" do not count, so a silent lead is not counted as lead), the share of words spoken by women, the gender of directors and writers, and the age gap in the main couple (G1).
- Where and when: the archive filters on country of origin (Wikidata P495; a co-production counts in each of its regions) and on period. These say where a film comes from, not how it is measured.
Signal on the film card: K2, sexual misconduct
- What: drugging (to enable sex), sexual assault, rape, sexual coercion (pressure, blackmail, deceiving someone about who you are, someone drunk or unconscious; forcing someone to undress always counts as coercion, whatever their age, decision Astrid 21 November 2026), unwanted touching, voyeurism (secretly watching or filming someone undressing, showering or having sex, or spreading such images), and abuse of a minor by an adult, also when the film presents it as consensual (category minderjarige, Astrid's decisions of 31 October and 2 November 2026): sexual contact between someone the card or subtitles show to be under 18 or in secondary school and an adult who is at least 3 years older or has authority over the minor (teacher, pastor, coach, parent or step-parent, boss, guardian); if the age gap is not stated, it counts when the film clearly shows the other person as an adult in another stage of life (a job, married, an own household) or with authority. Two minors together do not count. It counts when the film clearly shows or says that the sexual contact happens, not only when it is spelled out word for word; a teacher kissing a pupil counts as sexual contact (Astrid's decision 19 November 2026, Mean Girls); a case already counted as rape or coercion is not counted again; it is always a normal case, never a joke (
02-prompts/lichaam/lezer_minderjarig.md). Sex between an adult and a minor is always called abuse, never 'sex'. A minor with an adult who is less than 3 years older and has no authority, such as a 17-year-old with an 18- or 19-year-old, is called a questionable relationship (category dubieus, Astrid's decision 18 November 2026: a minor with an adult is always problematic; until then it did not count). The same holds when the partner is an adult but it is not established that the gap is 3 years or more or that there is authority (Astrid's decision 2 November 2026), and when it is not established that a pupil is still a minor (final year of secondary school, or a classmate is already 18), even with authority (Astrid's decision 21 November 2026). A questionable relationship counts in the K2 signal, is shown separately ('of which … a questionable relationship', like the jokes) and is never called abuse. Attempts count, and other K2 cases also count when played as a joke. - Source: the plot summary and candidate lines from the subtitles, with a quote per case; one run across the archive, a second on the sample of 327 films (
02-prompts/lichaam/lezer_plot.md). Result: 455 cases in 319 films, 50 played as a joke; it happens to a woman 361 times and to a man 77 times. Of these, 25 cases of abuse of a minor by an adult in 19 films (among them Het debuut, Fast Times at Ridgemont High, Trainspotting, An Education, Call Me by Your Name). Documentaries keep their K2 cases about real events; they are not shown on the site. On the sample the second reading agrees on case or no case in 313 of 327 films (α 0.87). - Shown as: "sexual misconduct, x times, of which y played as a joke", with the quote and who it happens to.
- Why not a ball: in the test it was stable (α 0.86), but most films have no case, it almost never falls her way (1 of 60 films), and it does not separate female-led from male-led films. As a ball it would pull LICHAAM towards 0 in most films. It leans his way in six of the nine broad comedies tested.
Not measured yet
| Question | Why not | What it needs |
|---|---|---|
| M2 power and agency in the language (Sap et al. 2017) | Too few verbs in plot summaries; on dialogue it does not separate | Script action lines |
| I1 defined through a partner ("his wife") | Too few films with enough cases | More text, or scripts |
| F3 who is harmed, who harms (risk ratio) | Plot summaries rarely describe violence between named women and men in detail | Scripts, or dialogue judgement |
| F2 punished for what the other gets away with | Does not separate clear cases | Rework |
| W2 care work | Answers in 22 of 59 films; mostly reflects that women are present | Dialogue or script action lines |
| BEELD screen time and close-ups | Needs a lawful copy of the film and a blind hand check | Decision by the project lead |
Part D. Checked against bechdeltest.com
Why the Bechdel test is not a ball. Its idea is already in the angle, made continuous and two-sided: St2 (do women talk with each other as often as men do, corrected for the cast) and L1 (who talks about the other as a partner, the test's third step). Adding the test itself would count that idea twice, and the test is a yes/no question about women only, with ratings entered by the site's visitors. So we use it the other way round: as an outside check on the angle.
How. The ratings (0 to 3; 3 = two named women talk to each other about something other than a man) come from bechdeltest.com, read from the site's year lists in October 2026 (its old API answers "410 Gone"). 243 films in the archive are listed; 221 of them have a verdict (not provisional) and are compared below. Ratings marked dubious by the site are kept. Angles in this part use the internal sign: positive is his way.
Result.
| Bechdel rating | Films | Mean angle |
|---|---|---|
| 3 (passes) | 161 | 2.8° |
| 2 (women talk, but only about a man) | 17 | 12.5° |
| 1 (women do not talk to each other) | 33 | 18.5° |
| 0 (fewer than two named women) | 10 | 16.8° |
- Films that fail lean 13.7° further his way than films that pass (95% interval 11.0° to 16.3°, bootstrap). The angle goes down with the rating (Spearman −0.57).
- Of the 36 films that lean her way (more than 5°), 35 pass; of the films in balance 90% pass; of the films that lean his way, 58%.
- On the first 129 films the difference stayed almost the same without the STEM dimension (14.1° against 14.8°), so the agreement does not come only from the questions that resemble the test.
What this does and does not show. The angle and a well-known, independent test point the same way, and the angle adds what the test cannot: degrees, both directions, and how much room a film gives, not only whether two women talk. It does not show that the angle is right film by film; a Bechdel pass is a low bar (161 of 221 films pass).
Part E. Known limitations
- The speaker attribution is done by a language model. It is checked by coverage, by quotes and, where they exist, by subtitles with speaker names (Status), not yet by a human coder on a sample.
- The plot summaries are Wikipedia's. They are free and cover almost every film, but they are written by editors with their own focus. Films without a summary get no plot questions (Spijt!).
- The plot judge is a language model too. The second run is blind to the first, but both are the same kind of judge; people check samples on an ongoing basis, since 19 November 2026.
- The archive is a choice, not a sample. Its average says something about these films, not about film in general.
- Gender only, and binary in the comparison. Non-binary characters are left out of the formulas, not reassigned.
- Outside checks so far: bechdeltest.com (Part D; also via the FiveThirtyEight data, CC BY 4.0, and the 2021 snapshot via TidyTuesday), The Pudding's screenplay counts, subtitles with speaker names (Status), Hollywood Age Gap (age in the main couple), the Cornell Movie-Dialogs Corpus and MovieSum (CC BY-NC 4.0) for the share of women's words, and the power and agency word list of Sap et al. (2017; riveter-nlp, GPL-3.0). That last one barely relates to the angle (r 0.11). The figures are on the site under How we measure (November 2026). None yet with human coding.
- The method changes. Scores can move when a rule is improved. Cite a result with the date you read it.
Appendix 1. Anchors for judgements (A5, F1)
| Value | Meaning |
|---|---|
| 1 | Clear and central: the pattern works against the women and drives the story, with no counterweight |
| 0.5 | Present but limited, or contradicted by other scenes |
| 0 | Absent, or equal on both sides |
| −0.5 | The reverse pattern is present but limited |
| −1 | The reverse pattern is clear and central |
Always one quote for the chosen value and one for the strongest argument the other way.
Appendix 2. Role and endearment words (I2)
The lists in 13-ondertitels/scripts/counts_film.py (ROLE) are the reference. Each word belongs to one gender, for example:
- EN, women: sweetheart, honey, baby, darling, girl, lady, miss, ma'am, wife, girlfriend, mother, daughter, sister, princess. Men: buddy, pal, son, boy, guy, sir, husband, boyfriend, father, brother, dude, prince.
- NL, women: schatje, lieverd, meisje, meid, mevrouw, vrouw, moeder, dochter, zus, vriendin. Men: jongen, meneer, vent, kerel, vader, zoon, broer, vriend, echtgenoot, maat.
- FR, women: chérie, mademoiselle, madame, fille, femme, mère, copine. Men: chéri, monsieur, garçon, homme, père, fils, copain, mec.
Appendix 3. Where the code is
The code is not public yet; it is available on request for research. For reference:
| Step | Script |
|---|---|
| Subtitles chosen and turned into transcripts | 13-ondertitels/scripts/choose.py, build_transcripts.py |
| Speakers and tags written per film | 17-nieuwe-films/scripts/lees.py, film.py (new films); annotation files for the first 60 |
| Counts and cast questions (I2, G1, coverage) | 13-ondertitels/scripts/counts_film.py, measure_film.py |
| P2, P3, St2, S1, G3 | 15-vragen-v05/scripts/auto_v05.py |
| Film cards for the plot judges | 15-vragen-v05/scripts/kaarten.py |
| Plot judgements checked and saved | 17-nieuwe-films/scripts/plotoordeel.py (prompt: 15-vragen-v05/prompts_v05.md) |
| Dimension scores and angle | 15-vragen-v05/scripts/hoek_v05.py, export_v05.py; the site recomputes the angle in dutch-angle-site/src/index.html (angleOf) |
| LOVE (R2, N1) | 41-liefde-archief/scripts/prep.py (cards), score_archief.py, export_liefde.py |
| Country of origin for the filters | dutch-angle-site/scripts/landen.py |
Appendix 4. Sources and inspiration
The classic tests are credited as inspiration only. Dutch Angle uses its own questions. - Anderson, H. & Daniels, M. (2016). Film Dialogue from 2,000 screenplays, Broken Down by Gender and Age. Polygraph / The Pudding. Summary: https://flowingdata.com/2016/06/27/gender-equality-in-the-movies-a-screenplay-analysis - Kagan, D., Chesney, T. & Fire, M. (2020). Using data science to understand the film industry's gender gap. Palgrave Communications 6. https://arxiv.org/abs/1903.06469v1 - Lauzen, M. (2026). It's a Man's (Celluloid) World: Portrayals of Female Characters in the Top Grossing U.S. Films of 2025. Center for the Study of Women in Television and Film, San Diego State University. https://womenintvfilm.sdsu.edu/wp-content/uploads/2026/02/2025-Its-a-Mans-World-Report.pdf - Lippman, J. R. (2018; online 2015). "I did it because I never stopped loving you": the effects of media portrayals of persistent pursuit on beliefs about stalking. Communication Research. - Sanders, W. & Becker, S. (2025). 'Gewoon' een Nederlandse vrouw. Universiteit Utrecht / Stichting Vrouwen in Beeld. https://www.vrouweninbeeld.nl/content/uploads/2025/10/Onderzoek-REPRESENTATIE-in-Beeld_251015.pdf - Sap, M., Prasettio, M. C., Holtzman, A., Rashkin, H. & Choi, Y. (2017). Connotation Frames of Power and Agency in Modern Films. Proceedings of EMNLP 2017, 2329–2334. https://aclanthology.org/D17-1247/ - bechdeltest.com: crowd-sourced Bechdel ratings, https://bechdeltest.com (read 17 October 2026). - Inspiration: the Bechdel-Wallace test (1985), the Mako Mori test (2013), the Sexy Lamp test (Kelly Sue DeConnick), "women in refrigerators" (Simone, 1999), the Smurfette principle (Pollitt, 1991). - Data: OPUS OpenSubtitles v2024 (opus.nlpl.eu); Wikidata; Wikipedia plot summaries (CC BY-SA).