Quotes

Among people quoted by name, women's share differs from their share of all gender-signalled people by −44.73 percentage points over the whole period (95% CI −45.40 to −43.99).

Women's share among people mentioned, and among people quotedBands: 95% article-cluster bootstrap CIs.

Table view

Year	people quoted (named speaker)	people quoted (incl. he/she said)	all gender-signalled people
1900.0	14.0% (11.1%–16.9%) · n=648	13.7% · n=901	43.9% (41.4%–46.2%) · n=10,179
1903.0	20.0% (16.6%–22.9%) · n=636	18.0% · n=881	47.2% (44.6%–49.9%) · n=9,710
1906.0	18.0% (14.8%–21.3%) · n=566	18.8% · n=796	51.2% (49.0%–53.3%) · n=10,076
1909.0	13.4% (10.8%–16.1%) · n=613	14.2% · n=878	51.9% (50.0%–53.9%) · n=10,407
1912.0	20.3% (17.1%–23.4%) · n=644	18.9% · n=903	57.3% (55.4%–59.5%) · n=10,932
1915.0	19.9% (16.1%–23.8%) · n=563	20.6% · n=744	61.7% (59.5%–63.8%) · n=10,758
1918.0	17.0% (13.8%–20.4%) · n=500	16.7% · n=652	62.0% (60.2%–64.1%) · n=9,877
1921.0	21.7% (18.7%–24.9%) · n=649	21.2% · n=848	64.4% (62.8%–65.9%) · n=12,646
1924.0	19.5% (16.4%–22.6%) · n=635	20.9% · n=843	63.4% (61.4%–65.2%) · n=11,238
1927.0	27.6% (24.0%–31.4%) · n=693	25.5% · n=903	67.8% (66.0%–69.5%) · n=12,593
1930.0	24.3% (21.4%–27.5%) · n=645	23.8% · n=841	69.6% (68.1%–71.2%) · n=12,738
1933.0	24.8% (21.4%–28.0%) · n=743	24.8% · n=987	67.5% (65.8%–69.3%) · n=13,605
1936.0	23.0% (19.7%–26.1%) · n=714	23.0% · n=932	68.1% (66.5%–69.7%) · n=13,908
1939.0	18.6% (16.0%–22.1%) · n=752	18.2% · n=937	70.4% (68.8%–71.9%) · n=15,002
1942.0	22.3% (19.0%–25.9%) · n=660	20.7% · n=831	67.3% (65.7%–68.7%) · n=11,785
1945.0	19.3% (16.2%–22.0%) · n=767	17.5% · n=960	65.2% (63.7%–66.9%) · n=11,206
1948.0	17.9% (15.4%–20.5%) · n=929	18.6% · n=1,131	66.4% (64.7%–68.0%) · n=14,117
1951.0	18.8% (16.4%–21.1%) · n=996	17.7% · n=1,233	67.3% (65.6%–69.0%) · n=14,259
1954.0	18.0% (15.2%–20.8%) · n=835	17.2% · n=1,027	68.5% (66.9%–70.2%) · n=13,714
1957.0	14.9% (12.5%–17.6%) · n=878	14.9% · n=1,051	71.5% (70.0%–72.8%) · n=15,212
1960.0	16.2% (13.5%–18.6%) · n=949	16.0% · n=1,163	70.2% (68.7%–71.6%) · n=15,445
1963.0	17.0% (14.6%–19.3%) · n=978	16.2% · n=1,183	65.0% (63.1%–66.7%) · n=12,309

AmericanStories (Dell et al. 2023, CC-BY-4.0), built on Chronicling America (Library of Congress). 10,000 sampled articles per year, 22 years.

Speaker share minus mention share

Attribution	Period	Difference (pp)	95% CI	Speakers	People
name	all	−44.73	95% CI −45.40 to −43.99	15,993	271,716
name	1900-15	−34.82	95% CI −36.36 to −33.78	3,670	62,062
name	1918-30	−43.30	95% CI −44.94 to −41.75	3,122	59,092
name	1933-45	−46.36	95% CI −47.68 to −44.69	3,636	65,506
name	1948-63	−51.16	95% CI −52.41 to −50.07	5,565	85,056
name+pronoun	all	−45.15	95% CI −45.68 to −44.49	20,625	271,716
name+pronoun	1900-15	−35.13	95% CI −36.54 to −33.65	5,103	62,062
name+pronoun	1918-30	−43.76	95% CI −45.43 to −42.25	4,087	59,092
name+pronoun	1933-45	−47.02	95% CI −48.07 to −45.71	4,647	65,506
name+pronoun	1948-63	−51.51	95% CI −52.49 to −50.28	6,788	85,056

Share of gender-signalled people who are quoted at least onceNamed-speaker attribution only.

Table view

Year	women	men
1900.0	2.0% (1.6%–2.5%) · n=4,468	9.8% (8.8%–10.6%) · n=5,711
1903.0	2.8% (2.2%–3.4%) · n=4,582	9.9% (9.0%–10.8%) · n=5,128
1906.0	2.0% (1.6%–2.4%) · n=5,162	9.4% (8.7%–10.3%) · n=4,914
1909.0	1.5% (1.2%–1.8%) · n=5,399	10.6% (9.7%–11.5%) · n=5,008
1912.0	2.1% (1.7%–2.5%) · n=6,267	11.0% (10.0%–12.0%) · n=4,665
1915.0	1.7% (1.3%–2.1%) · n=6,639	10.9% (10.0%–12.1%) · n=4,119
1918.0	1.4% (1.1%–1.7%) · n=6,122	11.1% (10.0%–12.1%) · n=3,755
1921.0	1.7% (1.4%–2.0%) · n=8,150	11.3% (10.4%–12.3%) · n=4,496
1924.0	1.7% (1.4%–2.1%) · n=7,125	12.4% (11.3%–13.4%) · n=4,113
1927.0	2.2% (1.9%–2.7%) · n=8,533	12.4% (11.2%–13.5%) · n=4,060
1930.0	1.8% (1.5%–2.1%) · n=8,867	12.6% (11.5%–13.7%) · n=3,871
1933.0	2.0% (1.7%–2.3%) · n=9,189	12.7% (11.7%–13.7%) · n=4,416
1936.0	1.7% (1.4%–2.1%) · n=9,475	12.4% (11.3%–13.5%) · n=4,433
1939.0	1.3% (1.1%–1.6%) · n=10,565	13.8% (12.7%–15.0%) · n=4,437
1942.0	1.9% (1.5%–2.2%) · n=7,934	13.3% (12.1%–14.4%) · n=3,851
1945.0	2.0% (1.7%–2.4%) · n=7,310	15.9% (14.6%–17.2%) · n=3,896
1948.0	1.8% (1.5%–2.1%) · n=9,375	16.1% (15.0%–17.2%) · n=4,742
1951.0	1.9% (1.7%–2.3%) · n=9,594	17.3% (16.2%–18.4%) · n=4,665
1954.0	1.6% (1.3%–1.9%) · n=9,400	15.9% (14.8%–17.0%) · n=4,314
1957.0	1.2% (1.0%–1.4%) · n=10,881	17.2% (16.0%–18.5%) · n=4,331
1960.0	1.4% (1.2%–1.7%) · n=10,844	17.3% (16.2%–18.4%) · n=4,601
1963.0	2.1% (1.7%–2.4%) · n=7,999	18.8% (17.5%–20.0%) · n=4,310

AmericanStories (Dell et al. 2023, CC-BY-4.0), built on Chronicling America (Library of Congress). 10,000 sampled articles per year, 22 years.

Of quoted people, the share quoted directly (in quotation marks)

Table view

Year	women	men
1900.0	51.6% · n=91	27.6% · n=557
1903.0	49.6% · n=127	32.6% · n=509
1906.0	43.1% · n=102	27.8% · n=464
1909.0	39.0% · n=82	24.5% · n=531
1912.0	36.6% · n=131	38.4% · n=513
1915.0	42.0% · n=112	31.9% · n=451
1918.0	34.1% · n=85	36.4% · n=415
1921.0	45.4% · n=141	34.4% · n=508
1924.0	42.7% · n=124	34.1% · n=511
1927.0	46.1% · n=191	37.3% · n=502
1930.0	44.6% · n=157	34.4% · n=488
1933.0	40.2% · n=184	35.2% · n=559
1936.0	47.6% · n=164	41.6% · n=550
1939.0	23.6% · n=140	35.8% · n=612
1942.0	35.4% · n=147	40.9% · n=513
1945.0	31.8% · n=148	45.4% · n=619
1948.0	24.7% · n=166	38.7% · n=763
1951.0	26.2% · n=187	41.8% · n=809
1954.0	31.3% · n=150	42.6% · n=685
1957.0	32.8% · n=131	40.4% · n=747
1960.0	31.2% · n=154	44.7% · n=795
1963.0	44.0% · n=166	45.2% · n=812

AmericanStories (Dell et al. 2023, CC-BY-4.0), built on Chronicling America (Library of Congress). 10,000 sampled articles per year, 22 years.

What counts as a quotation

A speech event is a verb from a fixed list (say, tell, declare, announce, testify, …) whose grammatical subject is a person mention. Quotation marks alone never count without an attributed speaker. A quotation is direct if the sentence contains quotation marks, or if a short attribution fragment follows a quoted sentence. The secondary rule also credits “he said” and “she said” to the nearest preceding salient person of the matching signalled gender.

How well the rule finds quoted people

Rule	Precision	Recall	F1	Quoted people in reference sample
E_llm	0.87	0.39	0.53	74
quote_rule_name	0.77	0.31	0.45	74
quote_rule_name+pronoun	0.74	0.33	0.46	74