# Method ## The base `data/journalists.csv.gz` holds 173,133 classified editorial rows. The published base is the 127,602 rows with `platform_domain == 0` and `media_typed == 1`: 12,540 rows are dropped because the vendor stored a link-in-bio, social or freelance-marketplace URL as the company website, and 32,991 because the outlet's own record does not type it as a newspaper, broadcaster, online outlet, wire agency or magazine. Every figure in this dataset is computed over all 127,602 rows, not a sample. The wider classification, the title lexicon, the roles, the editor kinds and the source tables are documented in `data/METHOD.md` in the repo, and the lexicons themselves are shipped here as `classifier_rules.txt`. The person rows come from `data.blitz_people` (Mongo, read-only), the same table `data/METHOD.md` names, pulled by `tools/pull_people.py` and `tools/pull_people_new.py` and classified by `data/classify.py`. ## The collection window The rows were not fetched in one pull. `fetched_at` on the published base runs from 2026-07-22T08:09:12 to 2026-09-15T10:07:27: | month of `fetched_at` | rows | share of 127,602 | |---|---|---| | 2026-07 | 45,211 | 35.4% | | 2026-08 | 82,383 | 64.6% | | 2026-09 | 8 | 0.0% | No row in the base is missing a `fetched_at`. A count here is therefore a state of the vendor's records over that window, not a single instant, and a person hired or promoted inside it can appear under either title. ## The outlets `tools/report_editors_per_reporter.py` imports `tools/export.py` and calls its `load_outlet_records`, `outlet_name`, `outlet_country_code`, `fallback_country_code`, `outlet_type` and `slugify`, so an outlet's published name, country, type and slug are the same strings the site shows. An outlet is publishable when it is not one of the twelve parent-company domains, and carries a name, a country and a type. The 40 outlets here are the 40 publishable outlets with the most editorial rows, the same 40 that have `/outlets//` pages. 8 of the 12 parent-company domains carry rows in the base, 3,313 rows in all: foxcorporation.com, gettyimages.com, hearst.com, iheartmedia.com, paramount.com, pearson.com, skygroup.sky and voxmedia.com. The other four (condenast.com, nbcuniversal.com, netflix.com, newscorp.com) carry none. Those 3,313 rows are excluded from every outlet table here, because a row at a holding company's own domain is not a newsroom row. ## The ranking rule `outlets_top40_editors_per_reporter.csv` is ranked by the exact, unrounded ratio `editors` / `reporters`, descending, with ties broken by `editorial_rows`, descending. The published `editors_per_reporter` column is that ratio rounded to two decimals, so two neighbouring rows can print the same number: News Corp Australia (154 / 203 = 0.7586) ranks above RTE (115 / 152 = 0.7566) and both print 0.76. `rank_narrow` applies the same rule to the narrow ratio, and is in the file so ranks are read off it rather than recounted. Under the narrow definition the BBC is rank 34, one place below Deutsche Welle (0.4644 against the BBC's 0.4604), where under the wide definition the BBC is rank 33 and Deutsche Welle 34. ## The ratio `editors_per_reporter` = rows with role `editor` / rows with role `reporter`, at the outlet, in the country cut, or in the whole base. The other six roles (correspondent, columnist, anchor/presenter, producer, critic, other editorial) are not in either side of the ratio: 27,041 of the 127,602 rows, 21.2%. Both directions are published, because at some outlets the readable number is reporters per editor. The narrow definition drops the editors whose `editor_kind` is video, photo, copy or audio and keeps the rest, which the classifier records as text. It is published beside the wide one in every table rather than replacing it. ## The type cut The type rows in `editors_per_reporter_by_type_and_country.csv` cover publishable outlets only, the same definition the 40 are drawn from: 3,062 newspapers, 3,115 broadcasters, 2,216 online outlets, 183 magazines and 59 wire agencies, 123,453 rows. An earlier version of this table counted every typed non-parent outlet, which added 377 rows at 99 outlets that carry a type but no publishable name or country, and moved the broadcaster ratio from 0.89 to 0.90; no other type ratio changed at two decimals. The 123,453 type rows fall 4,149 short of the base: 3,313 rows at the parent-company domains, 459 at non-parent outlets whose own record carries no type, and the 377 typed rows without a publishable name or country. ## The two headline outlets, by domain An outlet here is a canonical key that can gather more than one `company_domain`. Both outlets in the title do. | outlet | company_domain | rows | reporters | editors | editors per reporter | |---|---|---|---|---|---| | The New York Times | nytco.com | 1,801 | 566 | 963 | 1.70 | | The New York Times | nytimes.com | 367 | 66 | 181 | 2.74 | | BBC | bbc.co.uk | 3,192 | 1,734 | 635 | 0.37 | | BBC | bbc.com | 463 | 161 | 248 | 1.54 | | BBC | bbcstudios.com | 280 | 62 | 155 | 2.50 | nytco.com is the website the vendor records on the New York Times Company page, so the two New York Times domains are two company records for one employer, not two populations. On nytimes.com alone the ratio is 2.74 editors per reporter over 367 rows, 66 reporters and 181 editors; the outlet total, 1.81, is the figure the report uses. bbcstudios.com is BBC Studios, the BBC's commercial production and distribution arm, a separate company from BBC News; what its 280 rows do is not established here. Without it the BBC has 3,655 rows, 1,895 reporters and 883 editors, 2.15 reporters per editor, against 1.89 for the outlet as a whole. bbc.co.uk on its own, 3,192 rows, runs 2.73 reporters per editor. ## Where the outlet came from Every outlet row carries the table-A source it was attributed to, and the ratio differs across the four: | source of the outlet | rows | reporters | editors | editors per reporter | |---|---|---|---|---| | `pdl-core` | 75,142 | 26,048 | 32,313 | 1.24 | | `wikidata-org` | 23,169 | 9,523 | 9,497 | 1.00 | | `known-list` | 17,753 | 7,263 | 6,639 | 0.91 | | `wikidata-employer` | 11,538 | 3,179 | 6,099 | 1.92 | The four sum to the base. `known-list` is the 125 hand-listed outlets, the cleanest segment, and it runs below one editor per reporter; `wikidata-employer`, the widest and noisiest, runs highest. The sources and the priority order between them are documented in `data/METHOD.md`. ## Correspondents Role is assigned from the job title, and the classifier tests `correspondent` before `reporter`, so a title reading "Business Correspondent" is a correspondent and not a reporter. Houses differ on which word they use for the same job, which moves the reporter side of the ratio. `wire_vs_newspaper_correspondents.csv` is there so the size of that effect is visible rather than argued: the five largest wire agencies on file (of 59) put 22.8% of their editorial rows under correspondent titles, against 8.5% at the eight largest newspapers among the 40 (of 10 typed newspapers in the 40; USA Today and New York Post omitted) and 8.0% across all ten. Both newspaper totals are rows in the CSV. ## Cells under 5 Counts of 1 to 4 print "<5". No row in this dataset identifies a person; the smallest unit is an outlet and a role. ## Sanity check against public figures - The New York Times: 1,700 journalists, as of March 2023, attributed to deputy managing editor Sam Dolnick, alongside 5,800 company employees. (fetched 2026-09-15 through the Webshare gateway). The New York Times Company's own pages that were fetched (, ) carry no headcount, and the 2024 annual report PDF returned HTTP 500 through the gateway, so no company-published figure is recorded here. On file: 2,168 New York Times editorial rows, 128% of 1,700. The on-file count is not a headcount; it includes people who list the Times on a profile without being current newsroom staff, and misses staff with no profile. - BBC: 7,000 employees including 5,500 journalists, the BBC News infobox figure cited to bbc.co.uk. (fetched 2026-09-15 through the Webshare gateway). That 5,500 is BBC News's figure. The 3,935 rows on file are the BBC outlet key, which spans bbc.co.uk, bbc.com and bbcstudios.com, so it is not the same population: 3,935 is 72% of 5,500, and bbc.co.uk on its own carries 3,192 rows, 58% of it. Neither figure is a check on the ratio, only on the order of magnitude of the counts. ## Limitations 1. Titles are self-reported on profiles, not taken from an HR file, and they lag promotions and layoffs in both directions. 2. "Editor" covers commissioning, copy, picture, video and audio work. Both definitions are published; the kind patterns are English, so a non-English video-editing title is counted as a text editor. 3. Coverage is profile findability, not employment. Every count is a floor, and the shortfall is not uniform across countries. 4. A person is counted at the outlet the vendor records, so a stringer, a contractor and a staffer are one row each. 5. Country in the country cut is the person's location, not the outlet's, so a Reuters correspondent in Nairobi counts under Kenya. 6. The base was collected between 2026-07-22 and 2026-09-15, not at one instant; the month distribution is above. ## Licence CC BY 4.0. Cite as JournalistLabs, "Editors per reporter at the 40 largest newsrooms on file", 2026-09-15, rows collected 2026-07-22 to 2026-09-15.