How many independent centers of control sit behind a country's visible news brands?
Different websites can look like different choices. To understand how much choice they represent, we need to look behind the names, follow the recorded exposure and account for the owners we cannot identify.
Eight countries. One published source table. A test of what the evidence can establish, not a verdict on editorial independence.
25,631 source records · Measurement periods undisclosed.Tom T / CC BY 2.0. Printing press, 2009; cropped and darkened. Contextual image, not the merger below.
Jump to a section
Why this matters
Why look beyond the front page?
News is one of the ways a society observes itself. Through reporting, people learn what institutions are doing, which claims deserve scrutiny and which events might affect their own decisions. It is therefore worth asking how many genuinely separate organizations contribute to that picture, rather than stopping at the number of websites available.
The visible layer is made of brands, domains and individual stories. The ownership layer sits behind it. A group may operate several titles, while a single website can distribute work from many sources. Opening another domain therefore does not necessarily mean moving to another owner, just as reading from the same owner does not prove that two newsrooms make the same editorial decisions.
Ownership is worth investigating because it concerns a layer of institutional control: the resources available to an organization, its investment priorities and the arrangements under which editorial work takes place. Those questions matter, but this dataset does not observe newsroom decisions or measure the content of articles. The analysis below describes recorded ownership structure; it does not establish that ownership causes similar coverage.
There is also an economic measurement problem. A small local site and a large national service each count as one domain, although their recorded activity can differ greatly. We therefore ask the question three ways: by counting websites equally, by weighting them with Similarweb visits and by weighting them with GDELT article counts. The last two are alternative exposure proxies, not direct measurements of what people notice, trust or remember.
Finally, an ownership map can look complete precisely because its blank entries have been removed. That is the error this investigation refuses to make. Unidentified domains keep their recorded weight. We then test how far the answer moves when those domains are treated as separate owners or assigned to the largest known group. The aim is to distinguish a result supported by the records from a result supplied by an assumption.
Case in point / United Kingdom / 2018
The mastheads stayed. The ownership boundary moved.
A documented newspaper merger makes the distinction concrete before we turn to the country comparisons.
On 28 February 2018, Trinity Mirror, subsequently renamed Reach, acquired publishing assets from Northern & Shell. The Competition and Markets Authority recorded five UK national newspapers in the buyer's existing portfolio and four in the acquired company. The Daily Mirror, Daily Express and Daily Star thereby came under the same corporate roof. Distinct brands remained; the ownership map had changed.
This is why counting mastheads can miss consolidation. The important event was not an additional website appearing or an old title disappearing. Existing publications moved inside a larger ownership group. A list of brands alone would not reveal that change in control, and a count of domains would not tell us how much recorded exposure the combined group received.
The case also shows why concentration is a question to investigate rather than an automatic verdict. On 20 June 2018, the CMA cleared the merger, finding no realistic prospect of a substantial lessening of competition. The Secretary of State also accepted Ofcom's conclusions that the merger did not raise concerns about plurality of views or free expression of opinion, and did not refer it for a Phase 2 investigation.
This historical example comes from regulatory records, not from the undated ownership and exposure snapshot analyzed below. It does not explain that snapshot's UK HHI or establish a causal effect on editorial content.
Sources: CMA clearance decision, 20 June 2018. Paragraph 1: completion date and company renaming. Paragraphs 6-7: five existing and four acquired UK national titles. Summary, paragraphs 4-5: competition assessment and clearance. · Secretary of State's statement, 20 June 2018. Acceptance of CMA and Ofcom conclusions and the decision not to refer the merger for Phase 2.
The central questions
Four questions guide the investigation.
- 01
How much apparent choice survives ownership grouping?
Start with domains, then connect the owner names recorded behind them. Test whether linking shared names changes the picture compared with keeping each name set separate.
- 02
Do the busiest domains belong to fewer groups?
A country can have many small sites and still have concentrated recorded exposure. Compare an equal domain count with visits, then inspect article volume as a separate proxy.
- 03
How much could the unidentified owners change the answer?
Keep the same domains and recorded weights. Change only the ownership assignment of the missing names, and observe the resulting concentration range.
- 04
What does this table mean by a news environment?
Check which platforms survive the filters, distinguish institutional classifications and examine newspaper, television and web records separately.
Before the first chart
What we are actually investigating.
The News Ownership Dataset supplies 25,631 records. This story examines the records assigned to Germany, the United States, the United Kingdom, France, Italy, India, Japan and Canada. A country label belongs to the domain record; it does not identify where its visitors live. The result is a comparison of country-assigned source samples, not a complete census of national audiences.
The source is useful for an exercise in ownership concentration, but it is not ready-made ground truth. Its owner fields contain names rather than stable owner IDs; some entries list several names, and the published-name field can disagree with the primary field. The deposit also leaves measurement periods and several classification rules undocumented. Those limitations remain visible throughout the investigation rather than appearing only after the charts.
Germany has 681 retained domains in the default source-table view. Under recorded visit weights, the conditional range is 7.0–21.3 equally sized owner groups. These are different units, not contradictory counts.
Linked source-name groups; portals excluded. Separate unidentified owners versus all joining the largest group. The following steps explain how to read this result and what it cannot establish.Before calculating concentration, establish the population and the meaning of an owner group. The controls below let you challenge both choices while keeping the underlying source fixed.
First / establish the unit
Begin with a country, not a concentration score.
Look at its high-visit domains, reveal the recorded names and distinguish what is observed from what the grouping rule adds. Then keep that same country in view as the analysis moves from ownership to exposure and missingness.
Look for: Different domains that lead to the same recorded group, and domains whose owner remains unidentified.
Country means the source's corrected domain assignment, not the visitor's location. The geographic scope of visits is undocumented. Exclusion-flagged records, invalid hosts and source-labeled government nonmedia stay excluded. Other unflagged non-news sites can remain. Deposit: 16 September 2026. Measurement periods undisclosed.
Before treating these groups as independent owners
The owners field contains names, not stable owner IDs. Default linked grouping connects names co-listed in a country's non-excluded, non-government records, including portals, with transitive closure. This can connect parents, subsidiaries or historical owners and is not verified beneficial ownership. Exact-name-set sensitivity groups the normalized semicolon sets instead. No fuzzy matching or cross-country linkage is used. owners_published is retained separately and is never substituted for owners because it can conflict.
Blank medium is not recoded as web. Multi-medium records appear in every declared medium, so the three medium views overlap. Neither the grouping nor the source classifications are independently verified.
Look behind the mastheads.
Keep one country in view. First see the domains, then reveal the recorded owners, then decide what to assume about the names the dataset cannot supply.
Germany / All media / Source table (flags applied) / highest recorded visits
- bild.de191,022,833 recorded visits · newspaper;television;websiteLook behind the name →
- n-tv.de117,141,458 recorded visits · televisionLook behind the name →
- web.de91,917,962 recorded visits · medium unclassifiedLook behind the name →
- welt.de74,782,647 recorded visits · newspaper;television;websiteLook behind the name →
- tagesschau.de69,726,188 recorded visits · medium unclassifiedLook behind the name →
- spiegel.de68,252,173 recorded visits · websiteLook behind the name →
- kicker.de66,941,239 recorded visits · websiteLook behind the name →
- twitter.com66,047,999 recorded visits · televisionLook behind the name →
The 8 highest-visit retained domains in this view. This is a labeled specimen, not the full sample or a representative selection. Every concentration result uses all 681 retained domains; changing the weighting does not change this traffic-ranked shortlist.
The leading owner groups receive 35.9% of the recorded weight.
- 14.1%ARD [25 linked names]Recorded owner group
- 11.3%Axel Springer SE [2 linked names]Recorded owner group
- 6.0%RTL Group [3 linked names]Recorded owner group
- 4.4%Holtzbrinck Publishing Group [3 linked names]Recorded owner group
Weights: Similarweb visits. Each unknown is independent. Link shared owner names. Ranks are scenario-dependent; an unidentified domain can outrank a known group. These are source-name groups, not verified beneficial owners. Inspect every linked name.
Next / bound the answer
Start with a range before reading a ranking.
Keep each country's retained sample and recorded weights fixed. Give the unidentified domains separate owners at one end, then assign them to the largest known group at the other. This reveals how much the missing names alone can change the result.
Look for: Wide ranges indicate sensitivity to unidentified ownership. They are not sampling confidence intervals.
One country, more than one defensible answer.
The line joins two deliberately extreme assignments of unidentified domains. At the low-concentration end, each has a different owner. At the high-concentration end, they all join the largest known owner. The country sample and the amount of recorded exposure stay unchanged.
The Herfindahl–Hirschman Index adds up squared group shares. It approaches zero when exposure is spread across many small groups and equals one when a single group receives everything. Its inverse gives the number of equally sized groups that would produce the same concentration. Here the groups are constructed from recorded names, not verified legal owner IDs; the grouping control tests that additional assumption.
Source table (flags applied) / All media / Similarweb visits / Each unknown is independent / Link shared owner names / Portals excluded
HHI, 0–1. These are conditional ownership assignments, not statistical confidence intervals. An open owner field remains in the exposure denominator.
Inspect a country by hover, keyboard focus or tap. Your persistent country is Germany; sources and precise values remain available below.
Read the exact values
| Country | Domains | Known owners | Owner coverage | Weight coverage | Top owner | Top four | HHI | Effective owners | Lower HHI | Upper HHI | Missing weights | Zero weights |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Germany | 681 | 118 | 31.7181% | 80.0389% | 14.1250% | 35.9177% | 0.047058 | 21.2503 | 0.047058 | 0.142331 | 0 | 59 |
| United States | 9,246 | 1,587 | 47.2421% | 90.9313% | 37.4634% | 72.9903% | 0.209796 | 4.7665 | 0.209796 | 0.283067 | 14 | 904 |
| United Kingdom | 1,465 | 262 | 34.8805% | 91.4436% | 54.6564% | 79.3480% | 0.331335 | 3.0181 | 0.331335 | 0.431752 | 2 | 233 |
| France | 485 | 102 | 38.5567% | 77.6348% | 15.9958% | 34.3434% | 0.050604 | 19.7613 | 0.050604 | 0.169148 | 0 | 36 |
| Italy | 989 | 87 | 20.5258% | 58.7829% | 8.8965% | 27.5574% | 0.030550 | 32.7337 | 0.030550 | 0.269973 | 1 | 44 |
| India | 800 | 166 | 39.3750% | 80.8299% | 11.1033% | 37.6888% | 0.050883 | 19.6531 | 0.050883 | 0.128914 | 3 | 159 |
| Japan | 568 | 175 | 52.6408% | 85.8329% | 18.7275% | 46.8583% | 0.075050 | 13.3245 | 0.075050 | 0.146690 | 0 | 67 |
| Canada | 959 | 178 | 47.0282% | 82.3523% | 44.0157% | 61.0692% | 0.209371 | 4.7762 | 0.209371 | 0.394827 | 1 | 110 |
Next / change the measure
Keep the ownership assumption. Change what counts.
Now compare equal domain weights with recorded visits under the same grouping and missing-owner scenario. This separates the variety of available addresses from the concentration of the activity recorded against them.
Look for: Above the diagonal, visits are more concentrated than the domain count suggests. Below it, the reverse is true.
A website count and a visit count tell different stories.
Equal weights treat a small local website like a large national portal. Traffic weights instead ask where the recorded visits go. A country above the diagonal has greater concentration under the traffic proxy than under an equal count of domains, given the selected ownership assumption.
Neither axis measures the diversity of articles a person actually reads. Similarweb visits are an exposure proxy. GDELT article counts offer a different proxy, affected by publication volume and the outlets GDELT observes. Switching the weights is a sensitivity check, not a search for the one true audience measure.
Source table (flags applied) / All media / Domains versus Similarweb visits / Each unknown is independent / Link shared owner names / Portals excluded
A common 0–0.4 HHI scale. ○ Domains · ◆ Visits
Both axes always compare domains with visits under the selected grouping and missing-owner scenario. GDELT remains available in the other exhibits. The diagonal is an identity line, not a fitted model. Missing paired denominators are not plotted and remain unavailable in the table.
Inspect a country by hover, keyboard focus or tap. Your persistent country is Germany; sources and precise values remain available below.
Read the exact values
| Country | Domains | Visits | Articles |
|---|---|---|---|
| Germany | 0.007640 | 0.047058 | 0.016835 |
| United States | 0.018581 | 0.209796 | 0.039956 |
| United Kingdom | 0.003304 | 0.331335 | 0.046206 |
| France | 0.004885 | 0.050604 | 0.040485 |
| Italy | 0.002710 | 0.030550 | 0.080059 |
| India | 0.004203 | 0.050883 | 0.018102 |
| Japan | 0.015808 | 0.075050 | 0.035535 |
| Canada | 0.017603 | 0.209371 | 0.029545 |
Next / explain the uncertainty
Find out where the ownership map is incomplete.
The previous ranges tell us how sensitive concentration can be. This view explains the underlying gap: compare the share of domains with a recorded owner with the share of measured exposure those domains carry.
Look for: A large difference between the two dots means domain coverage alone gives a misleading impression of exposure coverage.
Count the missing names. Then weigh them.
A high share of identified domains need not mean high coverage of visits. One unidentified high-traffic website may matter more for concentration than many small identified outlets. Conversely, poor domain coverage can coexist with good coverage of measured exposure.
The unfilled part of each row remains visible. Missing traffic is not converted into zero traffic, and zero traffic does not erase a domain from the equal-weight calculation. Ownership missingness and exposure missingness are separate problems; the bounds address the first, conditional on the weights we have.
Source table (flags applied) / All media / Similarweb visits / Source coverage and classification / Link shared owner names / Portals excluded
Coverage is a share of the retained sample, not a claim that the source captured every news outlet. Hover, focus or tap a country for unrounded counts and missing-weight details.
Inspect a country by hover, keyboard focus or tap. Your persistent country is Germany; sources and precise values remain available below.
Read the exact values
| Country | Domains | Known owners | Owner coverage | Weight coverage | Top owner | Top four | HHI | Effective owners | Lower HHI | Upper HHI | Missing weights | Zero weights |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Germany | 681 | 118 | 31.7181% | 80.0389% | 14.1250% | 35.9177% | 0.047058 | 21.2503 | 0.047058 | 0.142331 | 0 | 59 |
| United States | 9,246 | 1,587 | 47.2421% | 90.9313% | 37.4634% | 72.9903% | 0.209796 | 4.7665 | 0.209796 | 0.283067 | 14 | 904 |
| United Kingdom | 1,465 | 262 | 34.8805% | 91.4436% | 54.6564% | 79.3480% | 0.331335 | 3.0181 | 0.331335 | 0.431752 | 2 | 233 |
| France | 485 | 102 | 38.5567% | 77.6348% | 15.9958% | 34.3434% | 0.050604 | 19.7613 | 0.050604 | 0.169148 | 0 | 36 |
| Italy | 989 | 87 | 20.5258% | 58.7829% | 8.8965% | 27.5574% | 0.030550 | 32.7337 | 0.030550 | 0.269973 | 1 | 44 |
| India | 800 | 166 | 39.3750% | 80.8299% | 11.1033% | 37.6888% | 0.050883 | 19.6531 | 0.050883 | 0.128914 | 3 | 159 |
| Japan | 568 | 175 | 52.6408% | 85.8329% | 18.7275% | 46.8583% | 0.075050 | 13.3245 | 0.075050 | 0.146690 | 0 | 67 |
| Canada | 959 | 178 | 47.0282% | 82.3523% | 44.0157% | 61.0692% | 0.209371 | 4.7762 | 0.209371 | 0.394827 | 1 | 110 |
Next / distinguish institutions
Ask what kind of organization is represented.
A similar concentration score can sit beside a different institutional composition. Separate commercial, public and state-sponsored labels without treating those categories as an assessment of editorial behavior.
Look for: Which source categories carry the recorded weight, and which categories or coding decisions require closer scrutiny.
The kind of owner is a separate question.
Commercial, public and state-sponsored classifications describe institutional categories. They do not rank journalistic quality, editorial independence or political influence. We retain the source's classifications rather than recoding outlets from their reputation.
The missing-owner scenarios do not rewrite these labels. Classification and ownership identity are separate source fields, so an unidentified owner can still have a recorded classification. Party press remains Other; source-labeled government nonmedia is excluded by an explicit analysis rule. None is silently labeled commercial.
Source table (flags applied) / All media / Similarweb visits / Source coverage and classification / Link shared owner names / Portals excluded
Each bar sums to 100% of the selected recorded weight within that country and medium. Blank exposure is not a zero-share category. Institutional labels are not ratings of editorial quality or independence.
Inspect a country by hover, keyboard focus or tap. Your persistent country is Germany; sources and precise values remain available below.
Read the exact values
| Country | Commercial | Public | State-sponsored | Other / party press | Unclassified |
|---|---|---|---|---|---|
| Germany | 94.6129% | 5.3546% | 0.0249% | 0.0076% | 0.0000% |
| United States | 99.9662% | 0.0260% | 0.0078% | 0.0000% | 0.0000% |
| United Kingdom | 83.0993% | 16.9000% | 0.0000% | 0.0007% | 0.0000% |
| France | 93.7871% | 6.2129% | 0.0000% | 0.0000% | 0.0000% |
| Italy | 96.6902% | 3.3098% | 0.0000% | 0.0000% | 0.0000% |
| India | 99.8688% | 0.0000% | 0.1312% | 0.0000% | 0.0000% |
| Japan | 99.6917% | 0.3083% | 0.0000% | 0.0000% | 0.0000% |
| Canada | 85.9922% | 14.0078% | 0.0000% | 0.0000% | 0.0000% |
Next / look inside the national total
Recalculate the same question within each medium.
A national aggregate can conceal different structures among newspaper, television and web records. Rebuild the traffic denominator inside each country-medium cell instead of treating the national score as an answer for every medium.
Look for: Differences across the three columns, alongside missing weights and overlapping medium labels.
The medium changes the denominator.
Newspaper, television and web domains need not have the same ownership structure. Here the traffic shares are recalculated within each country–medium cell before they are squared. A television result concerns the websites of television outlets; it is not a measure of television viewing.
The national HHI is not an average of these three indices. A multi-medium domain enters each declared medium, while the all-media result counts it once and also retains the unclassified records. The columns overlap. Blank medium is not silently interpreted as web.
| Country | Newspaper | Television | Web |
|---|---|---|---|
| Germany | |||
| United States | |||
| United Kingdom | |||
| France | |||
| Italy | |||
| India | |||
| Japan | |||
| Canada |
Traffic weights only, irrespective of the weighting control above. Bold: selected grouping and ownership scenario. Below: lower–upper conditional HHI and domain counts. “Weighted” means positive recorded visits. Multi-medium records appear in more than one column. Select a cell to inspect that population.
Finally / return to the question
Separate the result from the claim it cannot support.
Bring together grouping, exposure, missingness, institutional labels and the medium-specific results. The conclusion should reflect what survived those checks, not just the most dramatic country ranking.
Look for: The next records or audience measures that would most reduce uncertainty.
Plurality needs an ownership account.
The number of news websites may create an illusion of choice. A more informative question is how many owners sit behind the domains that receive most of the measured exposure, and how much of that exposure still lacks an identified owner.
For a researcher or regulator, the useful next step is therefore specific: investigate the unidentified domains that carry the most weight, check the grouping of known owners, and obtain audience measures with the right national scope. Collecting another list of website names would not resolve those uncertainties.
This exercise does not show that common ownership causes similar reporting. Establishing that would require article-level content, evidence about editorial decisions and, for a causal claim, a defensible research design such as an ownership change with a credible comparison.
Inspect the source and the assumptions ↓The final step / make the result inspectable
Reconstruct the comparison from its source.
The remaining section documents the source vintage, retained fields, filtering rules and calculations. Use it to check whether another reasonable definition would change the question, and download the exact results rather than estimating values from a chart.
Look for: A clear route from the original record to the displayed group, denominator and concentration measure.
What went into the calculation.
The source and the clock
News Ownership Dataset. Version-specific Zenodo deposit 22782826; no semantic version supplied. Retrieved . Deposit: 16 September 2026. Measurement periods undisclosed.
Country means the source's corrected domain assignment, not the visitor's location. The geographic scope of visits is undocumented.
- Similarweb
- The numeric traffic_visits field is used as supplied. The deposit does not document the observation period, aggregation, geographic scope or desktop/mobile coverage. Explicit zeros are not certified zero audience; missing values are not replaced by zero.
- GDELT
- The gdelt_article_count field is an alternative exposure proxy. Collection dates, query, deduplication and missing-value meaning are not supplied. Blank cells remain missing, not zero articles.
- Classification
- Source-assigned labels: commercial, public_service, state_sponsored, government_nonmedia and party_press. Government nonmedia are excluded by our explicit analysis rule. Party press remains Other. Criteria are undocumented; for example, nhk.or.jp is supplied as commercial. These labels are not independently verified.
- License
- CC BY 4.0. Anonymous for peer review (2026), News Ownership Dataset, Zenodo. Analysis and transformations by Michael Schymura. The deposit does not separately document underlying third-party data rights.
Deposit description · Original source file · CC BY 4.0
The deposit supplies one CSV, but no codebook, measurement-period documentation or downloadable companion paper. Its publication date is not an observation date. For example, the supplied classification marks NHK as commercial; this story preserves, rather than endorses, that label.
Version and source fingerprint
340d76aae1ac8d2b4420ac35261b1480a4f6c4d6159b64fa34beda688da22991
SHA-256 of the unmodified downloaded input. This identifies the data vintage, not a live register of ownership.
Full-source validation
These counts describe the input across all countries, before the eight-country comparison. Exclusion and portal flags can overlap.
| Source rows | 25,631 |
|---|---|
| Unique domains | 25,630 |
| Duplicate domain keys | 0 |
| Exact duplicate rows removed | 0 |
| Invalid domain/host records | 5 |
| Missing owner identities | 15,870 |
| Zero recorded visits | 2,670 |
| Missing recorded visits | 33 |
| Zero GDELT article counts | 0 |
| Missing GDELT article counts | 8,447 |
| Exclusion flag set | 16 |
| Portal flag set | 76 |
| Both flags set | 0 |
| Retained nonportal outlets, all countries | 25,480 |
| Government nonmedia (also omitted) | 54 |
| Multiple primary owner names | 1,572 |
| Missing medium labels | 13,431 |
| Conflicting/different nonblank owner-name sets | 1,701 |
| Only a published owner name, primary blank | 32 |
| Country assignments corrected | 1,560 |
Population and missing weights
Use the corrected country assignment; blank corrected codes remain unassigned. Exclusion-flagged records, invalid hosts and government_nonmedia never enter the analysis. Portal domains are omitted by default and can be included as a sensitivity check. Unflagged non-news or platform domains can remain, so this is a source-table exercise, not a verified national news market. Duplicate domains must be reconciled before aggregation.
Every retained domain has equal weight in the domain calculation. A missing or zero proxy contributes no measured exposure to that proxy's total, but those are reported separately. We do not impute missing visits or articles, and the source does not establish that literal zero means observed zero audience. The ownership bounds are conditional on recorded weights, not bounds for all possible audience exposure.
Owner grouping and concentration
The owners field contains names, not stable owner IDs. Default linked grouping connects names co-listed in a country's non-excluded, non-government records, including portals, with transitive closure. This can connect parents, subsidiaries or historical owners and is not verified beneficial ownership. Exact-name-set sensitivity groups the normalized semicolon sets instead. No fuzzy matching or cross-country linkage is used. owners_published is retained separately and is never substituted for owners because it can conflict.
The display label of a linked group is its most frequently recorded member name, followed by its number of linked names. It is not a claim that this member ultimately controls the others. The original owner fields and complete group membership are downloadable.
HHI = sum of squared recorded-group shares
Effective groups = 1 / HHI
Top-group and top-four shares use the full retained proxy denominator, including unidentified exposure. The known-only calculation has a different denominator and is available in the complete results. HHI is on 0–1, not 0–10,000. An empty denominator produces “Not available,” not zero concentration.
Two bounds and one additional scenario
Separate owners: each unidentified domain is a distinct owner. Largest known owner: all unidentified exposure joins the largest known owner under the selected weighting. If none is identified, the upper case is one hypothetical owner of all domains.
Proportional: distribute unidentified weight fractionally across known owners in the same country and source medium, in proportion to their observed weight. A cell without positive known-owner weight stays unresolved. This is a modeled allocation, not evidence of who owns a domain. Because it splits domain weight, it can fall below the lower bound for indivisible domains; it is not promised to lie between the two endpoints.
Country–medium decomposition
For newspaper, television and web separately:
HHI(c,m) = sum_j [Traffic(j,c,m) / sum_k Traffic(k,c,m)]²
Each cell has its own denominator. A multi-medium domain enters each of its declared media, so these columns overlap. Blank medium remains unclassified, never automatically web. The all-media result includes it once. Proportional allocation uses exact recorded medium combinations, including a separate unclassified stratum. The country HHI is not an average of these cells; television websites are not broadcast viewing.
Country corrections, media, classifications and source flags
| Original | Corrected | Domains |
|---|---|---|
| (blank) | ALB | 2 |
| (blank) | AUS | 1 |
| (blank) | AUT | 14 |
| (blank) | AZE | 5 |
| (blank) | BDI | 1 |
| (blank) | BEL | 10 |
| (blank) | BGR | 1 |
| (blank) | BIH | 2 |
| (blank) | BLR | 1 |
| (blank) | BRA | 2 |
| (blank) | BRN | 1 |
| (blank) | CAN | 32 |
| (blank) | CHE | 1 |
| (blank) | CHL | 1 |
| (blank) | CHN | 59 |
| (blank) | CIV | 2 |
| (blank) | COD | 1 |
| (blank) | CPV | 3 |
| (blank) | CUB | 1 |
| (blank) | CYM | 1 |
| (blank) | CZE | 51 |
| (blank) | DEU | 3 |
| (blank) | DNK | 1 |
| (blank) | ESP | 9 |
| (blank) | EST | 23 |
| (blank) | FIN | 1 |
| (blank) | FRA | 15 |
| (blank) | GBR | 2 |
| (blank) | GEO | 2 |
| (blank) | GRC | 2 |
| (blank) | HKG | 1 |
| (blank) | HRV | 9 |
| (blank) | HUN | 3 |
| (blank) | IDN | 1 |
| (blank) | IND | 2 |
| (blank) | IRL | 10 |
| (blank) | ISL | 1 |
| (blank) | ITA | 54 |
| (blank) | JPN | 3 |
| (blank) | KAZ | 1 |
| (blank) | KGZ | 1 |
| (blank) | LKA | 1 |
| (blank) | LTU | 3 |
| (blank) | LVA | 1 |
| (blank) | MAR | 1 |
| (blank) | MDA | 1 |
| (blank) | MKD | 2 |
| (blank) | NAM | 3 |
| (blank) | NLD | 4 |
| (blank) | NOR | 1 |
| (blank) | PER | 1 |
| (blank) | POL | 49 |
| (blank) | PRT | 3 |
| (blank) | PSE | 7 |
| (blank) | ROU | 3 |
| (blank) | RUS | 58 |
| (blank) | SRB | 3 |
| (blank) | SVK | 16 |
| (blank) | SVN | 3 |
| (blank) | SWE | 4 |
| (blank) | TUR | 31 |
| (blank) | TWN | 1 |
| (blank) | UKR | 28 |
| (blank) | USA | 60 |
| (blank) | VNM | 17 |
| ALB | GRC | 1 |
| ALB | MKD | 2 |
| ALB | USA | 1 |
| ARG | BRA | 1 |
| ARG | ITA | 1 |
| ARM | ARG | 1 |
| ARM | AZE | 2 |
| ARM | CZE | 1 |
| ARM | TUR | 1 |
| ARM | USA | 2 |
| AUS | ARG | 1 |
| AUS | EST | 1 |
| AUS | FJI | 1 |
| AUS | GBR | 1 |
| AUS | HUN | 1 |
| AUS | NZL | 1 |
| AUS | QAT | 1 |
| AUS | USA | 1 |
| AUT | NLD | 1 |
| BEL | ITA | 1 |
| BIH | HRV | 1 |
| BLR | RUS | 1 |
| BRA | ARG | 1 |
| BRA | CHN | 1 |
| BRA | CPV | 1 |
| BRA | JPN | 1 |
| BRA | SLB | 1 |
| CAN | CHN | 1 |
| CAN | DEU | 1 |
| CAN | GBR | 1 |
| CAN | IND | 1 |
| CAN | JPN | 1 |
| CAN | PSE | 1 |
| CAN | USA | 7 |
| CHN | ARE | 1 |
| CHN | BEL | 1 |
| CHN | CAN | 1 |
| CHN | CHE | 1 |
| CHN | CHL | 1 |
| CHN | CZE | 4 |
| CHN | DNK | 1 |
| CHN | ESP | 1 |
| CHN | FRA | 3 |
| CHN | GBR | 7 |
| CHN | GRC | 2 |
| CHN | HKG | 11 |
| CHN | HRV | 1 |
| CHN | HUN | 2 |
| CHN | ISR | 1 |
| CHN | ITA | 4 |
| CHN | JPN | 4 |
| CHN | KOR | 1 |
| CHN | MAC | 4 |
| CHN | MKD | 5 |
| CHN | NLD | 1 |
| CHN | PER | 1 |
| CHN | PHL | 1 |
| CHN | POL | 2 |
| CHN | RUS | 3 |
| CHN | SGP | 1 |
| CHN | SRB | 1 |
| CHN | SVN | 2 |
| CHN | TWN | 1 |
| CHN | USA | 13 |
| CMR | BEL | 1 |
| COG | FRA | 1 |
| COL | BEL | 1 |
| CUB | USA | 1 |
| CZE | DEU | 1 |
| DEU | AUT | 9 |
| DEU | CHE | 7 |
| DEU | ESP | 1 |
| DEU | FRA | 1 |
| DEU | LIE | 1 |
| DEU | LUX | 1 |
| DEU | MKD | 1 |
| DEU | NAM | 1 |
| DEU | POL | 1 |
| DEU | USA | 1 |
| DMA | CZE | 1 |
| DNK | FRO | 4 |
| DNK | USA | 1 |
| EGY | DEU | 1 |
| EGY | LBN | 1 |
| EGY | MYS | 1 |
| EGY | PSE | 2 |
| ESP | ALB | 1 |
| ESP | AND | 1 |
| ESP | BRA | 1 |
| ESP | CHL | 1 |
| ESP | CZE | 1 |
| ESP | GBR | 2 |
| ESP | HUN | 2 |
| ESP | MKD | 5 |
| ESP | MYS | 1 |
| ESP | POL | 1 |
| ESP | RUS | 1 |
| ESP | USA | 1 |
| ETH | BEL | 1 |
| ETH | ZAF | 1 |
| FIN | ALA | 2 |
| FIN | USA | 1 |
| FRA | BEL | 3 |
| FRA | CHN | 1 |
| FRA | COD | 1 |
| FRA | CZE | 1 |
| FRA | DEU | 3 |
| FRA | DNK | 1 |
| FRA | GBR | 6 |
| FRA | ITA | 2 |
| FRA | JPN | 1 |
| FRA | LUX | 1 |
| FRA | MCO | 1 |
| FRA | MMR | 1 |
| FRA | NLD | 2 |
| FRA | PYF | 2 |
| FRA | ROU | 1 |
| FRA | SDN | 1 |
| FRA | TUR | 1 |
| GBR | AUS | 3 |
| GBR | BEL | 2 |
| GBR | BMU | 2 |
| GBR | CAN | 2 |
| GBR | CHE | 2 |
| GBR | DEU | 5 |
| GBR | ESP | 1 |
| GBR | FLK | 1 |
| GBR | FRA | 2 |
| GBR | GGY | 1 |
| GBR | GIB | 1 |
| GBR | IMN | 1 |
| GBR | IRL | 5 |
| GBR | ITA | 10 |
| GBR | MKD | 1 |
| GBR | NLD | 2 |
| GBR | NZL | 1 |
| GBR | USA | 4 |
| GBR | ZAF | 1 |
| GEO | RUS | 2 |
| GIN | GNQ | 1 |
| GRC | CAN | 1 |
| GRC | CYP | 1 |
| GRC | ROU | 2 |
| GRC | SRB | 2 |
| GRL | ATG | 1 |
| HKG | CHN | 5 |
| HKG | USA | 1 |
| HRV | BIH | 2 |
| HUN | MAR | 1 |
| IDN | CHN | 1 |
| IDN | GBR | 1 |
| IDN | MYS | 1 |
| IDN | PER | 1 |
| IDN | USA | 2 |
| IND | AUS | 2 |
| IND | BEL | 1 |
| IND | BGD | 1 |
| IND | CAN | 2 |
| IND | CHN | 1 |
| IND | DEU | 1 |
| IND | ESP | 1 |
| IND | GBR | 1 |
| IND | ITA | 1 |
| IND | MYS | 1 |
| IND | NPL | 4 |
| IND | RUS | 1 |
| IND | SGP | 1 |
| IND | USA | 3 |
| IRL | USA | 3 |
| IRN | CZE | 1 |
| IRN | USA | 2 |
| IRQ | USA | 1 |
| ISR | ARE | 2 |
| ISR | AUS | 4 |
| ISR | BEL | 4 |
| ISR | CAN | 3 |
| ISR | CHE | 11 |
| ISR | CZE | 1 |
| ISR | DEU | 2 |
| ISR | FIN | 1 |
| ISR | FRA | 8 |
| ISR | GBR | 7 |
| ISR | GRC | 2 |
| ISR | IRN | 3 |
| ISR | IRQ | 1 |
| ISR | ITA | 6 |
| ISR | LBN | 3 |
| ISR | MAR | 1 |
| ISR | MKD | 1 |
| ISR | NLD | 1 |
| ISR | NOR | 1 |
| ISR | NZL | 2 |
| ISR | PSE | 12 |
| ISR | SWE | 1 |
| ISR | TUR | 6 |
| ISR | USA | 13 |
| ISR | ZAF | 3 |
| ITA | AUS | 1 |
| ITA | BRA | 1 |
| ITA | CHE | 1 |
| ITA | DEU | 2 |
| ITA | FRA | 2 |
| ITA | GBR | 1 |
| ITA | IRL | 1 |
| ITA | MKD | 1 |
| ITA | SMR | 2 |
| ITA | UKR | 1 |
| ITA | VAT | 3 |
| JPN | AUS | 1 |
| JPN | BEL | 1 |
| JPN | CHN | 1 |
| JPN | DEU | 3 |
| JPN | FRA | 1 |
| JPN | GBR | 5 |
| JPN | HKG | 1 |
| JPN | ISR | 1 |
| JPN | ITA | 1 |
| JPN | NLD | 1 |
| JPN | SHN | 1 |
| JPN | USA | 6 |
| KAZ | RUS | 1 |
| KEN | CHE | 1 |
| KOR | JPN | 2 |
| KOR | USA | 2 |
| KWT | TTO | 1 |
| MAR | FRA | 2 |
| MAR | NLD | 1 |
| MDA | RUS | 2 |
| MEX | PRI | 1 |
| MEX | USA | 6 |
| MKD | CHE | 1 |
| MYS | BRN | 2 |
| MYS | SGP | 1 |
| NGA | CHE | 1 |
| NGA | USA | 2 |
| NLD | ABW | 1 |
| NLD | BEL | 1 |
| NLD | FRA | 1 |
| NOR | ASM | 1 |
| NZL | AUS | 1 |
| PAK | FRA | 1 |
| PAK | IND | 1 |
| PAN | CHL | 1 |
| PHL | AUT | 1 |
| PHL | CHN | 1 |
| PHL | USA | 2 |
| POL | CZE | 1 |
| POL | DEU | 2 |
| POL | UKR | 2 |
| POL | USA | 1 |
| PRT | BRA | 1 |
| PRT | CPV | 2 |
| PRT | ESP | 1 |
| PRT | LUX | 1 |
| PRY | FRA | 1 |
| QAT | CAN | 1 |
| QAT | GBR | 1 |
| QAT | TUR | 2 |
| ROU | HUN | 1 |
| RUS | AFG | 1 |
| RUS | ARG | 1 |
| RUS | AUS | 2 |
| RUS | AZE | 3 |
| RUS | BEL | 1 |
| RUS | BGR | 2 |
| RUS | BLR | 2 |
| RUS | BRA | 1 |
| RUS | CAN | 1 |
| RUS | CHN | 2 |
| RUS | CRI | 2 |
| RUS | CZE | 7 |
| RUS | DEU | 3 |
| RUS | ESP | 2 |
| RUS | EST | 1 |
| RUS | GBR | 1 |
| RUS | GRC | 2 |
| RUS | HRV | 1 |
| RUS | IRN | 1 |
| RUS | KAZ | 1 |
| RUS | KGZ | 1 |
| RUS | MDA | 2 |
| RUS | MKD | 2 |
| RUS | POL | 2 |
| RUS | SRB | 1 |
| RUS | SVK | 1 |
| RUS | UKR | 2 |
| RUS | USA | 1 |
| RUS | UZB | 1 |
| SAU | ARE | 1 |
| SDN | ITA | 1 |
| SDN | USA | 2 |
| SOM | USA | 1 |
| SRB | BIH | 3 |
| SRB | CZE | 1 |
| SRB | FRA | 1 |
| SRB | MNE | 1 |
| SSD | SLE | 1 |
| SWE | BEL | 2 |
| SWE | DNK | 2 |
| SWE | MEX | 1 |
| SWE | NOR | 3 |
| SWE | URY | 1 |
| SWE | USA | 1 |
| SYR | CHE | 1 |
| SYR | ITA | 1 |
| THA | NLD | 1 |
| TKM | RUS | 1 |
| TUR | DNK | 1 |
| TUR | HRV | 1 |
| TUR | ROU | 1 |
| TUR | USA | 1 |
| TWN | CHN | 2 |
| TWN | MDG | 2 |
| UGA | GBR | 1 |
| UGA | USA | 1 |
| UKR | AUT | 1 |
| UKR | CZE | 1 |
| UKR | GRC | 1 |
| UKR | HUN | 1 |
| UKR | ITA | 3 |
| UKR | LUX | 1 |
| UKR | MAR | 1 |
| UKR | POL | 2 |
| UKR | RUS | 1 |
| UKR | SVK | 2 |
| UKR | USA | 1 |
| USA | (blank) | 1 |
| USA | ARE | 1 |
| USA | ARG | 1 |
| USA | AUS | 3 |
| USA | AUT | 3 |
| USA | BEL | 4 |
| USA | BGR | 3 |
| USA | BMU | 1 |
| USA | BRA | 2 |
| USA | BRB | 2 |
| USA | CAN | 10 |
| USA | CHE | 2 |
| USA | CHL | 2 |
| USA | CHN | 8 |
| USA | CRI | 2 |
| USA | CZE | 2 |
| USA | DEU | 13 |
| USA | DNK | 4 |
| USA | DOM | 8 |
| USA | ESP | 8 |
| USA | FIN | 2 |
| USA | FRA | 12 |
| USA | GBR | 47 |
| USA | GEO | 6 |
| USA | GRC | 4 |
| USA | HKG | 2 |
| USA | HTI | 1 |
| USA | HUN | 23 |
| USA | IDN | 1 |
| USA | IMN | 1 |
| USA | IND | 4 |
| USA | IRL | 3 |
| USA | IRN | 1 |
| USA | ISL | 1 |
| USA | ISR | 4 |
| USA | ITA | 9 |
| USA | JPN | 3 |
| USA | KOR | 1 |
| USA | LBN | 1 |
| USA | MAR | 1 |
| USA | MEX | 4 |
| USA | MKD | 2 |
| USA | MYS | 1 |
| USA | NIC | 1 |
| USA | NLD | 18 |
| USA | NOR | 2 |
| USA | NZL | 1 |
| USA | PER | 1 |
| USA | PHL | 1 |
| USA | POL | 11 |
| USA | PRI | 8 |
| USA | PRT | 6 |
| USA | ROU | 3 |
| USA | RUS | 9 |
| USA | SGP | 1 |
| USA | SLV | 1 |
| USA | SRB | 1 |
| USA | SWE | 4 |
| USA | TCA | 1 |
| USA | THA | 1 |
| USA | TUR | 5 |
| USA | TWN | 2 |
| USA | TZA | 1 |
| USA | UKR | 2 |
| USA | URY | 1 |
| USA | VCT | 1 |
| USA | ZAF | 1 |
| UZB | RUS | 1 |
| VEN | ESP | 2 |
| VEN | USA | 2 |
| VNM | CHN | 1 |
| VNM | USA | 3 |
| XKX | MKD | 1 |
| YEM | KAZ | 1 |
| ZAF | CHE | 1 |
| ZAF | PSE | 1 |
| ZMB | MEX | 1 |
| ZWE | GBR | 1 |
| (blank) | 13,431 |
|---|---|
| newspaper | 5,843 |
| television | 4,720 |
| website | 1,026 |
| newspaper;website | 396 |
| television;website | 117 |
| newspaper;television | 86 |
| newspaper;television;website | 12 |
| commercial | 24,972 |
|---|---|
| public_service | 304 |
| state_sponsored | 285 |
| government_nonmedia | 54 |
| party_press | 16 |
| False | 25,555 |
|---|---|
| True | 76 |
| False | 25,615 |
|---|---|
| True | 16 |
| (blank) | 15,870 |
|---|---|
| wikidata_p127_infobox | 7,424 |
| wikidata_p123_p749_p137 | 1,059 |
| wikipedia_category_tree | 873 |
| subdomain_rollup | 377 |
| redirect | 14 |
| wikidata_p127_current | 13 |
| domain_normalisation | 1 |
| cctld | 9,579 |
|---|---|
| shallow | 7,596 |
| gdelt | 4,446 |
| infobox | 2,538 |
| published | 1,443 |
| (blank) | 21 |
| deep | 8 |
Take the evidence with you.
The complete results include all eight countries, four medium views, three weighting systems, all three assignments and portal sensitivity.
Complete concentration table (CSV)Original unfiltered source (25,631 rows; CSV)Linked owner-name groups (CSV)Validation report (JSON)Source and output manifestRebuild from the frozen input with npm run news-ownership:data. The repository retains the input, normalization rules and arithmetic tests separately from generated public assets.