The largest published measurement puts the gap at 33 to 1: the Centre for Media Monitoring reported in 2025, from more than 35,000 BBC articles and broadcast segments over twelve months, that Israeli deaths received 33 times more coverage per fatality than Palestinian deaths in BBC articles. A second organisation counting the same broadcaster over an overlapping window reports the imbalance running the other way: CAMERA found 35% of more than 2,500 BBC headlines critical of Israel against 11% critical of Hamas. Both counts stand, because neither measures what the other measures.
Key Findings
- Israeli deaths received 33 times more coverage per fatality than Palestinian deaths in BBC articles, and 19 times more in BBC television and radio, in a study of more than 35,000 BBC articles and broadcast segments over twelve months, on the Centre for Media Monitoring’s 2025 count, against tolls of 42,010 Palestinians and 1,246 Israelis.
- CAMERA coded the headlines of more than 2,500 BBC News articles published between 7 October 2023 and 7 October 2025 and reported 35% critical of Israel against 11% critical of Hamas, on CAMERA’s own definition of critical.
- The Gaza Ministry of Health’s list to 30 June 2024 carried 28,185 named records with complete data, 22,368 from its own hospital records and 5,817 from notification by relatives, with a further 9,715 deaths the ministry stated were not yet complete.
- The figure most often quoted for the accuracy of Gaza casualty reporting, a discrepancy of 1.5% to 3.8% in previous conflicts, appears in a 2024 Lancet correspondence in a sentence that carries no citation, and reproduces exactly from three dated number pairs printed in an Associated Press explainer of 26 October 2023 that prints no percentage at all.
- Total media exposure to the Israeli-Palestinian conflict did not directly predict ethnic stereotypes among 89 Jewish American and 180 Arab American high-school students surveyed by Huesmann and colleagues in 2012; the association ran through identification with the groups as portrayed.
How much coverage did each side’s deaths receive per fatality?
Israeli deaths received 33 times more coverage per fatality than Palestinian deaths in BBC articles, on the Centre for Media Monitoring’s 2025 study of more than 35,000 BBC articles and broadcast segments across twelve months of the war. The Centre is a British media-monitoring organisation applying its own coding frame, and the figure is its own calculation rather than a peer-reviewed measurement.
The same study puts the figure for BBC television and radio at 19 times. The denominators matter as much as the numerator. The Centre’s ratio divides coverage by deaths, and the death figures it used were 42,010 Palestinians and 1,246 Israelis for 7 October 2023 to 6 October 2024, counts it did not measure itself. That is a fatality ratio of about 34 to 1 for the same period.
A per-fatality ratio behaves differently from the volume figures most readers encounter. It rises when coverage of one side rises and it rises when deaths on the other side rise, so the same output can produce a very different ratio if the casualty figures are revised. The Centre’s 33 is therefore a statement about the relationship between two series, not about the quantity of BBC coverage of Gaza, which was large in absolute terms across the period.
Three further limits sit on the figure. It covers one broadcaster, so it establishes nothing about the New York Times, CNN, the Associated Press or Reuters. It covers twelve months, and the Centre published no continuation, so it speaks to no later period. And it counts items, which is a unit that treats a two-minute bulletin and a long feature alike, and which says nothing about placement, length, headline wording or the agent named in a headline.
Is it often said Western outlets humanise Israeli victims more?
It is often said that Israeli victims are named while Palestinian deaths are counted. A 2026 study in TRT Akademi, the peer-reviewed academic journal of Turkey’s public broadcaster, coded 160 items from the New York Times, BBC, CNN and the Guardian published between October 2024 and October 2025 and reported depersonalised language, passive construction and selective sourcing around Palestinian casualties. It publishes no numeric ratio.
Two further named studies run in the same direction. AlShebli, Casara and Maass (2025), in a preprint posted as arXiv:2510.06453 that has not been peer-reviewed, analysed 14,280 articles from the New York Times, BBC, CNN and Al Jazeera English across the war’s first year and found the odds of a story individualising a victim 14.7% lower for Palestinians than Israelis at the BBC, 28.7% lower at CNN and 41.3% lower at the New York Times. Thussu and Chatterjee (2026), in Third World Quarterly, is peer-reviewed, and its argument is about framing and context rather than a count of items; the casualty figures it works with are figures it relays from UN agencies and press reporting rather than figures it measured.
So the direction the named studies run in is consistent, and stating that is not the same as stating a magnitude. The TRT Akademi study’s 160 items across four outlets over thirteen months is a sample about one two-hundredth the size of the Centre for Media Monitoring’s, and its method is qualitative coding of language features rather than enumeration, which is why it reports patterns and not a rate. A finding that passive construction recurs is a finding about grammar in a sample. It cannot be converted into a claim that Israeli victims were named some number of times more often per death: the preprint counts individualised mentions, but no study read here divides namings by deaths.
The Centre for Media Monitoring’s version of the humanisation claim is quantified too, and it concerns victim profiles rather than language.
Do story counts track the death counts?
No, on the two pairings published with both halves in view. The Centre for Media Monitoring reported in 2025 that the BBC ran almost equal numbers of humanising victim profiles for Israelis and Palestinians, 201 and 279, across a twelve-month period in which the casualty figures it worked from stood at about 34 Palestinian deaths for every Israeli death.
The widely circulating version of this claim is the second pairing, and it comes from the 2025 preprint named above. AlShebli, Casara and Maass report that the BBC, CNN and the New York Times ran a close to one-to-one ratio of Palestinian to Israeli casualty stories, against roughly 20 Palestinian casualties for every Israeli casualty by the end of the first year. Two details travel with it: “casualties” there means deaths and injuries, not deaths alone, and the stories counted are victim narratives within articles, not whole articles.
The Centre’s two findings are arithmetically the same observation seen from two sides. If deaths differ by a factor of about 34 and item counts or profile counts do not differ by anything like that factor, then coverage per fatality differs by approximately the same factor. The Centre’s near-parity in profiles and its 33 to 1 coverage ratio are one measurement, not two independent confirmations of a pattern.
What a story count cannot register is also worth stating plainly. It treats every item as one unit, so it is blind to length, placement, repetition within a bulletin and the number of readers or viewers reached. Two outlets producing identical item counts can produce entirely different volumes of attention, and none of the studies read here measures audience reach.
Which monitoring groups find the bias runs the other way?
CAMERA, the Committee for Accuracy in Middle East Reporting in America, coded the headlines of more than 2,500 BBC News articles published between 7 October 2023 and 7 October 2025 and reported 35% critical of Israel against 11% critical of Hamas. CAMERA performed the coding itself and states an advocacy mission, so the share rests on its own judgment of what counts as critical.
The same 2025 study counts 45 headlines over those 24 months featuring claims of genocide, famine or starvation in Gaza, and 21 headline uses of the words starving or starvation. Those are raw counts within the corpus, which CAMERA puts at 2,533 stories, and CAMERA reports them as counts rather than shares.
CAMERA’s 2023 study of CNN takes a different approach again. It counted the 99 new articles in CNN’s Middle East section between July and September 2023 and divided regional population figures from the World Bank by that count, reporting 324,316 Israelis and Palestinians per CNN Middle East article. That is CAMERA’s own calculation from its own article count set against a third party’s population data, and it measures coverage volume per head of population; the article also sets article counts against ACLED conflict fatalities without printing a per-fatality figure for Israel.
NGO Monitor’s Annual Report 2024, published in 2025, works on a different object entirely. It reports 157 groups in the United States linked to what it identifies as an anti-Israel network, with over 3,000 connections, and 111 groups in Canada with over 1,800 connections, alongside $28 million in government funding to NGOs that it says was cut as a direct result of its own work. Those are self-reported counts of organisations, with inclusion criteria the report does not state, and they are not a measurement of any outlet’s coverage.
Are the two sets of studies counting the same thing?
No. The measurements named in this report use six different units: coverage items per fatality, share of headlines, population per article, coded passages in a small sample, individualised victim mentions, and personal details per post. None shares a denominator with another, so none of their figures can be added, divided into one another, or ranked.
| Body and year | What it counted | The denominator | Source type |
|---|---|---|---|
| Centre for Media Monitoring, 2025 | More than 35,000 BBC items over 12 months | Deaths, 42,010 and 1,246 | Monitoring organisation |
| CAMERA, 2025 | Headlines of more than 2,500 BBC articles | Headlines in the corpus | Monitoring organisation, advocacy mission |
| CAMERA, 2023 | 99 CNN Middle East articles, July to September 2023 | Population, from World Bank figures | Monitoring organisation, advocacy mission |
| TRT Akademi, 2026 | 160 items from four outlets, October 2024 to October 2025 | None published | Peer-reviewed journal |
| AlShebli and colleagues, 2025 | Victim mentions in 14,280 articles, four outlets | Casualties, deaths and injuries | Preprint |
| Elmasry, 2024 | 436 Instagram posts by five broadcasters, 7 to 16 October 2023 | Posts, not deaths | Peer-reviewed journal |
That table explains how two organisations can count the same broadcaster and publish opposite findings without either arithmetic being wrong. A per-fatality ratio is driven hard by the fatality denominator, which differed by a factor of about 34. A share-of-headlines figure is driven by the headline denominator and by whoever defines the coding category. Coverage can be weighted toward Israeli deaths on a per-death basis while a larger share of the headlines that assign criticism at all assign it to Israel. Neither result refutes the other, because they answer separate questions.
Nothing in the record establishes a shared codebook across these bodies. Each built its own categories, and none publishes a mapping onto another’s, which means the figures cannot be pooled into a single estimate of coverage asymmetry. This is the recurring problem with what published counts actually measure in this subject: the disagreement between two totals is frequently a disagreement about the unit, visible only when both denominators are stated.
How do outlets attribute the Gaza casualty figure?
To the Gaza Ministry of Health, which publishes the underlying records. Its list to 30 June 2024, the ministry stated on 24 July 2024, carried 28,185 named records with complete data.
Of those, 22,368 came from its own hospital records and 5,817 from notification by relatives, and in the same statement the ministry put a further 9,715 deaths as not yet complete.
The ministry is a party to the war reporting its own count, and it publishes its own exclusions. In a statement of 7 January 2024, never withdrawn, it recorded five: the missing and bodies trapped under rubble are not counted; the unidentified who reached hospitals are not counted; the unidentified whose bodies Israel handed over are not counted; those buried by their families without passing through a hospital are not counted; and victims in Gaza and North Gaza after the information system stopped in November 2023 are not counted. What the ministry publishes about its own count is therefore narrower than a census of the dead, by its own account.
One published assessment exists. Huynh, Chin and Spiegel published a Lancet correspondence in 2024, editorially reviewed rather than peer-reviewed, titled No evidence of inflated mortality reporting from the Gaza Ministry of Health. Its title is an absence claim and not a verification, and its own text states that mortality “might have been under-reported by the Gaza MoH due to decreased capacity.”
The sentence most often lifted from that letter is the one with the weakest support in it. It reports discrepancies between ministry reporting and independent United Nations analyses of 1.5% to 3.8% in previous conflicts, and that sentence carries no citation, though the letter cites the Associated Press explainer below for the sentence before it. Those percentages reproduce exactly from three dated pairs printed in an Associated Press explainer of 26 October 2023, which prints no percentage of its own: 1,440 against 1,385 in 2008, 2,310 against 2,251 in 2014, and 260 against 256 in 2021. Each divides on the ministry’s own count as denominator, and the letter’s 8.0% figure for 2014 comes from the same page’s Israeli Foreign Ministry comparison.
No count exists, in any study read here, of how often a given outlet attaches a caveat to the figure.
Does the asymmetry appear in images as well as words?
One peer-reviewed content analysis of the war’s imagery appears among the sources carried here: Elmasry (2024), in Journalism & Mass Communication Quarterly, coded 436 Instagram posts by five Western broadcasters in the war’s first ten days. It reports 0.47 personal details per post about Israeli victims against 0.14 about Palestinian victims, and no coverage-per-fatality comparison.
That is a different object of measurement in three respects at once. The denominator is posts rather than deaths, so a result cannot be expressed per fatality. The unit is a coded post, with its caption and video, rather than an article or broadcast segment, so it cannot be set beside the Centre for Media Monitoring’s item count. And the platform is a social feed with its own distribution and its own moderation, so an outlet’s Instagram output is not its editorial output, and a finding about one does not transfer to the other.
The two text-side studies measure different things again. The TRT Akademi study codes language features, passive construction and sourcing, in its 160-item sample. The Centre for Media Monitoring counts humanising victim profiles, which are items rather than photographs. Neither counts images.
So the humanisation argument, as it is usually made, spans three separate measurements that have never been run against a common denominator: whether a victim is named, whether a victim is pictured, and how many items each death produces. Naming has been counted, by the 2025 preprint and the Instagram study, but only the third has been quantified per fatality, and only for one broadcaster over one year. The photographic version of the claim is not measured at the per-death level anywhere in this record, and recording that is the accurate answer to the question rather than an extrapolation from the verbal findings to the visual ones.
Where do the content analyses disagree about method?
Sample sizes across these studies differ by a factor of more than two hundred: the Centre for Media Monitoring’s 2025 study codes more than 35,000 BBC items, against the 160 items across four outlets in the 2026 TRT Akademi study. Their designs differ as sharply as their scale.
Taken in turn, and by name. The Centre for Media Monitoring is a monitoring organisation, not peer-reviewed, coding to a frame of its own, with casualty denominators taken from a body that measured them separately. CAMERA’s 2025 and 2023 studies are advocacy-mission monitoring, self-coded, with “critical” defined by the coder in one case and a population denominator imported from the World Bank in the other. The TRT Akademi study of 2026 is peer-reviewed, is qualitative, publishes no ratio, and rests on a sample small enough that a handful of items shifts any pattern it reports. The AlShebli preprint of 2025 has not been through peer review and counts victim mentions against a casualty baseline that includes injuries. The Instagram content analysis of 2024 is peer-reviewed and measures a different medium. The Third World Quarterly article of 2026 is peer-reviewed and is an argument about framing that relays casualty figures rather than counting coverage.
The claims that survive each design are narrower than the claims the studies are usually cited for. A per-fatality volume asymmetry survives for one broadcaster across one twelve-month window. A headline-criticism asymmetry survives for the same broadcaster across a two-year window on one organisation’s coding. A qualitative pattern of depersonalisation survives for a 160-item sample from four outlets over thirteen months.
None of the three supports a statement about Western media as a whole, and the two BBC studies do not converge on a single number for the BBC, because they never counted the same thing. Where the literature is thin is also visible: no study read here covers the wire services, Reuters and the Associated Press, whose copy supplies much of what smaller outlets publish.
What has been measured about exposure to this coverage?
One measurement of exposure and attitudes appears in the peer-reviewed literature carried here, and it found no direct effect. Huesmann, Dubow, Boxer, Souweidane and Ginges surveyed 89 Jewish American and 180 Arab American high-school students in a 2012 study and reported that total media exposure to the Israeli-Palestinian conflict did not directly predict ethnic stereotypes.
What the study did find was an indirect association. The relationship ran through identification with the groups as portrayed in the media rather than through the volume of exposure itself. Both groups held more negative stereotypes about the out-group than the in-group, with ethnic identity correlating at r = -0.70 with an implicit measure of negative stereotypes about Arab versus Jewish youth. The authors report associations, not causes, and the design is cross-sectional.
Two limits bound how far that finding reaches. It measured adolescents in two specific communities in the United States, not a general audience. And it was published in 2012, more than a decade before October 2023, in a media environment without the platforms and volumes that the Instagram and wire-era studies above describe, so it cannot speak to exposure to coverage of the current war.
That leaves a gap worth stating directly. The coverage asymmetries measured by the Centre for Media Monitoring in 2025 and by CAMERA in 2025 have not been paired, in any study read here, with a measurement of what audiences exposed to that coverage came to believe about Jews, Israelis or Palestinians. The bias literature measures output. The attitudes literature measures people. Nothing in this record joins them for this war, and the one study that attempted the join for an earlier period found the path ran through who the viewer already identified with.
Methodology and limitations
This report draws on the source types below, each labelled where it appears. Monitoring organisations supply the two largest coverage counts: the Centre for Media Monitoring’s 2025 study of more than 35,000 BBC items, and CAMERA’s 2025 and 2023 studies of BBC headlines and CNN articles, with CAMERA and NGO Monitor stating advocacy missions and coding their own material. Peer-reviewed journals supply the TRT Akademi study of 2026, the Instagram content analysis of 2024, the Third World Quarterly article of 2026 and the Huesmann study of 2012. One preprint, AlShebli, Casara and Maass (2025), arXiv:2510.06453, is labelled as such at every use. One item, the Huynh, Chin and Spiegel piece of 2024, is Lancet correspondence, editorially reviewed rather than peer-reviewed, and PubMed carries no abstract for it.
Casualty figures throughout are attributed to the bodies that published them. The 42,010 and 1,246 used as denominators in the coverage ratio are figures the Centre for Media Monitoring used without measuring them, and its published summary does not name their source. The record counts of 28,185, 22,368, 5,817 and 9,715 are the Gaza Ministry of Health’s own published counts, from a party to the war, with the ministry’s own five stated exclusions attached.
The scope is the coverage and the counting. What is measured here is published output, the unit each counting body used, and the denominator each ratio rests on. Whether any particular editorial judgment was correct is a different question from how many items were produced, and no figure in this report bears on it.
Three limits are visible in the evidence itself. The per-fatality and headline-share studies both examine the BBC; the 2025 preprint adds the New York Times and CNN against a casualty baseline but publishes no per-fatality ratio, and no figure exists for Reuters or the Associated Press. The windows end by 2025 in every case. And the widely repeated pairing of a roughly one-to-one story ratio against a roughly 20 to 1 ratio comes from a preprint, compares casualty-related victim narratives rather than whole articles, and uses casualties, deaths and injuries together, rather than deaths.
Conclusion
Two organisations counted the same broadcaster’s Gaza coverage and published findings that point in opposite directions. Both are right about what they counted. The Centre for Media Monitoring divided coverage by deaths across more than 35,000 BBC pieces and got 33 to 1 for articles against a fatality ratio of about 34 to 1. CAMERA divided coded headlines by headlines across more than 2,500 BBC articles and got 35% against 11%. Change the denominator and the direction of the answer changes with it.
What the measurements establish, taken together, is narrower and firmer than the argument that surrounds them. Coverage volume on one broadcaster did not scale with the death counts over one twelve-month window, and the same study found near-parity in humanising victim profiles across a period whose casualty figures differed by a factor of about 34. A smaller qualitative sample of 160 items from four outlets found depersonalised language around Palestinian casualties and published no rate. A separate body of counting finds criticism concentrated in the other direction at the headline level. The casualty figure every outlet attributes is the Gaza Ministry of Health’s own, published in name-by-name lists whose completed portion split 22,368 to 5,817 by channel at the end of June 2024, alongside the ministry’s own statement of what it does not count. And the accuracy claim most often attached to that figure, 1.5% to 3.8%, traces to an uncited sentence that reproduces three number pairs from a news explainer.
The Institute’s subject is what feeds antisemitism, and this is where the measurement stops short of the question people are actually arguing about. Huesmann and colleagues found in 2012 that exposure did not predict stereotyping directly, and that the association ran through who the viewer already identified with. If that holds for coverage of this war, then a per-fatality ratio describes what was broadcast and not what was absorbed, and the argument over media bias is partly an argument about arithmetic standing in for a harder one. No study in this record has yet put the next question to a sample: which of these counts, if any, changes what a reader ends up believing about Jews, Israelis or Palestinians?
Frequently Asked Questions
How is coverage per fatality calculated?
It divides a count of coverage by a count of deaths. The Centre for Media Monitoring’s 2025 figure of 33 divides BBC article coverage by the tolls of 42,010 Palestinians and 1,246 Israelis, so it moves when either series moves, and a revision to the casualty figures changes the ratio without any change in what was broadcast.
Which body’s casualty figures do the media-monitoring studies use?
The Centre for Media Monitoring used 42,010 Palestinian and 1,246 Israeli deaths without measuring them, and its published summary does not name their source. The Palestinian toll in circulation rests on the Gaza Ministry of Health’s published records, which the ministry compiles as a party to the war and publishes with its own stated exclusions.
Is the Gaza Ministry of Health’s count a complete record of the dead?
No, on the ministry’s own account. It stated in January 2024 that it does not count the missing, bodies under rubble, the unidentified who reached hospitals, those buried by families outside hospitals, or deaths in Gaza and North Gaza after its information system stopped in November 2023, and at 30 June 2024 it reported 9,715 further deaths as not yet complete.
What is the difference between the source types cited here?
The 2024 Instagram content analysis in Journalism & Mass Communication Quarterly and the 2012 Huesmann study are peer-reviewed. The 2025 AlShebli, Casara and Maass paper, arXiv:2510.06453, is a preprint that has not been peer-reviewed. The 2024 Huynh, Chin and Spiegel item is Lancet correspondence, which is editorially reviewed rather than peer-reviewed.
Do any of these studies cover outlets other than the BBC?
Five do. The 2026 TRT Akademi study covers the New York Times, BBC, CNN and the Guardian across 160 items, CAMERA’s 2023 study covers 99 CNN articles, the 2025 preprint covers the New York Times, CNN and Al Jazeera English alongside the BBC, the Instagram study covers five broadcasters’ accounts, and the Third World Quarterly article reads the New York Times. No study read here measures Reuters or the Associated Press.
Does any of this evidence cover the period after 2025?
No. The Centre for Media Monitoring’s window is twelve months and CAMERA’s runs to 7 October 2025, and neither organisation published a continuation in the record carried here, so no per-fatality or headline-share figure exists for coverage after those dates.
Which of the counting bodies state an advocacy mission?
CAMERA and NGO Monitor both state advocacy missions and code their own material, which is why their figures are labelled as monitoring counts rather than measured shares. The Centre for Media Monitoring is a media-monitoring organisation applying its own frame, and TRT Akademi is the peer-reviewed academic journal of Turkey’s public broadcaster.
Sources
- AlShebli, B., Casara, B. G. S., and Maass, A., 2025. Media Coverage of War Victims: Journalistic Biases in Reporting on Israel and Gaza. arXiv:2510.06453 (v3, 25 February 2026). Preprint.
- Associated Press, 2023. DeBre, I. What is Gaza’s Ministry of Health and how does it calculate the war’s death toll? 26 October 2023. International press.
- CAMERA, 2023. Tracking CNN’s Bias: CNN’s Obsession with Israel. camera.org. Monitoring organization (advocacy mission).
- CAMERA, 2025. Analysis of BBC News headlines shows scale of anti-Israel bias. camera.org. Monitoring organization (advocacy mission).
- Centre for Media Monitoring, 2025. BBC On Gaza-Israel: One Story, Double Standards. 16 June 2025. cfmm.org.uk. Monitoring organization.
- Elmasry, M. H., 2024. Images of the Israel-Gaza War on Instagram: A Content Analysis of Western Broadcast News Posts. Journalism & Mass Communication Quarterly. DOI 10.1177/10776990241287155. Peer-reviewed.
- Huesmann, L. R., Dubow, E. F., Boxer, P., Souweidane, V., and Ginges, J., 2012. Foreign Wars and Domestic Prejudice: How Media Exposure to the Israeli-Palestinian Conflict Predicts Ethnic Stereotyping by Jewish and Arab American Adolescents. Journal of Research on Adolescence. DOI 10.1111/j.1532-7795.2012.00785.x. Peer-reviewed.
- Huynh, B. Q., Chin, E. T., and Spiegel, P. B., 2024. No evidence of inflated mortality reporting from the Gaza Ministry of Health. The Lancet, 403(10421), pp. 23 to 24. DOI 10.1016/S0140-6736(23)02713-7. Correspondence, editorially reviewed.
- Ministry of Health, Gaza, 2024. Published martyr lists to 30 June 2024, and the completeness statements of 7 January and 24 July 2024. Official record (a party to the war reporting its own count).
- NGO Monitor, 2025. Annual Report 2024. ngo-monitor.org. Monitoring organization (advocacy mission).
- Thussu, D. K., and Chatterjee, A., 2026. The geopolitics of reporting a genocide: an analysis of the coverage of the Israeli invasion of Gaza in Foreign Affairs and the New York Times. Third World Quarterly. DOI 10.1080/01436597.2026.2623199. Peer-reviewed.
- TRT Akademi, 2026. Doğutaş, A. Selective Humanitarianism: Western Media Bias and the Silencing of Gaza. TRT Akademi 11(26). DOI 10.37679/trta.1827180. Peer-reviewed journal.
