The Gaza Ministry of Health publishes a cumulative war fatality total and a list of names, and neither carries a field recording who caused a death: on the list the ministry published on 6 April 2024, 11,371 of 33,091 recorded fatalities, about 34%, carried incomplete data by the ministry’s own account. Because the record has no attribution field, the totals neither exclude deaths caused by Palestinian armed groups nor itemise them. The Henry Jackson Society’s 2025 analysis states the data contains deaths of people killed by Hamas itself, and publishes no figure for that component; no other named body publishes one either.
Key Findings
- The Gaza Ministry of Health’s published fatality lists record a person’s name, identity number, sex, age and date of death. None of those five fields records a cause of death or an attributing party, so no subset of the total can be sorted by who caused it.
- On the ministry’s own list of 6 April 2024, 11,371 of 33,091 recorded fatalities, about 34% of entries, carried incomplete data, the ministry’s own count of the unresolved portion of its own registry.
- The Henry Jackson Society (2025), a policy institute and not a peer-reviewed publisher, states the ministry data “contains deaths of those killed by Hamas itself” and makes “no mention of Hamas combatant fatalities,” and separately estimates upwards of 17,000 combatant fatalities unaccounted for. It publishes no figure for the Hamas-caused component.
- The Israel Defense Forces stated at a press briefing on 29 February 2024 that it had eliminated over 13,000 people it describes as terrorists in Gaza, a party’s own count, not independently verified, with no counterpart field in the ministry’s list.
- Elmas and colleagues (2026), in a preprint covering 4,570,633 tweets from 1,521,433 users between 1 September and 15 December 2023, identified 11 coordinated groups involving 541 accounts, of which 12 of 191 highly amplified posts were assessed as misleading.
Does the ministry attribute each death to a specific perpetrator?
Five fields appear against each name on the Gaza Ministry of Health’s published fatality lists: name, identity number, sex, age and date of death. None of them records a cause, a weapon or a responsible party. The ministry publishes one cumulative war total built from those entries, so the record it releases cannot be sorted by who caused a death.
That structural point governs everything downstream. A registry with no attribution field does not screen anything out. It also does not fold anything in as a deliberate act of classification, because no classification of that kind is performed at all. Deaths recorded during the war period enter the list on the same terms regardless of origin, and the published output carries no marker that would let a reader, a relaying agency or the ministry itself separate them afterwards.
A second omission sits alongside the first. Associated Press reporting records that the ministry does not distinguish civilian from military fatalities in its records. BBC reporting states that the casualty list does not differentiate between civilians and members of Palestinian armed groups, and that Gaza officials describe their lists as subject to verification and revision. The two absences are distinct: one concerns who caused a death, the other concerns the status of the person who died, and neither is resolved inside the published file.
The figures reach international readers through relay rather than independent enumeration. The UN Office for the Coordination of Humanitarian Affairs, in its own casualty-data note, identifies the Gaza Ministry of Health as its source for fatalities in Gaza and sets out the limits of what it can verify. A relayed figure inherits the fields of the record it came from, which is why the published counts that circulate carry no cause-of-death breakdown either.
So the direct answer to the question as asked is that the ministry does not account for deaths caused by Hamas, and does not account against them. The instrument was not built to make that distinction, and its output does not contain it.
Is it often said Hamas-caused deaths are folded in?
It is often said that the Gaza figures quietly absorb deaths that Hamas itself caused. The Henry Jackson Society’s 2025 report, a policy-institute analysis and not peer-reviewed, states the ministry data “contains deaths of those killed by Hamas itself,” alongside deaths from natural causes and deaths that occurred before the war, and observes that the same data makes “no mention of Hamas combatant fatalities.” The report attaches no number to the Hamas-caused component.
The directionally correct part of the claim follows from the previous section and can be stated flatly. No mechanism in the ministry’s published record removes such deaths, because no field distinguishes them. Anyone asserting that the totals exclude deaths caused by Palestinian armed groups would need a source in the record for that exclusion, and the record contains none.
The reframe is the quantity. The same 2025 report that asserts inclusion estimates upwards of 17,000 Hamas and affiliated combatants unaccounted for in media use of the totals, which is a claim about combatant status rather than about who fired. Those are two different quantities: one asks whether a person recorded as a fatality was a combatant, the other asks whether the fatal event was caused by Israeli or Palestinian fire. A combatant killed by an Israeli strike is not a Hamas-caused death. A civilian killed by a rocket that fell short inside Gaza is not a combatant. The estimate speaks to the first question and leaves the second where it was.
What exists in the record, then, is an assertion of inclusion without a measured share, published by a named institute with a stated policy mission, set against a ministry registry that cannot confirm or refute it from its own fields. That is the state of the evidence at the point the question is usually settled in one direction or the other, and both directions run past it.
What does Israel’s own record say about misfired rockets?
The Israel Defense Forces stated at a press briefing by Rear Admiral Daniel Hagari on 29 February 2024 that it had eliminated over 13,000 people it describes as terrorists in Gaza since the beginning of the war. That is the IDF’s own count, published by a party to the events it describes and not independently verified, and the classification of those counted is the IDF’s own characterisation.
On deaths from launches that failed inside Gaza, the Israeli record is specific in instances and silent in aggregate. Israel attributed the blast at al-Ahli Arab Hospital on 17 October 2023 to a failed rocket launch by Palestinian Islamic Jihad; Palestinian authorities attributed the same event to an Israeli strike, and the two accounts have never been reconciled in any jointly accepted record. Beyond named incidents of that kind, the IDF publishes no cumulative count of rockets that fell short into Gaza, and no count of deaths it associates with them.
The IDF’s stated precautions carry the same shape. Its published materials describe telephone warnings, text messages, leaflets, roof-knocking before strikes on buildings, aborted strikes and numbered evacuation areas, and its published position holds that Hamas sites military infrastructure in residential buildings, hospitals, mosques and schools. None of those pages attaches a number to warnings issued, strikes aborted or launches intercepted inside the Strip.
Where Israel does publish counts on conduct, they are counts of process. The Military Advocate General’s Corps reported that, as at 3 August 2024, 74 criminal investigations had been opened into alleged misconduct in the Gaza war, with over 1,000 incidents referred to the General Staff’s Fact-Finding and Assessment Mechanism for factual assessment. The Corps states its own threshold in the same document: in armed conflict, “death or injury to a civilian or harm to civilian objects, while tragic and regrettable, in most cases do not by themselves indicate a reasonable suspicion of criminal conduct.” Those are counts of investigations opened, not findings, and they run parallel to the fatality registry rather than into it. Two parties publish two sets of numbers, and neither set has a field the other can be joined on.
Has Hamas acknowledged killings it carried out in Gaza?
The Palestinian Central Bureau of Statistics reported 70,942 deaths in Gaza by the end of December 2025, a figure it attributes to the Palestinian Ministry of Health rather than measuring itself. No published accounting sets any killing acknowledged by Hamas against that figure, and no release states whether such deaths sit inside it, outside it, or in the unresolved portion of the registry.
Acknowledgements of internal killings do exist in the public record, separately from any count. Hamas has publicly announced executions of people it accuses of collaboration, and reporting by Euronews in 2025 and by Deutsche Welle covered armed operations conducted by Hamas forces against clans and rival groups inside Gaza during 2024 and 2025. The Jerusalem Post reported in 2025 on Hamas statements concerning Palestinians recorded as killed by Israel. Each of those is press reporting or a party’s own statement, and none carries a reconciliation to the fatality list.
The gap between the two records is the finding. An announced execution and a name on a war fatality list are entries in different systems maintained by the same governing authority, and no published document links them at the level of the individual record, which the Institute does not attempt either, since aggregate figures are the only unit at which this question can be answered. There is no published subtraction, no annotation, and no stated rule for how the fatality registry treats a death the authority has itself announced.
That absence is what the question runs into. The reader asking whether Hamas accounts for deaths it caused is asking for a reconciliation, and the reconciliation is not published by the ministry, by the Bureau, or by any body relaying either. The Bureau’s own December 2025 brief describes the wider demographic change in Gaza, about 254,000 fewer people against pre-aggression population estimates, as a net decline attributable to war-related deaths, missing persons, reduced births and forced out-migration, and it does not decompose that decline by responsible party either.
How much of the ministry’s own data is incomplete?
On the fatality list the Gaza Ministry of Health published on 6 April 2024, 11,371 of 33,091 recorded fatalities carried incomplete data, about 34% of entries, by the ministry’s own publication. That share is the ceiling on any claim about what the registry contains, because a third of the entries were not, on the ministry’s own account, fully resolved at the point of publication.
What “incomplete” covers is set by the ministry, not by an external standard, and the ministry did not publish a breakdown of which fields were missing from which entries. An entry lacking an identity number and an entry lacking a date of death are both incomplete, and the two carry different implications for whether the person can be matched against any other record.
The share moved across releases. Analysis published by the Washington Institute in 2025 tracked how the fully identified proportion of the Gaza fatality figures shifted between successive ministry and United Nations releases, and how the composition of the identified portion changed with it. That analysis is a policy-institute publication and not peer-reviewed, and it measures record-level identification rather than total mortality.
The consequence for the question is direct. A registry with a large unresolved fraction cannot answer a question about the origin of any subset within it, in either direction. The incomplete third is not evidence that Hamas-caused deaths are inside the total, and it is not evidence that they are absent. It fixes how much of the record was, at that date, not yet in a state where any such assessment could be made from the file itself.
Why do names appear, disappear and reappear on the lists?
Entries have been removed from Gaza fatality lists between releases and, in some cases, restored on later ones. Euronews reported in 2025 on research alleging manipulation of the toll after thousands of names moved between successive publications, and analyses published by the Washington Institute and by the American Enterprise Institute in 2025 documented the same pattern of revision across ministry and United Nations releases.
Two readings of the same movement circulate, and they measure different things. Gaza officials, as recorded in BBC reporting, describe the lists as provisional and subject to verification and revision, which is the ordinary behaviour of a registry that publishes before identification is complete. The named analyses above read the same removals and restorations as evidence of record-level unreliability. Both readings apply to the same file, and the published data does not include a change log that would separate a correction from anything else.
Revision is also a predictable consequence of the completeness figure in the previous section. A list of which roughly a third of entries carried incomplete data at publication is a list that will change, because resolving an incomplete entry either completes it, merges it with a duplicate, or removes it. What the published releases do not carry is a field distinguishing those three outcomes from one another.
The distinction that matters for the question asked is that list revision and cause-of-death attribution are separate problems. A name removed and later restored tells a reader something about the integrity of the record-keeping. It tells them nothing about whether the death was caused by Israeli fire, by a launch that fell short, or by an internal killing, because that field was never in the file to begin with. A report that treats revision as evidence about attribution has substituted one question for another, and the record supports neither substitution.
Where do statisticians disagree about the registry’s reliability?
The Jerusalem Center for Security and Foreign Affairs published in 2025 a ceiling of at most 41,000 civilians in Gaza who have died of war-related causes, derived by subtracting estimated natural deaths and Israel’s count of Hamas fighters killed from the Gaza Ministry of Health total. The Center states on the same page that neither the Hamas claim of 67,000 total deaths nor the Israeli estimate of 21,000 fighters killed has been independently confirmed.
That is the shape of the disagreement: the disputing parties are measuring different quantities, and the quantities are being compared as though they were one. Le Monde reported in 2024 on researchers who assess the ministry count as broadly reliable as an estimate of total mortality, an assessment about the aggregate. The Henry Jackson Society’s 2025 report assesses record-level integrity, meaning what the individual entries contain and how they change. The Washington Institute’s 2025 analyses assess the demographic composition of the identified entries, including the argument that men are undercounted relative to the population structure.
Set side by side, those three assessments are not contradictory in the way a public dispute implies. A registry can be a defensible estimate of how many people died and a poor instrument for establishing who they were or who killed them, and the named parties are each speaking to one of those properties. None of the three publishes a measured figure for deaths caused by Palestinian armed groups, and none claims to.
The tier labels matter for reading them. The Le Monde piece relays the assessments of named researchers rather than measuring mortality itself. The Henry Jackson Society, the Washington Institute, the American Enterprise Institute and the Jerusalem Center are policy institutes publishing outside peer review, several with stated missions. The ministry’s own figures are a party’s own count. The documentary record on civilian harm in this conflict is assembled almost entirely from those categories, and the strongest available assessments sit at the level of total mortality, which is not the level at which this question is asked.
How does the casualty-figure dispute travel online?
Elmas and colleagues (2026) identified 11 coordinated groups driving the casualty-figure dispute across X, according to their preprint.
The preprint by Elmas and colleagues (2026) drew on 4,570,633 tweets from 1,521,433 users on X between 1 September and 15 December 2023, and found that these coordinated groups involved 541 accounts.
Of 191 highly amplified coordinated posts, 12 were assessed as misleading, concentrated in 3 of the 11 groups; the authors describe the landscape as fragmented, without centralised control, and note the campaigns received little engagement from non-coordinated accounts.
The volume around the conflict was large from the outset. Chen and colleagues (2024), in a preprint dataset drawn from 25 conflict-centric and 75 conflict-inclusive subreddits between August and November 2023, recorded 412,258 submissions and 8,089,095 comments, of which the Zionism and antisemitism topic set held 126,107 submissions and 2,516,114 comments. Guerra and colleagues (2024), in a preprint scoring over 450,000 posts and comments across four subreddits between 7 October 2023 and 3 January 2024, recorded daily post volume of around 10,000 at the conflict’s onset, with a seven-day moving average above 8,000 through November before levelling to about 6,000 in December.
Those studies measure the volume and the coordination structure of the conversation. None of them measures the accuracy of a casualty figure, and none was designed to. Elmas and colleagues note that the coordinated groups they identified relied on low-complexity tactics, retweet amplification and copy-paste diffusion, and that most of the groups were engaged in advocacy, religious solidarity or humanitarian mobilisation rather than misleading content. Guerra and colleagues report that peaks in their extremity score coincided with offline conflict events, a coincidence in time and not a measured cause.
What the three corpora establish is that a contested count circulates at a scale where the underlying record’s fields are not visible. A figure that carries no attribution column in its source file travels as a plain number, and the constraints described in the sections above do not travel with it.
Methodology and limitations
This report measures what the published records contain and what each named body states, not the number of people killed in Gaza and not the cause of any death.
Sources are drawn from four tiers, each labelled where it appears. Parties’ own counts: the Gaza Ministry of Health’s published fatality lists, including the list of 6 April 2024; the Israel Defense Forces press briefing of 29 February 2024 and the Military Advocate General’s Corps publication of 3 August 2024; the Palestinian Central Bureau of Statistics brief of December 2025, which attributes its Gaza death figure to the Palestinian Ministry of Health rather than measuring it. Each is published by a party to the events it describes and is not independently verified. Policy-institute analyses, none peer-reviewed and several with stated missions: the Henry Jackson Society (2025), the Washington Institute (2025), the American Enterprise Institute (2025) and the Jerusalem Center for Security and Foreign Affairs (2025). Press reporting: the Associated Press, the BBC, Le Monde (2024), Euronews (2025), Deutsche Welle and the Jerusalem Post (2025), each cited for what it reported and dated. Peer-reviewed and preprint research on the online discourse: Elmas and colleagues (2026), Chen and colleagues (2024) and Guerra and colleagues (2024), all preprints, cited only for their own measured volume, coordination and extremity metrics.
Where the literature disagrees, the competing assessments are named and the quantity each measures is stated: total mortality (researchers summarised by Le Monde, 2024), record-level integrity (Henry Jackson Society, 2025), and demographic composition of the identified entries (Washington Institute, 2025). Those are three properties of one registry and are not interchangeable.
Two limits bound every figure here. No source located for this report publishes a measured count, share or estimate of deaths within the Gaza fatality total caused by Palestinian armed groups, so no such quantity appears here as an established figure. And on the ministry’s own 6 April 2024 publication, about 34% of entries carried incomplete data, which caps what can be established from the file at the level of the individual record. The discourse studies are correlational: their extremity and volume peaks coincided with offline events, and the authors claim no causal relationship.
Conclusion
The question asks whether Hamas accounts for deaths caused by Hamas itself, and the record answers it structurally rather than numerically. The Gaza Ministry of Health’s published list carries a name, an identity number, a sex, an age and a date, and no field for cause or actor. A count built that way cannot exclude such deaths and does not itemise them, and no release from the ministry, the Palestinian Central Bureau of Statistics or any relaying agency supplies the missing column.
Everything else in the record sits beside that fact rather than resolving it. About 34% of entries on the 6 April 2024 list were incomplete on the ministry’s own account. The Israel Defense Forces publishes its own count of over 13,000 people it describes as terrorists killed, and 74 criminal investigations opened by August 2024 under a threshold it states plainly, with no field joining either to the fatality registry. The Henry Jackson Society asserts inclusion of Hamas-caused deaths and publishes no share. The Jerusalem Center derives a ceiling of at most 41,000 civilians from inputs it says are unconfirmed. Named researchers summarised by Le Monde find the aggregate broadly reliable as total mortality while other named analyses find the individual entries unstable, and both can hold at once because they measure different properties. Two parties keep two sets of books, and the books have no common key.
That is the condition in which the number travels. Elmas and colleagues counted 4,570,633 tweets in fifteen weeks and 541 coordinated accounts inside them; Chen and colleagues counted 8,089,095 Reddit comments in four months. At that volume a figure circulates stripped of the fields its source file never had, and each side of the argument reads its own certainty into an attribution column that does not exist. When a count carries no record of who caused what, the certainty has to come from somewhere, and the tropes about who is lying and who is counting bodies for advantage are old and already in circulation. What does a public argument do when its central number is structurally incapable of settling it?
Frequently Asked Questions
Who compiles the Gaza fatality figures that international agencies publish?
The Gaza Ministry of Health compiles them, and international agencies relay them. The UN Office for the Coordination of Humanitarian Affairs identifies the ministry as its source for Gaza fatalities in its own casualty-data note and sets out the limits of what it can verify, so a figure appearing under an agency’s name is generally the ministry’s count carried forward.
Does the ministry’s total include deaths from causes other than the war?
The Henry Jackson Society’s 2025 analysis states that the ministry data contains deaths from natural causes and deaths that occurred before the war, in addition to war deaths. The ministry’s published lists carry no cause-of-death field, so no released file separates those categories, and the institute attaches no figure to either.
Has Israel published its own figure for combatants killed in Gaza?
The Israel Defense Forces stated at a press briefing on 29 February 2024 that it had eliminated over 13,000 people it describes as terrorists in Gaza. The Jerusalem Center for Security and Foreign Affairs (2025) uses a later Israeli estimate of 21,000 fighters killed in its own arithmetic, and states on the same page that this estimate has not been independently confirmed.
What did the Palestinian Central Bureau of Statistics report about Gaza’s population?
The Bureau reported Gaza at about 2.13 million at the end of December 2025, a decline of approximately 254,000 people, which it puts at 10.6% against its pre-aggression population estimates. It attributes the change to war-related deaths, missing persons, reduced births and forced out-migration, and describes it as a net demographic decline rather than a casualty count.
What is the argument that men are undercounted in the figures?
The Washington Institute published analysis in 2025 arguing that the demographic composition of the identified entries in the Gaza fatality data shows fewer adult men than the population structure would predict. It is a policy-institute analysis rather than peer-reviewed work, and it addresses the composition of the identified portion, not total mortality.
How do researchers study the online conversation about the war?
By collecting large platform corpora and scoring them. Chen and colleagues (2024) assembled 412,258 Reddit submissions and 8,089,095 comments from 100 subreddits between August and November 2023, and Guerra and colleagues (2024) scored over 450,000 posts and comments from four subreddits for extremity. Both are preprints and measure discussion volume and tone, not the accuracy of any claim within it.
Sources
- American Enterprise Institute, 2025. Don’t Fall for Hamas’ Numbers Game. AEI op-ed. aei.org. Policy institute, not peer-reviewed.
- Associated Press. Reporting on the Gaza Ministry of Health’s casualty records and their treatment of civilian and military fatalities. Press reporting.
- BBC News, 2024. Reporting on the Gaza casualty lists, their categories and their revision. bbc.com. Press reporting.
- Begin-Sadat Center for Strategic Studies, 2025. The Gaza Terror Offensive, situation report. besacenter.org. Research institute.
- Chen, He, Burghardt, Zhang, Lerman, 2024. IsamasRed: A Public Dataset Tracking Reddit Discussions on the Israel-Hamas Conflict. arXiv:2401.08202. Preprint.
- Deutsche Welle. Reporting on armed operations by Hamas forces inside Gaza and on the Gaza death toll. dw.com. Press reporting.
- Elmas, Silva, Pote, Dey, Chang, Ye, Luceri, Buntain, Ferrara, Flammini, Menczer, 2026. Israel-Hamas War on X: A Case Study of Coordinated Campaigns and Information Integrity. arXiv:2604.10566. Preprint.
- Euronews, 2025. Reporting on research alleging manipulation of the Gaza death toll. euronews.com. Press reporting.
- Gaza Ministry of Health, 2024. Fatality list published 6 April 2024. A party’s own published record.
- Guerra, Lepre, Karakuş, 2024. Quantifying Extreme Opinions on Reddit Amidst the 2023 Israeli-Palestinian Conflict. arXiv:2412.10913. Preprint.
- Henry Jackson Society, 2025. Hamas Casualty Reports are a Tangle of Technical Problems. henryjacksonsociety.org. Policy institute, not peer-reviewed.
- Israel Defense Forces, 2024. Press Briefing by IDF Spokesperson, Rear Admiral Daniel Hagari, 29 February 2024. idf.il. Official record, a party’s own count.
- Israel Defense Forces, Military Advocate General’s Corps, 2024. Addressing Alleged Misconduct in the Context of the War in Gaza, updated 3 August 2024. idf.il. Official record, a party’s own count.
- Jerusalem Center for Security and Foreign Affairs, 2025. The Death of Genocide. jcfa.org. Research institute, not peer-reviewed.
- Jerusalem Post, 2025. Reporting on Hamas statements concerning Palestinians recorded as killed. jpost.com. Press reporting.
- Le Monde, 2024. Why the Gaza Health Ministry’s death count is considered reliable. lemonde.fr. Press reporting summarising named researchers.
- Palestinian Central Bureau of Statistics, 2025. A Brief on the Status of the Palestinian People at the End of 2025. pcbs.gov.ps. Official record.
- Washington Institute for Near East Policy, 2025. Analyses of the revised Gaza fatality data and its demographic composition. washingtoninstitute.org. Policy institute, not peer-reviewed.