The nearest measured figure in the peer-reviewed record is 59.1%, and it does not count what the claim counts. Jamaluddine and colleagues, in a capture-recapture study published in The Lancet in 2025, report 16,699 of the 28,257 Gaza deaths for which age and sex data were available as women, children under 18 and older people 65 and over, over 7 October 2023 to 30 June 2024 and for traumatic injury alone. The figure that explains why any such share moves sits in the same paper: 9,692 decedents in the Gaza Ministry of Health’s own 30 June 2024 update, 26% of its cumulative total, carried no identity, and a decedent recorded without an identity carries no recorded age and no recorded sex. It is often said that 72% of those killed in Gaza are women and children; that claim names two groups, while the published figures name other combinations, over other windows, on other bases, which is why the shares in circulation disagree.
Key Findings
- Jamaluddine and colleagues report 16,699 of 28,257 deaths, 59.1%, as women, children under 18 and older people 65 and over, in a peer-reviewed Lancet capture-recapture study covering 7 October 2023 to 30 June 2024 and traumatic injury only.
- The Lancet study records 9,692 unidentified decedents in the Gaza Ministry of Health’s 30 June 2024 update, 26% of that ministry’s cumulative total, plus 930 people reported missing from the survey list; neither group can enter a demographic share in either direction.
- The study’s central estimate of deaths from traumatic injury in that window is 64,260, with a 95% confidence interval of 55,298 to 78,525, against the 37,877 decedents the Gaza Ministry of Health had reported by 30 June 2024, of whom 28,185 were identified (Jamaluddine and colleagues, 2025).
- Crawford’s 2025 Costs of War paper records 67,075 killed and 169,430 injured in Gaza as of 3 October 2025 as the Gaza Ministry of Health’s own count rather than an independently verified one, and states that the ministry publishes no combatant-noncombatant breakdown at all.
- The Israel Defense Forces stated on 29 February 2024 that it had eliminated over 13,000 people it describes as terrorists in Gaza, its own reported count on its own classification, written as a floor rather than a total.
What is the closest measured figure to the 72% claim?
59.1% is the closest checkable figure, and it does not measure the same quantity the claim names. Jamaluddine and colleagues, in a peer-reviewed capture-recapture study published in The Lancet in 2025, report 16,699 of the 28,257 Gaza deaths for which age and sex data were available as women, children and older people, over the window from 7 October 2023 to 30 June 2024.
It is often said that 72% of those killed in Gaza are women and children, and the question people ask is whether that figure is accurate. The nearest measurement in the peer-reviewed literature is the study above, and the two are not accounts of the same thing.
Jamaluddine and colleagues built their estimate by capture-recapture, a method that compares separate lists of the dead and uses their overlap to estimate how many deaths none of the lists caught. The three lists were the Palestinian Ministry of Health’s hospital morgue records, carrying 22,368 identified decedents; the same ministry’s rolling online mortality survey, launched on 1 January 2024, carrying 5,817 individual decedents; and obituary pages scraped by hand from social media. The Lancet paper reports a London School of Hygiene and Tropical Medicine ethics reference, publishes its analysis code and anonymised data, and declares no funding.
The scope is narrower than the question usually assumes. The Lancet study estimates deaths due to traumatic injury only, and its authors state that it excludes deaths from healthcare disruption, food insecurity, water and sanitation failure, and disease. The window runs from 7 October 2023 to 30 June 2024 and stops there, so nothing in it describes the composition of the dead after that date.
Two of the three input lists are self-reported, and the hospital list comes from the health ministry of a party to the war, so this is an independent academic estimate built on those inputs rather than an independent count of the dead. The 59.1% is a share of the deaths inside the study for which age and sex were recorded. It is a real, checkable, peer-reviewed figure, and it answers a question about a wider category, a narrower cause of death and a shorter window than the claim it is usually set against (Jamaluddine and colleagues, 2025).
Which three groups does the peer-reviewed share actually combine?
The study’s category holds three groups, not two: women, children defined as under 18, and older people defined as 65 and over. Jamaluddine and colleagues publish no women-and-children-alone share anywhere in the result, so the 59.1% cannot be restated as one, and a reader setting it against a claim about women and children is setting a three-category share against a two-category claim.
The Lancet paper reports the category as a single combined group and publishes no breakdown inside it. There is no women-only figure, no children-only figure and no older-people-only figure in the reported result, so the share cannot be split, and it cannot be narrowed to the two groups the claim names.
The definitional choices inside the category govern what the share means. Children are defined as people under 18, which places 16-year-olds and 17-year-olds inside the category. Older people are defined as those 65 and over, a group the claim about women and children does not mention at all. Men aged 18 to 64 form the complement, and the paper does not report them separately either.
The denominator governs as much as the category does. The 59.1% is a share of 28,257, the deaths in the study for which age and sex data were available. It is not a share of the study’s own central estimate of the toll for the same window, and multiplying that estimate by 59.1% produces a figure that appears nowhere in the paper (Jamaluddine and colleagues, 2025).
So a reader who sets 59.1% beside a claim about women and children is not comparing a lower estimate with a higher one. The comparison runs between a three-category share, computed over deaths with age and sex recorded in a window of just under nine months, and a two-category claim whose own base is usually unstated. Two figures built that differently can disagree without either being an account of the other, and neither becomes the other’s correction.
How many of the dead carried no recorded age or sex?
9,692 decedents carried no identity. Jamaluddine and colleagues record that number in the Gaza Ministry of Health’s own 30 June 2024 update, 26% of that ministry’s cumulative total at the date, and exclude them from the hospital list used in the analysis. A decedent recorded without an identity carries no recorded age and no recorded sex.
The Lancet paper records 930 people reported missing from the survey list, likewise excluded. Neither exclusion is a comment on the ministry’s record-keeping; both are a property of what a death record can hold. A decedent entered without a name has no date of birth attached and no sex attached, so that death cannot be placed on either side of a demographic share.
The consequence is arithmetic, and it is the Institute’s own arithmetic on the study’s published figures rather than a finding of the study. Any share of women and children is computed over the identified portion of a count. When the identified portion changes, whether because more bodies are identified, because a reporting body changes what it publishes, or because the record for an earlier period is completed late, the share computed over it changes without anything about the dead having changed.
That is why a composition share for Gaza is a moving quantity rather than a fixed one, and why two shares published months apart can differ without either being a revision of the other. The split below is the one the ministry’s own update recorded at the end of June 2024.
Both parts of that split are the ministry’s own categories in the ministry’s own update, not an independently verified classification, and the study takes them as it finds them. A quarter of the cumulative total sits in the part that no demographic share can reach (Jamaluddine and colleagues, 2025).
How much of the estimated toll does that share cover?
64,260 is the study’s own central estimate of deaths from traumatic injury over the window, with a 95% confidence interval of 55,298 to 78,525. Jamaluddine and colleagues set that estimate against the 37,877 decedents the Gaza Ministry of Health had reported by 30 June 2024, the ministry’s own count.
The 28,257-death base under the 59.1% is therefore smaller than the study’s own central estimate for the same window, and smaller than the count the ministry itself had published by the same date. All three numbers are printed here; any ratio between them is the Institute’s own arithmetic and is labelled as such.
The gap between the estimate and the reported count is 26,383 deaths, and The Lancet paper puts that gap at 41% under-reporting. The 41% is a share of the 64,260 estimate. It is not a multiplier on the 37,877 the ministry reported, and the same gap read against the ministry’s figure is 70% of it, so the two ways of stating one gap are not interchangeable.
The estimate’s own bounds are wide. The interval runs from 55,298 to 78,525, a span of more than 23,000 deaths, and The Lancet paper reports the point estimate with that interval rather than as a settled total.
And the estimate covers traumatic injury alone. Deaths from healthcare disruption, from food insecurity, from water and sanitation failure and from disease sit outside it entirely, which means the toll it estimates is narrower than the toll most readers have in mind when they ask what share of the dead were women and children.
How does the ministry’s own count reach a reader?
67,075 killed and 169,430 injured in Gaza as of 3 October 2025 is the figure Crawford’s 2025 Costs of War paper records, and it is entered there as the Gaza Ministry of Health’s own count rather than a measurement of the project’s own.
The Costs of War footnote records the route the pair travelled: the ministry published it, a United Nations casualties dashboard carried it, and the paper compiled it. The body that produced the figure is the Gaza Ministry of Health, a party to the war, and the dashboard and the paper carried it forward without measuring it.
A second relayer carries a later figure. The Palestinian Central Bureau of Statistics reports 70,942 deaths in Gaza by the end of December 2025 in its brief on the status of the Palestinian people, attributed to the Palestinian Ministry of Health and not measured by the Bureau. Two publishers, one reporting body, and both label the figure as the ministry’s own rather than as an independently verified one.
The two counts are about three months apart and differ by 3,867. Neither publisher measured either figure, and neither presents it as verified. What a reader receives from either is the same object: a count kept by a health ministry during a war, carried forward by a body that names where it came from (Crawford, 2025).
Crawford also relays a reading of that count which is not his project’s own. He writes that “Recent analysis by public health experts suggests that the number of fatalities reported by the Gaza Ministry of Health, which faces many obstacles to making a full account of the deaths, may be a significant undercount of the violent deaths.” The sentence is a report of what other work suggests, and the paper states it that way rather than as a finding of its own; the argument over whether the registry undercounts or overcounts the dead runs on a different set of figures from the ones read here.
Does the ministry publish a combatant or civilian breakdown?
The Gaza Ministry of Health publishes no combatant-noncombatant breakdown at all, in the words of Crawford’s 2025 Costs of War paper, so a demographic share is not a civilian share and cannot be converted into one. The about-80% civilian figure printed in that paper’s own Table 1 is not measured there.
What the ministry publishes is a running toll of deaths and injuries. What it does not publish is any division of that toll into combatants and noncombatants, which the Costs of War paper states in terms.
The civilian share in the paper’s table is therefore not a measurement of the paper’s own. Crawford derives it from a newspaper’s reading of Israeli military data, a party’s own data, and he flags the assumption in his own footnote (Crawford, 2025). A combatant-civilian split belongs to that reporting and to the military data underneath it, not to this paper, and the paper is not the place to cite one from.
The distinction bears directly on the claim, because a demographic share and a civilian share are different quantities that are routinely read as one. A count of women, children and older people is a count by age and sex, taken from a death record. A count of civilians is a count by status, and status is not recorded on a death certificate. Nothing in the age-and-sex data published for Gaza assigns status to anyone, and nothing in the status figures that circulate is broken down by age and sex.
So the 59.1% is not a civilian share, a claim about women and children is not a claim about civilians, and the body publishing the toll publishes no civilian share at all. A reader who converts one into the other has crossed between two categories that no published body in this record maps onto each other, and has done so without a bridge that any of them supplies.
What does the other party count, and by whose classification?
Over 13,000 people it describes as terrorists is what the Israel Defense Forces stated on 29 February 2024 it had eliminated in Gaza since the beginning of the war. That is the force’s own reported count on its own classification, not independently verified, and the page writes it as over, so it is a floor rather than a total.
The classification is that party’s own characterisation rather than a neutral or verified one, and combatant and civilian counts in this war are contested. Nothing in the figure records age or sex, and nothing in it distinguishes the people counted from any of the people counted in the ministry’s toll.
Two properties keep the figure separate from the Lancet study’s base. Its window ends on 29 February 2024, four months before the study’s window closes on 30 June 2024, so the two describe different periods of the war. And its category is one the Gaza Ministry of Health does not publish, so neither the ministry’s record nor the study’s demographic share is its complement.
That is why the complement of a demographic share cannot be derived by subtraction. Taking one party’s count of the people it classifies as combatants, subtracting it from another body’s cumulative total, and reading the remainder as women, children and civilians combines counts from bodies with different definitions, different universes and different closing dates, and produces a number neither body published.
The Israel Defense Forces published no age or sex breakdown alongside that figure, and the Gaza Ministry of Health publishes a running toll with no status breakdown. The two records were built for different purposes by parties on opposite sides of the same war, and neither is the other’s remainder. Any percentage that treats them as complements has assumed the bridge rather than found it.
Does any primary record count the children killed in Gaza?
The peer-reviewed paper closest to this subject in the record read here counts nothing itself (Hamamra, Mahamid and Mayaleh, 2025): every figure in it is relayed and belongs to the body that produced it. No primary record read for this report counts the children killed by military violence in Gaza.
The study is qualitative. Its authors conducted 30 semi-structured interviews with displaced adult caregivers of children under 18 in the Rafah displacement camps, 15 women and 15 men, analysed by thematic content analysis, with thematic saturation reported at the 25th interview. They report six themes: widespread malnutrition, diseases and epidemics in shelters, unsafe environments, weak immune systems and growth, inaccessibility of healthcare, and daily struggles for basic needs (Hamamra, Mahamid and Mayaleh, 2025). They examine no patients, sample no population and report no clinical series.
The one figure in it that touches child mortality belongs to another body. The paper attributes to the Palestinian Central Bureau of Statistics a count of 28 children who died due to malnutrition and dehydration in the first 6 months of the war. That is a statistical office’s figure carried by the paper, it counts deaths from malnutrition and dehydration rather than from military violence, and it covers a six-month period rather than the war.
The authors’ stated conclusion, that the crisis is a deliberate outcome of Israeli policies designed to undermine civilian survival, is their reading of interview themes rather than a measured attribution of cause, and the characterisation in the paper’s own title is likewise theirs.
So the published work read here yields no child-death total for Gaza, and this report gives none. That absence is not an argument that no children have been killed, and it is not a claim about the size of any figure in circulation. It is a statement about what this record contains, which is a study of caregivers, running tolls with no age breakdown attached to them, and one peer-reviewed demographic share in which children are one of three groups counted together.
Which published body reports women and children as one category?
Jamaluddine and colleagues combine three age-and-sex categories, women, children under 18 and people 65 and over, and publish no two-category share; Crawford relays the ministry’s totals with no demographic breakdown at all. None of the bodies read for this report publishes women and children as one share of the Gaza dead.
The Palestinian Central Bureau of Statistics carries a total attributed to the ministry and measures something different itself: a net demographic change for Gaza. It reports a decline of approximately 254,000 people, which it states is a decrease of 10.6% against pre-aggression population estimates, and attributes that change to war-related deaths, missing persons, reduced births and forced out-migration. It is a population accounting rather than a casualty count, and it cannot be converted into one.
So the quantity the question names, women and children as a single share of the Gaza dead, is published by no body in the record read for this report. The nearest peer-reviewed figure combines three groups. The running tolls carry no demographic breakdown. The population brief measures a different thing entirely, and says so in its own text.
That absence is worth recording plainly, because it accounts for much of the disagreement. When a widely repeated share has no publishing body behind it, every version of it is somebody’s arithmetic on somebody else’s base, and two people quoting different percentages can each be quoting accurately from a different denominator over a different window. The same property runs through every casualty figure for Gaza, where what each body counts differs before any two of them disagree. Not all of the disagreement is about the dead. A great deal of it is about what each number was a share of, and that part of it is settled by reading the base rather than by arguing about the total.
Methodology and limitations
This report reads what each published body counted, on what base and over what window, and it draws on five source types, each labelled where it is used.
One peer-reviewed journal article supplies the composition figure: Jamaluddine and colleagues in The Lancet, a three-list capture-recapture estimate. Its three inputs were the Palestinian Ministry of Health’s hospital morgue records, the same ministry’s rolling online mortality survey, and manually scraped social media obituary pages. Two of the three are self-reported and the hospital list comes from the health ministry of a party to the war, so the result is an independent academic estimate built on those inputs, not an independent count. The window closes on 30 June 2024, which means nothing here speaks to the composition of the dead after that date, including anything after the ceasefire.
Three arithmetic limits travel with that study. The 59.1% has a denominator of 28,257 and never of 64,260, so applying it to the estimate yields a figure the paper does not report. The 41% under-reporting figure is a share of the 64,260 estimate and never a multiplier on the ministry’s 37,877. And the estimate covers traumatic injury only.
One university research-project report, the Costs of War paper (Crawford, 2025), is not peer-reviewed and compiles casualty counts rather than measuring them; its Gaza figures are the Gaza Ministry of Health’s own. One national statistical office publication, the Palestinian Central Bureau of Statistics brief for the end of 2025, relays a ministry total and separately measures a net population change. One official record was published by a party to the conflict: the Israel Defense Forces stated its figure at a briefing of 29 February 2024, that force’s own count on its own classification, given as a floor. One further peer-reviewed article, the Rafah caregiver study (Hamamra, Mahamid and Mayaleh, 2025), is qualitative and relays every figure it carries.
Two absences are stated rather than filled. No source read here gives Gaza’s own age structure, so this report cannot say what share of the living population is under 18, which is the baseline against which any composition share should be read: the Palestinian Central Bureau of Statistics counted 1,899,291 residents of the Gaza Strip on 1 December 2017 in the most recent full census, and no age breakdown from it is carried here. And no source read here publishes women and children as a single category.
Counts from bodies with different definitions, universes and dates are not summed or subtracted anywhere in this report. Where two figures are placed side by side, the comparison is stated as the Institute’s own arithmetic.
Conclusion
So is 72% accurate? The peer-reviewed record does not contain that quantity to be accurate or inaccurate about. What it contains is 16,699 of 28,257, 59.1%, for women, children under 18 and older people 65 and over, over a window that closed on 30 June 2024, for deaths from traumatic injury alone, computed over a base from which 9,692 unidentified decedents and 930 people reported missing were absent by construction (Jamaluddine and colleagues, 2025).
The rest of the record is a set of counts that do not line up, and each one says plainly what it is. The Costs of War paper records 67,075 dead and 169,430 injured to 3 October 2025 as the ministry’s own count, and states that the ministry publishes no combatant-noncombatant split at all. The Palestinian Central Bureau of Statistics carries 70,942 to the end of December 2025, attributed to the same ministry and measured by nobody else, and measures a population decline of about 254,000 that it does not present as a death toll. The Israel Defense Forces stated over 13,000 killed in a category of its own definition, four months before the study’s window even closes. Not one of these bodies publishes women and children as a single share, and the closest peer-reviewed figure covers three groups on a base that leaves out a quarter of the ministry’s own cumulative total.
A percentage is a promise about a denominator. Every share of Gaza’s dead that travels through public argument arrives with the denominator stripped off, and the denominator is where most of the disagreement actually lives: which deaths, over which months, identified by whom, and recorded by a body with what stake in the recording. When a quarter of a cumulative total carries no name, no age and no sex, what is actually being handed over when a single percentage is handed over, and who decided what it was a share of?
Frequently Asked Questions
Which deaths does the Lancet estimate leave out?
Jamaluddine and colleagues estimate deaths due to traumatic injury only. The Lancet paper states that deaths from healthcare disruption, food insecurity, water and sanitation failure and disease sit outside the estimate, so 64,260 is not a figure for all war-related mortality in the window.
What does the 41% under-reporting figure actually measure?
It measures the gap between the study’s estimate and the Gaza Ministry of Health’s reported count as a share of the estimate. Jamaluddine and colleagues estimate 64,260 deaths against 37,877 reported by 30 June 2024, a gap of 26,383, which is 41% of the estimate and 70% of the ministry’s figure.
How fast were people dying, on the study’s own rate?
Jamaluddine and colleagues report an annualised crude death rate of 39.3 per 1,000 people, with a 95% confidence interval of 35.7 to 49.4. They put that at a rate ratio of 14.0, with a confidence interval of 12.8 to 17.6, against all-cause mortality in 2022.
Where did the study’s three lists come from?
The first was the Palestinian Ministry of Health’s hospital morgue records, carrying 22,368 identified decedents. The second was the same ministry’s rolling online mortality survey, launched on 1 January 2024, carrying 5,817 individual decedents. The third was obituary pages scraped by hand from social media.
How many people were injured in Gaza on the ministry’s own count?
Crawford’s 2025 Costs of War paper records 169,430 injured as of 3 October 2025 alongside 67,075 killed, both attributed to the Gaza Ministry of Health as that ministry’s own count. Crawford compiled the pair rather than measuring it.
How many people did the last census count in the Gaza Strip?
The Palestinian Central Bureau of Statistics counted 1,899,291 residents of the Gaza Strip on the reference date of 1 December 2017, in the third round of the Palestinian census. It is the most recent full census, so every later total for the territory is a projection off that measured base.
What did the Rafah caregiver study actually measure?
The Rafah caregiver study conducted 30 semi-structured interviews with displaced adult caregivers and coded six themes by thematic content analysis (Hamamra, Mahamid and Mayaleh, 2025). It reports no prevalence of its own, and the figures inside it, including a count of 28 child deaths from malnutrition and dehydration, belong to the bodies it relays them from.
Does the Palestinian statistical office measure the death toll itself?
No. The Palestinian Central Bureau of Statistics reports 70,942 deaths in Gaza by the end of December 2025 and attributes that figure to the Palestinian Ministry of Health. What the Bureau measures itself is a population decline of approximately 254,000 people, 10.6% against pre-aggression population estimates, which it attributes to deaths, missing persons, reduced births and out-migration.
Sources
- Crawford, Neta C., 2025. The Human Toll of the Gaza War: Direct and Indirect Death from 7 October 2023 to 3 October 2025. Costs of War project, Watson School of International and Public Affairs, Brown University, 7 October 2025. costsofwar.watson.brown.edu. University research-project report, not peer-reviewed.
- Hamamra, Bilal, Mahamid, Fayez, and Mayaleh, Asala, 2025. Childhood under siege: battling malnutrition, disease, and despair during genocide in Gaza. Journal of Health, Population and Nutrition, 44:321. DOI 10.1186/s41043-025-00948-6. Peer-reviewed journal, qualitative interview study.
- Israel Defense Forces, 2024. Press Briefing by IDF Spokesperson, Rear Admiral Daniel Hagari, February 29th, 2024. idf.il. Official record, self-reported by a party to the conflict.
- Jamaluddine, Zeina, Abukmail, Hanan, Aly, Sarah, Campbell, Oona M. R., and Checchi, Francesco, 2025. Traumatic injury mortality in the Gaza Strip from Oct 7, 2023, to June 30, 2024: a capture-recapture analysis. The Lancet, 405(10477), pp. 469 to 477. DOI 10.1016/S0140-6736(24)02678-3. PMID 39799952. Peer-reviewed journal.
- Palestinian Central Bureau of Statistics, 2017. Population, Housing and Establishment Census 2017 (Third Round): Final Results. Ramallah, State of Palestine. pcbs.gov.ps. Official record, national census.
- Palestinian Central Bureau of Statistics, 2025. A Brief on the Status of the Palestinian People at the End of 2025. Ramallah. pcbs.gov.ps. Official record, national statistical office.