An Engineered Tragedy: Statistical Analysis of Casualties in the Palestinian-Israeli Conflict

International Policy Institute for CounterTerrorism ^ | June 24, 2002 Updated: July 7, 2002 | Don Radlauer

[SJackson]

<p>

June 24, 2002

Updated: July 7, 2002

An Engineered Tragedy

Statistical Analysis of Casualties in the Palestinian - Israeli Conflict,
September 2000 - June 2002

Don Radlauer
ICT Associate

Introduction

For the last 21 months, Israel and the Palestinian Arabs have been engaged in a “low-intensity conflict” generally referred to as the “al-Aqsa Intifada”. This conflict has caused around two thousand deaths so far; and yet, its most significant aspect seems to be the struggle for international public opinion rather than any effort at gaining a conventional military victory. New reports of death and injury appear almost daily in the world’s news media, generally accompanied by the current tally of the total number of people killed on each side. Pundits and laymen read these reports and draw conclusions from the simplistic statistics they convey.

Before many months of this conflict had passed, it became apparent to some observers that the “fatality scorecard” commonly included in coverage of the al-Aqsa conflict was painting an oversimplified and deceptive picture of a complex reality. A more thorough accounting and analysis of the conflict’s incidents and casualties should enable a better understanding of the true nature of the conflict. Accordingly, the International Policy Institute for Counter-Terrorism (ICT) launched the Al-Aqsa Casualties Database Project to provide such an accounting. (See Project Summary, for a brief review of the project’s findings.)

A Note on Terminology

The word “intifada” is properly translated as “uprising”. The use of either of these terms implies a judgement as to the nature of the conflict – specifically that, like the earlier Intifada of 1987-1991, this “intifada” is a spontaneous and authentic expression of “popular rage at Israeli occupation”. Some supporters of Israel, realizing the political significance of the common, often unthinking use of such terms, have attempted to substitute other names for the conflict, such as “the Oslo War”. None of these alternative names has gained general currency.

As this study aims to investigate the true nature of the ongoing Israeli/Palestinian conflict, it seems most appropriate to avoid the use of all such value-laden terms. Accordingly, we decided to use the more neutral and general term “al-Aqsa conflict” to describe the events which began in September 2000. Even this usage can be criticized, in that it suggests some specific causal connection between the conflict and the al-Aqsa Mosque; but it seems to be a useful compromise.

Summary of Findings

Our research and analysis shows that the al-Aqsa conflict is different in many respects from what it is generally believed to be. Among our findings are the following:

• The al-Aqsa conflict is not a single, homogenous event. To describe it glibly as “an uprising against Israeli occupation” grossly oversimplifies and distorts our view of the conflict, and obscures much of the significance of its history. The conflict has, in fact, gone through three distinct phases from its beginning until now. Judging by fatality statistics, only the first of these phases, which lasted less than three months, strongly resembled a true “intifada”. The especially high rate of death among the Palestinians during this phase probably created the general public impression of the entire conflict as an overwhelming slaughter of overmatched Palestinians.

• The usual fatality count quoted in news articles presents an inaccurate and distorted picture of the al-Aqsa conflict, exaggerating Israel’s responsibility for the death of noncombatant civilians. For example, our database shows a total of 561 Israelis killed, compared to around 1499 Palestinians, up to 30 June 2002 – numbers in general agreement with media reports (see Graph 1.1).
  

  But such numbers distort the true picture: They lump combatants in with noncombatants, suicide bombers with innocent civilians, and report Palestinian “collaborators” murdered by their own compatriots as if they had been killed by Israel. Correcting for such distortions, we can arrive at a figure of 579 Palestinian noncombatants killed by Israel, compared to 433 Israeli noncombatants killed by Palestinians (see Graph 1.2). While Israelis account for a little more than 27 percent of the total fatalities as generally reported, they represent some 43 percent of these noncombatant victims. There are a number of valid ways of arriving at such corrected figures to compare the extent to which each side has been responsible for the killing of noncombatants; they all show a much less lopsided picture of the conflict than the raw totals do.

  

  
  

• While Israeli fatalities in the al-Aqsa conflict have consisted of almost 80 percent noncombatants (and over 80 percent before the substantial IDF casualties suffered during the Jenin incursion of April 2002), Palestinian fatalities have consisted of more combatants than noncombatants (see Graph 1.3). The proportion of combatants among Palestinians killed has increased significantly over time, from around 40 percent to nearly 55 percent. (It is also worth noting that the figure for Palestinian noncombatants includes a large number of fatalities for which combatant status could not be determined – see “Combatants and Noncombatants”, below.)

  

  
  

• The proportion of females among Israeli fatalities was relatively low in the early months of the conflict, and gradually rose to a level of around 30 percent; since June 2001, this proportion has remained fairly stable. Palestinian fatalities, in contrast, have been consistently and overwhelmingly (over 95 percent) male (see Graph 1.4).

If we restrict our view to each side's noncombatants killed by the opposing side, the gap in the percentage of females among those killed is even wider: 39 percent of Israeli noncombatants killed by Palestinians have been female, compared to 7 percent of Palestinian noncombatants killed by Israel (see Graph 1.5).

• In absolute terms, many more Israeli females have been killed than Palestinian females. If we include combatants and fatalities for whom responsibility is unclear, 61 Palestinian females have been killed; the corresponding Israeli figure is 170 (see Graph 1.6).

Restricting ourselves to cases where clear responsibility can be reliably assigned for noncombatant deaths, we see that Israel has been responsible for killing 43 Palestinian noncombatant females, while Palestinians have killed 168 Israeli noncombatant females – almost four times as many (see Graph 1.7).






• More than twice as many Israeli noncombatants aged 40 and over have been killed as Palestinian noncombatants 40 and over (see Graph 1.8). Israelis have killed 70 noncombatant "mature" Palestinians, while Palestinians have killed 167 "mature" Israeli noncombatants.
  

• The number of Palestinians killed by actions of their own side – including suicide bombers, “work accidents”, collaborators, and people killed in intra-Palestinian fighting – has increased strongly over time. So far, some 189 Palestinians have died in this way; they represent more than twelve percent of the total Palestinians killed (see Graph 1.9). (Note that this figure does not include Palestinian “suicide shooters” – gunmen who carry out attacks on Israelis with the expectation that they will be killed in the course of their attacks.)

• The second phase of the al-Aqsa conflict lasted for nine months; this phase was characterized by a rough parity in each side’s killing of the other side’s noncombatants. The third phase (which is apparently still in progress) appears to be characterized by a higher (and more erratic) level of violence than the second phase; even so, noncombatant fatalities in the third phase are much more balanced between Israelis and Palestinians than they were in the first phase.

• Palestinian fatalities, including those of noncombatants, have shown extremely regular age distributions. In combination with the fact that almost all Palestinians killed in this conflict have been male – and absent any other reasonable explanation for such a non-random pattern of fatalities – this suggests that large numbers of Palestinian men and teenaged boys made a choice to confront Israeli forces, even after many of their compatriots had been killed in such confrontations. The pattern of Palestinian deaths is completely inconsistent with accusations that Israel has committed random attacks on residential areas, mixed-sex crowds at roadblocks, or other ordinary civilian gatherings.

• Israeli fatalities do not show such a high regularity in their age or gender distribution. Population segments like women or older people would not normally be specifically targeted for their military importance; thus their higher relative prevalence among Israeli fatalities is an indication of the degree to which these deaths were caused by random attacks on a civilian population.

Data Gathering

The greatest care and effort in carrying out this project has been spent in gathering and evaluating enough information on each fatal incident to enable accurate classification of each claimed fatality. Reliable and detailed data on Israeli casualties of the al-Aqsa conflict has been relatively easy to find, as this information is extensively reported in Israeli and foreign newspapers, as well as various official and unofficial websites. Palestinian Arab fatalities present much greater difficulty, for several reasons:

• Arab names are often long and complex; in many cases different sources give different casualty names for the same incident, and it is difficult to ascertain whether the different names in fact refer to the same person.

• Detailed reports of Palestinian casualties are generally provided by Palestinian organizations and individuals; in some cases these reports are “slanted” or even fraudulent, due to pressures to paint Israel in as negative a light as possible.

• The Israel Defense Forces do not keep a precise record (or at least have not so far provided such a record to us) of every time, place, and circumstance where weapons have been used; thus, certain Palestinian reports of fatalities due to Israeli fire cannot be confirmed or refuted. In many cases, we have given the Palestinian account of events the benefit of the doubt, even though the casualties may have resulted from Palestinian actions, rather than Israeli actions. In cases where the cause of death as reported by Palestinian sources is very much open to question, we have assigned a Low confidence Level to that particular casualty.

We have made extensive use of mainstream media outlets, both in Israel and abroad, for the details of al-Aqsa conflict incidents. Information on Palestinian casualties has been gathered from Arabic-language newspapers, cross-correlated with reports from human-rights organizations in Israel and the Palestinian Authority. Unfortunately, these sources generally disagree on many significant details, including the name, age, and circumstances of death of victims. It should be noted that, since no Israeli official body has been keeping records of Israeli actions and their results, the information reported by the Western media has come almost exclusively from Palestinian sources.

Categorization

In order to provide a better understanding of the nature and significance of al-Aqsa conflict incidents and casualties, we have classified them according to several criteria. This classification system is a reflection of our desire to improve upon the usual reportage, which – to give an extreme example – treats the death of a suicide bomber as equivalent to the death of one of his victims. A good system of classification, combined with carefully-gathered data, provides rich opportunities for analysis.

As the conclusions reached in this study are strongly influenced by the categories we have chosen for our analysis, it is important to understand these categories in some detail.

Incident Organization		This category is used to classify incidents, such as terror attacks, which can be attributed to a particular terrorist organization; it is used as well to identify the groups targeted by Israeli “targeted killings”.
Incident Type		This category provides a broad description of the type of incident that occurred; examples include “Terror Attack”, “Roadblock Confrontation”, and so on.
Attack Type		This is a more specific category, identifying the particular form of violence which led to injury or loss of life. Examples include “Car Bomb”, “Suicide Bomb”, and “Lynching”.
Incident Target		The Incident Target represents the immediate goal of an attack, as opposed to any broader political aims. Examples include “Civilian”, “Vehicle”, “Hotel”, and “Militant”.
Incident Confidence Level		This category reflects our considered opinion as to the reliability of a report. Confidence Level as applied to incidents represents the degree of certitude that the incident itself took place; applied to individual casualties, it represents the level of certainty that a particular individual was injured or killed in the incident. It is possible that an incident itself may have a high Confidence Level while some of the casualties attributed to the incident have a lower Confidence Level. This is the case where the incident itself was well attested, but some of the reports of casualties attributed to the incident were less reliable. Confidence levels are as follows: Extremely Low Low Questionable High Extremely High
Side Responsible for Incident		This category assigns responsibility for initiating an incident – for example, “Israel,” “Palestinians,” “Probably Israel,” “Probably Palestinians,” or “Unclear.” In the case of a terrorist attack on Israeli civilians by Palestinians, the Side Responsible is listed as “Palestinians”. The same is true of “work accidents.” In the case of a targeted killing of a Palestinian militant carried out by Israel, the Side Responsible is listed as “Israel.” In the majority of incidents—Roadblock Confrontations, Violent Clashes, etc—the Side Responsible is listed as “Unclear.” Note that Side Responsible in this study refers to physical responsibility only, and does not indicate a moral judgement.
Side Responsible for Casualty		In the majority of incidents, assigning responsibility for an incident to one side or another is insufficient for purposes of analysis. An incident may be initiated by Palestinians—for instance, an armed attack by Hamas militants on an Israeli bus—but end with the militants being killed by Israelis. In such a case, the Side Responsible for Incident would be “Palestinians”, while the Side Responsible for Casualty would be “Israel” in the case of each militant, and “Palestinians” in the case of each Israeli casualty. There are also cases in which Palestinians killed Palestinians and Israelis killed Israelis. As with Side Responsible for Incident, Side Responsible for Casualty represents only the physical responsibility for causing death or injury, and expresses no judgement as to the appropriateness of that action.
Combatant Level		This is one of our most significant ways of classifying casualties, and represents the degree to which someone killed or injured during the course of the al-Aqsa conflict can be considered an “innocent victim”. Our decisions in defining the Combatant Level categories and assigning casualties to these categories are made with reference to ICT’s published definition of terrorism, explained in Boaz Ganor’s article, “Terrorism: No Prohibition Without Definition”. Non-Combatant A non-combatant is an innocent bystander – a person whose death or injury has no justification other than nationality or ethnicity. Health Related A “health related” fatality is someone who died from a cause only indirectly related to violence – for example, due to a heart attack following an incident, tear-gas inhalation, or a roadblock delay that prevented an ilperson from receiving medical treatment in a timely manner. Probable Combatant A “probable combatant” is someone killed at a location and at a time during which an armed confrontation was going on, who appears most likely – but not certain – to have been an active participant in the fighting. For example, in many cases where an incident has resulted in a large number of Palestinian casualties, the only information available is that an individual was killed when Israeli soldiers returned fire in response to shots fired from a particular location. While it is possible that the person killed had not been active in the fighting and just happened to be in the vicinity of people who were, it is reasonable to assume that the number of such coincidental deaths is not particularly high. Where the accounts of an incident appear to support such a coincidence, the individual casualty has been given the benefit of the doubt, and assigned a non-combatant status. The status of “probable combatant” has also been assigned to people who knowingly took some action which would lead to increased danger, such as entering an area in which fighting was going on or which security officials had declared off-limits. Uniformed Non-Combatant A “uniformed non-combatant” is a non-civilian, but is not actively involved in the conflict. This category can include civil police as well as soldiers in uniform but not at their post. Violent Protester A “violent protester” may be a civilian, but has chosen to take an active and violent part in the conflict – such as rioting or vigilante activity. Protestor Unkown A “protestor unknown” is anyone who was killed during a protest, for whom information as to violent behavior is unavailable. Full Combatant A “full combatant” is a soldier on active duty, an active member of a terrorist group, or a civilian independently choosing to perpetrate an armed attack on the opposing side. In general, rock-throwers are not considered to be combatants; an exception to this generalization would be, for example, someone dropping large rocks from a bridge onto fast-moving traffic. A rioter throwing "Molotov cocktails", grenades, or the like can be considered a full combatant. Mere possession of a weapon does not imply combatant status. A civilian driving with a weapon in his/her car, or a pedestrian with a holstered pistol, is normally considered a noncombatant. However, a civilian who encounters a terror attack in progress and draws his/her weapon in an attempt to stop or prevent the attack is a combatant once the weapon is out of its holster. Suspected Collaborator This is a special category for Palestinian or Israeli Arabs targeted by militants who suspect them of aiding Israel. Unknown In a large number of cases, the information at hand has been insufficient to decide the circumstances of death for a given Palestinian casualty. As the project continues and additional information is accumulated, we believe that the number of “unknowns” will steadily decrease. When we divide fatalities into the broader categories of “Combatant” and “Noncombatant”, only “Full Combatants”, “Probable Combatants”, and “Violent Protestors” are counted as combatants. All others, including “Unknowns”, are considered to be noncombatants.
Casualty Confidence Level		This category reflects our considered opinion as to the reliability of a report of an individual death. A Confidence Level is applied to individual casualties as well as to incidents because we found that in many cases an incident itself was well attested, but some of the reports of casualties attributed to the incident were less reliable. Confidence levels are as follows: Extremely Low Low Questionable High Extremely High All the casualty statistics mentioned, along with the associated charts and graphs in this report, are based upon casualty reports with a confidence level of High or Extremely High. Unless specically indicated, Questionable, Low-confidence, and Extremely Low-confidence casualty reports have been filtered out.
Age		Where known, the age of the casualty is entered into the database. There are a number of casualties, particularly on the Palestinian side, where the age is unknown, or uncertain.
Gender		The gender of casualties has turned out to be central to our analysis, as discussed below.
Nationality and Secondary Nationality		The casualty’s nationality is obviously at the heart of any comparison of casualties between the two sides. In some cases, the casualty had dual citizenship. For this reason, we’ve included a field for Secondary Nationality.
Residence		This category was added in order to serve as a basis for a geographical analysis of both casualties and incidents.
Casualty Type		Casualty Type describes whether the casualty was killed or injured, and if the latter, the extent of the injury. The Casualty Types are: Injured Lightly Injured Moderately Injured Seriously Injured Unclear Killed For this first phase of the project only fatalities have been entered into the database.
Organizational Affiliation		This category describes the casualty’s membership in an existing non-governmental organization. This is particularly useful in the case of Palestinian combatant fatalities, who were largely active members of known terrorist groups, rather than individuals acting on their own.

A note on “Combatants” and “Civilians”

Media reports frequently discuss the fatalities of the al-Aqsa conflict in terms of the number of “civilian fatalities” on each side. We have deliberately avoided this usage. In any conflict between a country with conventionally-organized military and police forces and an opposing force mostly composed of non-uniformed “irregulars”, the uniformed forces cannot avoid killing a disproportionate number of “civilians” – since even their most deadly opponents are usually not members of an official military, and in many cases have perfectly respectable “day jobs”.

In the current Israeli-Palestinian conflict, the distinction between Palestinian “civilians” and members of the plethora of official Palestinian Authority security forces is even harder to make, since many Palestinian policemen (and members of the other P.A. uniformed forces) combine their official service with membership in one or more unofficial groups such as Hamas or the various arms of Fatah. When Palestinians in this situation have killed Israeli noncombatants, they have generally done so in their “civilian” capacity.

At first glance, it should be easier to determine which Israeli fatalities are “civilians”. However, even here the distinction between “civilians” and members of official security forces paints a somewhat distorted picture. A substantial number of Israeli fatalities, especially those killed inside “Israel proper”, have been members of the civil police, or noncombatant members of the Israel Defense Forces – such as office workers and mechanics.

As a result of all these factors, dividing this conflict's fatalities into “civilians” and “non-civilians” over-emphasizes the “civilian” status of many of the Palestinian victims, and to a degree distorts the significance of Israeli fatalities as well. At best, such categorization paints an inaccurate picture of the conflict; and in some instances, those who use these categories are clearly being disingenuous. (As an extreme example, one report in a Saudi newspaper contrasted some 1,400 Palestinian “civilians” killed with about 530 Israeli “soldiers and settlers”.)

For this reason, we chose to classify those killed by their actual combatant status, according to the criteria laid out in the “Combatant Level” section above. While this method requires a degree of judgement in categorizing those killed, it offers some hope of making sense of an assymmetrical conflict; whereas the alternative system, while easier to apply, cannot provide meaningful results.

Results and Trends

1. General Trends in Overall Fatalities

The first impression conveyed by the standard “Intifada body count” report is that people on both sides of the conflict have been getting killed at a more or less steady pace, with Palestinian fatalities outnumbering Israeli fatalities by a factor of almost three to one. A glance at Graphs 2.1 and 2.2 quickly dispels this illusion.

These graphs display all Israeli and Palestinian fatalities month-by-month, with no categorization or qualification. They show that Israeli casualties have varied widely from one month to the next, but have shown a general upward trend. (This trend is somewhat masked by the especially high death toll of March 2002, which “flattens” the rest of the graph.) Palestinian fatalities, on the other hand, were very high for the first few months of the conflict, then remained at a lower level – although still generally above the level of Israeli casualties. They increased again starting in September 2001 – possibly as a result of new, more aggressive Israeli counter-terrorism tactics adopted after the September 11 terror attacks on the United States. Large numbers of Palestinians were killed in March and April 2002, in the course of Israel’s “Operation Protective Shield” incursions into Palestinian-Authority-ruled cities; this operation was a response to a rash of major Palestinian terror attacks.

In order to show these trends more clearly, it is helpful to correct (or “normalize”) for the fact that the al-Aqsa conflict began on 27 September 2000^* – so that the first “month” shown on our graphs is in fact only four days long. By multiplying by 30/4, we can correctly display a value corresponding to the rate at which people were being killed during these first four days. Graphs 2.3 and 2.4 show this correction, as well as trend-lines to clarify the changes in the rates of death.

* Note that we consider the beginning of the “al-Aqsa Intifada” to have been on 27 September 2000, the date of the first Palestinian attack on Israelis carried out by official Palestinian Authority personnel.

While the monthly number of Israeli fatalities is rather chaotic, there has clearly been an upward trend in the fatalities during both “good” and “bad” months (that is, higher peaks and higher troughs in the graph), especially since December 2000 – January 2001. Palestinian fatalities, on the other hand, trended downward from a very high beginning, then picked up somewhat from September 2001 onwards – although they did not approach the levels of the first months of the conflict until March 2002.

Graph 2.5 compares these overall trends in fatalities suffered by the two sides. While it is evident that the overall level of Israeli fatalities has been consistently lower than that of Palestinian fatalities, the gap between the trend-lines lessened over the first few months of the al-Aqsa conflict and then remained roughly steady until September 2001. Since then, the gap between fatalities incurred by the two sides has been fluctuating erratically.

2. Refining the Trends – Responsibility and Combatant Status

So far we have looked only at overall fatality figures, without regard to any difference between one fatality and another. There are several ways in which we can refine our view. One obvious approach is to classify fatalities by which side caused them, rather than by the nationality of the deceased. By counting the people killed by the actions of each side rather than simply those who died on each side, we now classify suicide bombers and people killed while preparing explosives (“work accidents”) as part of their own side’s tally of deaths, rather than as apparent victims of the other side. Graph 2.6 shows the trends in deaths caused by each side, ignoring cases in which responsibility for death was unclear.

While the general appearance of this graph is similar to what we have seen before, it is significant that when we make these adjustments, the figures for fatalities caused by each side were actually quite similar, on average, for all but two months between January 2001 and January 2002. (Note that, as mentioned above, selecting fatalities by the side “responsible” for them does not imply that the responsible side is at fault; “responsible” in this context refers only to physical rather than moral responsibility.)

Our view can be refined still further (in the sense of selecting those deaths which are “politically significant”) by ignoring the deaths of combatants. This eliminates from consideration those fatalities who can be considered to have been actively involved in the fighting, and thus legitimate targets for attack: soldiers at their posts, active members of terrorist groups, suicide bombers, and so on. Those left – even though many of them, such as stone-throwing protestors, may have knowingly put themselves in harm’s way – are considered “noncombatants”. (Note that for our purposes, “Violent Protestor”, “Probable Combatant”, and “Full Combatant” are treated as combatants; all other classifications, including “Unknowns”, are considered noncombatants.)

Noncombatant status is significant in several ways: First, attacks by non-governmental groups on noncombatants are categorized as terror attacks, as opposed to attacks on combatants which are categorized as guerilla attacks – a more legitimate form of “resistance”. Second, large numbers of noncombatant fatalities inflicted by Israeli forces are a possible sign of problems in field tactics, as noncombatants are not normally considered to present an immediate danger to soldiers on duty. Third, noncombatant deaths are normally considered more deserving of sympathy, and thus have a special significance in a conflict being fought as much in the news media as on the ground. In short, combatants on both sides are generally accepted as legitimate military targets, of less political significance; while there is a moral onus attached to the killing of noncombatants.

Graph 2.7 shows the trends in noncombatant deaths inflicted by each side on "non-nationals" – i.e. citizens of the opposing side and third-country citizens. This graph shows the same general trends as Graph 2.6, although (as Palestinian combatant fatalities are no longer included) the vertical scale is somewhat lower. Also, this graph shows even more clearly the trend for a rough parity between the two sides after December 2000.

3. Identifying Phases of the Conflict

Graph 2.7 suggests that the al-Aqsa conflict, up to the end of June 2002, can be divided into three phases:

• An initial phase characterized by very high Palestinian fatalities;

• A middle phase of near-parity between the two sides (when we adjust for either combatant status or responsibility);

• A third phase in which fatalities on both sides fluctuated rather wildly.

A closer look at our casualty data enabled a more precise selection of starting and ending points for each phase – although any such exact dates are inevitably somewhat arbitrary.

The first phase of the al-Aqsa conflict began on 27 September 2000, and ended before the end of December 2000. At that time Palestinian fatalities tapered off sharply, and remained generally lower until the next September. December 21, 2000 has been chosen as the last day of this first phase. As a first approximation, we can label this phase of the conflict the “real or apparent popular uprising” phase (leaving room for uncertainty as to extent to which this “uprising” was manufactured by Palestinian leaders), as most of the fatalities would appear to have occurred as the result of Palestinian mass demonstrations or riots, and the Israeli response to them. (A more detailed breakdown of these fatalities by Incident Type remains to be done.)

The second phase began on 22 December 2000, and lasted until September 2001. It was characterized by rough parity between the two sides when fatalities are measured either by responsibility or by noncombatant status. It also featured a general rising trend in Israeli fatalities, as well as in deaths to both sides caused by Palestinians. A final date of 11 September 2001 has been chosen for this period, because changes in Israeli policy and tactics resulting from that day’s terror attacks on the United States appear to have ushered in the next phase.

The third phase of the al-Aqsa conflict began on 12 September 2001 – again, a date chosen somewhat arbitrarily, as the first day “post 9/11”. This phase is characterized by an initial significant increase in Palestinian fatalities, in contrast to the period of rough parity between the two sides in noncombatant fatalities suffered and fatalities caused. Over the course of this phase, fatalities on both sides have fluctuated dramatically from month to month, with Palestinian and Israeli trends seemingly closely correlated.

More work remains to be done in order to match the trends of fatalities with corresponding developments on the political scene. However, one can already point to several significant points: for example, a dip in Israeli fatalities in July 2001 (as a result of international pressure on the Palestinians after the June 2001 Dolphinarium attack); and sharp dips in both Palestinian and Israeli casualties in January 2002. The latter decrease in casualties corresponds to Yasser Arafat’s “cease-fire calls” to his own side on 16 December 2001.

Further sections will explore the three phases of the al-Aqsa conflict in more detail.

4. Combatants and Noncombatants

As noted above, the classification of victims into combatants and noncombatants is important in evaluating both their tactical and political significance. Obviously it is worthwhile to examine this classification in more detail.

Graphs 2.8 shows the distribution of Palestinian fatalities among Combatant Level categories (including only high-confidence reports).

Note that a substantial portion – almost 55 per cent – of Palestinian fatalities are either Full Combatants, Probable Combatants, or Violent Protestors. Of the remaining fatalities, a substantial slice are classified as “Unknown”, meaning that their Combatant Level has not yet been determined. Ongoing review of existing reports, as well as addition data to be gathered, will, we hope, reduce the number of Unknowns.

Graph 2.9 shows the equivalent distribution for Israeli fatalities. The most obvious feature of this graph is the overwhelming preponderance of noncombatants over combatants. According to our practice of classifying only Full Combatants, Probable Combatants, and Violent Protestors as combatants, about one in five Israeli fatalities have been combatants. Even were we to include such categories as civil police (i.e. Uniformed Noncombatants) among combatants, some 70 per cent of Israelis killed have been noncombatants.

The apparent reason for the lopsided distribution of Israeli fatalities is that Israeli combatants are members of a well-trained and equipped modern army – and more specifically, one that goes to unusual lengths to minimize its casualties. This has two implications: first, that Palestinians will generally prefer to attack civilian targets, or alternatively members of the military who are not on active duty; and second, that most Palestinian attacks on Israeli military patrols or outposts are unlikely to cause extensive Israeli fatalities.

Graph 2.10 shows trends in the balance of Palestinian fatalities between combatants and noncombatants, month by month. Note that Phase 1 and the first month of Phase 3 are both characterized by surges in noncombatant deaths, while other periods show either parity between combatants and noncombatants, or else a preponderance of combatant deaths.

Graph 2.11 shows the trend in Palestinian fatalities caused by Palestinian actions – including suicide bombings, “work accidents”, internecine struggles, and so on.

The strong upward trend in Palestinian fatalities due to Palestinian actions suggests several possible explanations: increased suicide bombings, occasional clashes brought on by efforts by Palestinian Authority security forces to exert its authority over the various Islamist groups, and a general breakdown in law and order in Palestinian areas. The large “spike” in Palestinian noncombatants killed by Palestinians in April 2002 appears to represent the large number of “collaborators” killed in the aftermath to Israel’s “Defensive Shield” operation.

Graph 2.12 is the Israeli equivalent to Graph 2.10, showing Israeli combatant versus noncombatant fatalities.

The preponderance of noncombatant over combatant casualties is immediately obvious. The irregular but generally increasing trend in Israeli noncombatant fatalities is also apparent, along with the gradual increase in Israeli combatant fatalities since the beginning of 2002, leading up to the substantial losses suffered during Operation Protective Shield. (Note also December 2001, which included several major Palestinian terrorist attacks perpetrated during American envoy Anthony Zinni’s visit to the area.) As we saw above, the extremely high number of Israeli noncombatants killed in March 2002 "flattens" the rest of the graph, making the overall increasing trend appear less significant than it otherwise would.

Finally, Graph 2.13 analyzes the relationship between each month's Palestinian combatant and noncombatant fatalities over time. (Compare this to Graph 1.3, which shows the cumulative trend rather than each month's ratio individually.)

The pronounced increase over time in the percentage of combatants among Palestinian fatalities appears to result from a combination of several possible factors:

• more efficient Israeli tactics which cause fewer noncombatant fatalities;

• an increase in the number of Palestinian combatants killed due to the actions of their own side, or killed by Israeli actions;

• possible changes in the behavior of Palestinian protestors, resulting in fewer and/or less violent demonstrations.

At the same time, note that the major incursions of Israeli forces into Palestinian Authority-controlled areas in September 2001 and March-April 2002 resulted in significant “dips” in the monthly ratio of combatants killed. Graph 1.3 shows that these “dips” represented a pause in a long-term trend, rather than any reversal of that trend.

5. Gender War

The issue of gender has not been widely discussed in relation to the al-Aqsa conflict. Investigation of the sexual composition of fatalities on both sides of the al-Aqsa conflict reveals some striking facts.

Graph 2.14 shows Israeli fatalities month by month, with male victims separated from female victims. While there is an overall preponderance of male Israelis killed (slightly over two thirds of the victims were male), the pattern is not consistent – there were some months when females outnumbered males among the victims – or extreme. As will be seen below in the section dealing with Age Distribution, the excess of males is not consistent across ages; fatalities among young people and the elderly are balanced between the sexes.

Graph 2.15, displaying the equivalent data for Palestinians, shows a dramatically different picture. Palestinian fatalities, it turns out, have been consistently and overwhelmingly male. In total, Palestinian women account for fewer than five percent of all Palestinian fatalities.

To eliminate any possible distortion caused by including combatants – almost all male on both sides – in our picture, we can generate similar graphs illustrating the preponderance of males and females among the noncombatants on each side who were killed by the other side. Graphs 2.16 and 2.17 show that even when we restrict ourselves to the noncombatant victims of the conflict, the Palestinians killed have been almost entirely male.

Graph 2.18 reveals yet another fact that has received scant media attention: If we look only at females killed, Israeli fatalities have far outnumbered Palestinian fatalities. This graph includes only each side's noncombatant females killed by the opposite side; but even if we include Palestinian combatant females and others whose death is not reliably attributable to Israeli actions, the ratio of Israeli females killed to Palestinian females killed is 2.8 to 1. (Using the more stringent criteria of noncombatants killed by the opposite side, the ratio is 3.9 to one.)

6. Age Distributions

Yet another area that has received inadequate attention is the age distribution of victims of the al-Aqsa conflict. Again, analysis of the data yields some surprises.

Graph 2.19 shows the age distribution of Palestinian female noncombatant victims. Since the number of such victims is comparatively small, cases where responsibility for death is unclear have been included. Note that this graph displays no strong trend – the distribution of deaths across age categories appears essentially random.

This graph is included mainly to provide a contrast to those that follow.

Graph 2.20, on the other hand, showing high-confidence Palestinian noncombatants of both sexes (which, as we’ve seen before, consist almost entirely of males), displays a highly non-random distribution.

The median age for Palestinian victims is roughly 23; some 37% of them were under twenty years old, and 72% under thirty. The most striking features of the graph, however, are the dramatically high number of teenage noncombatants killed and the almost perfect “textbook” skewed distribution.

Graph 2.21 shows the age distribution of Palestinian high-confidence combatant fatalities. Again, this graph shows a highly non-random distribution – but with significant differences from Graph 2.20. Palestinian combatant deaths are even more concentrated in a small age-range than Palestinian noncombatant deaths are, and combatants on average were somewhat older than noncombatants.

Teenagers are much less prevalent among Palestinian combatants; only around 18% of all Palestinian combatants killed were under twenty years old. (This reflects, to a minor degree, the fact that as a matter of principle we have classified essentially all Palestinians under the age of 13 as non-combatants.) The median age for Palestinian combatants is about 24 years. On the other hand, around 80% of the Palestinian combatants killed were under thirty – a slightly higher percentage of under-age-thirty deaths than among Palestinian noncombatants.

The distribution of Palestinian fatalities across age and gender demonstrates a simple but important fact: Palestinians killed in the al-Aqsa conflict have been overwhelmingly male, and almost as overwhelmingly teenaged or in their twenties. (Note, though, that the number of children killed under the age of ten is very low – under five percent of noncombatants.) This is highly significant, as it is very different from the results one would expect from random Israeli fire into inhabited neighborhoods, or other forms of indiscriminate killing of which Israel has been accused. These graphs suggest that there was a strong element of selection of those who would eventually be killed – basically, teenagers and young men decided, or were encouraged, to confront Israeli forces and, in all too many cases, “achieve martyrdom”. In this context, it is unsurprising that the element of selection – showing up as a more “focused” distribution – is even stronger for Palestinian combatants.

Graph 2.22 and 2.23 provide separate age breakdowns for males and females, for Israeli noncombatants killed by Palestinians and Palestinian noncombatants killed by Israel. The Israeli figures for both males and females display patterns consistent with a population subject to terror attack – an essentially random distribution, with a slight prevalence among ages when people are more likely to be “out and about”. (It is interesting to note that female deaths are roughly equal to male deaths for ages below 20 and above 59; the lower number of female deaths between the ages of 20 and 59 presumably reflects more time spent indoors.)

The Palestinian pattern is very different. Palestinian female noncombatants (as more easily seen in Graph 2.19, above) show a fairly random distribution; but Palestinian male noncombatants display an age distribution completely unlike any of the other noncombatant groups. It is obvious from these graphs that the Israeli killing of Palestinian noncombatant males is a very different phenomenon from the killing of other noncombatants in this conflict.

Graph 2.24 shows another interesting aspect of the age distribution of Palestinian fatalities. Here noncombatant Palestinians killed by Israel have been divided up according to the al-Aqsa conflict phases described above.

In order to make accurate comparisons among phases of different lengths, the data has been “normalized” into deaths per 30-day month. When the data is graphed in this way, several things become apparent:

• Each phase of the conflict shows a distinct difference in the overall rate at which noncombatants were killed. This is most obvious in the case of Phase 1, during which many noncombatants were killed over a relatively short time-span. Phase 2 shows a dramatic reduction in the rate at which noncombatants were killed – a reduction of over 75 percent. Phase 3 displays a rate of noncombatant death at least twice that of Phase 2, but still much less than that of Phase 1. (Note that the graph somewhat understates Phase 3 deaths, as we don't yet know the ages of some 66 of the Palestinian noncombatants killed by Israel during this phase – some 20% of the total.)

• While Phase 2 saw a major reduction in the rate of noncombatant death, the age distribution of noncombatant fatalities was somewhat similar to that during Phase 1. This implies that a substantial proportion of the noncombatants killed in Phase 2 were probably killed in similar circumstances as those in Phase 1 – that is, with the same selection process for potential victims.

• Phase 3 displays a significantly different pattern. While Phase 1 noncombatant deaths show a strong peak in the teenage years (with a median age at death of around 18-19), Phase 3 noncombatants killed were, on average, significantly older. The greatest number of deaths occurred between the ages of 25 and 29, and the median age of death was between 24 and 25. On the other hand, Phase 3 is the only phase in which fatalities among 10-14-year-olds exceeded those among older teenagers. If this trend continues, it will obviously be cause for concern.

The above observations imply that Phase 2 differed from Phase 1 (at least in terms of the death of Palestinian noncombatants) primarily in intensity. Phase 3, on the other hand, saw a significant difference in the ages at which noncombatants were killed – implying that these deaths were due to tactical changes on one or both sides of the conflict. One significant factor in this change may be Israel’s series of incursions into Palestinian Authority territory during Phase 3.

Graphs 2.25, 2.26, and 2.27 show total Israeli noncombatant fatalities for each phase of the al-Aqsa conflict. (Three separate graphs are shown here for readability; combining them as was done to create Graph 2.22 resulted in a graph that was colorful but difficult to interpret.)

Four facts are immediately apparent:

• Phase 1 shows almost all casualties (the sole exception is one person aged between 60 and 64) between 15 and 44 years old. The fact that almost all Phase 1 Israeli deaths took place within these ages – a very different pattern than is seen for later phases of the conflict – suggests that the incidents resulting in these fatalities were different from those that came later. (This should be clarified by future analysis by Incident Type.)

• Phases 2 and 3 show that Israeli teenagers were killed more than any other age bracket. This high number of teenaged fatalities may be the result of suicide bombings, many of which targeted groups of young people. Again, it is hoped that analysis by Incident Type will clarify this issue.

• Phases 2 and 3 show a considerable number of fatalities among people aged 40 and over. The general pattern appears to be much more random than what was seen for Palestinian noncombatant fatalities. This suggests that the element of selection of victims by their own side which appears to account for the age distribution of Palestinian fatalities is not present for Israeli noncombatant fatalities. Once more, this is an area for further investigation.

• The rate of Israeli fatalities has increased each phase, and the age distribution of these fatalities has become broader as well.

Graph 2.28 displays all Israeli fatalities by age, with combatants separated from noncombatants.

In addition to showing once more that the vast majority of Israelis killed have been noncombatants, this graph contrasts the very orderly distribution of Israeli combatant deaths – expected for a uniformed army with reserve service generally continuing until the mid-40’s – with a much broader distribution of noncombatant fatalities. (Note that the small number of Israeli combatants appearing in the “15-19 Years” category represents young conscripts, normally recruited at the age of 18. In fact, no Israeli 18-year-old combatants have been killed in the conflict so far; the youngest killed have been 19 years old.)

Graph 2.29, the equivalent graph for Palestinian fatalities, shows a completely different picture. Palestinian combatant fatalities, like those on the Israeli side, are concentrated in a narrow age range – although this concentration is slightly less pronounced. (This is unsurprising, given that Palestinian combatants are mostly members of unofficial terrorist/guerilla organizations.) Palestinian noncombatant fatalities, however, show an age distribution completely unlike that on the Israeli side. Instead of a “sloppy” distribution over a broad range of ages, Palestinian noncombatant fatalities are heavily concentrated among teenagers and young adults.

Graph 2.30 focuses specifically on noncombatants on both sides in their forties and older.

While overall Palestinian deaths outnumber Israeli deaths by almost three to one, Israeli “mature noncombatant” deaths are almost double the equivalent Palestinian fatalities. (If we omit Palestinian noncombatants, such as “collaborators”, killed by Palestinians, the ratio is slightly over 2:1.)

Graph 2.31 shows a detailed breakdown of Israeli and Palestinian young noncombatants killed by the opposite side, by age and gender. These fatalities display a rather strange pattern. Among both Palestinians and Israelis, the number of young children (under the age of 8-9 years old) is comparatively small (although more young Israeli children were killed as a proportion of total fatalities). The number of Palestinian children killed begins to increase at about 10 years of age, and jumps up dramatically between the ages of 12 and 14. However, the increase consists entirely of boys – the number of Palestinian girls killed shows no age-trend, and is very low for all ages.

Young Israelis killed by Palestinians show a different profile: Both boys and girls show an increase starting at age 14 (perhaps a year earlier for boys), and just as many teenaged girls were killed as teenaged boys.

What is significant in all these comparisons is, again, the contrast between the randomness of the pattern of Israeli fatalities and the more non-random distribution of Palestinian deaths. The random distribution is typical of terrorist attacks, which, though sometimes carried out in places frequented by young people, e.g. the Dolphinarium disco attack, may equally target restaurants or buses which are used by a wide spectrum of the population. Some of the most frequent targets of Palestinian terror attacks, such as open-air markets and public buses, are used disproportionately by the most vulnerable segments of society: women, the elderly, and the poor.

The fact that Palestinian deaths caused by Israeli actions do not, as a rule, follow the same pattern would seem to undermine claims that Israel deliberately targets Palestinian civilians.

---

Home | Spotlight | International Terrorism | Counter-Terrorism | Arab-Israeli Conflict | Search

[sanchmo]

One last statistic:

Who started it:
Israel: 0%
Palestinians: 100%

[Black Agnes]

bttt, Interesting statistical analysis.

[Uncle Miltie]

Fantastic Analysis! I have always maintained that the fatality statistics quoted by AP and Reuters were flawed due to Palestinian deaths composed of suicide bombers, armed terrorists KIA, and collaborator murders.

Now we have it: 69% of Israelis killed have been Non-Combatants, 13.6% of Palestinians killed have been Non-Combatants. This documents the moral distinction between evil Palestinian terrorists targeting unarmed innocent women and children, and good uniformed Israeli soldiers trying to avoid killing innocents.

Will any mainstream news organization pick this up and explain it to people?

[headsonpikes]

A tour de force!!!

[MosesKnows]

Who started it:
Israel: 0%
Palestinians: 100%

In order for your conclusion to be valid, you would need to define the time over which you accumulated the data. Without a time reference, your conclusions have no merit.

Just for kicks do the data reduction again and this time start with one of the milestones dates in Middle East history. Perhaps the Balfour Declaration in November of 1917 or maybe, if you are real gutsy, begin accumulating your data early in the morning of April 9, 1948.

Tally your results and post them here. I will look forward to seeing them.

[staytrue]

I think the time frame is correct for the present. You could go back to the 0 AD for example. But the proper thing to do now is to make the best of the current situation. I think the Israelis are doing a better job of that right now than the Palis are. I'm glad the statistics were so thorough. However the Palis can argue that the only effective weapons that they have are the suicide bombers while the Israelis have tanks, jets, etc. which are more discriminating weapons than the plain jane bomb. Bottom line, though, I still think the jews are doing better than the palis. Of course, neither side is completely in the right.

[Saundra Duffy]

I'm impressed. Victory for Israel!

[PeoplesRepublicOfWashington]

Good stuff.

[Pearls Before Swine]

Now we have it: 69% of Israelis killed have been Non-Combatants, 13.6% of Palestinians killed have been Non-Combatants. This documents the moral distinction...

Will any mainstream news organization pick this up and explain it to people?

No.

[Senator Pardek]

Superlative post. I hope that next someone tackles the "U.S. starves 3 trillion Iraqi babies a week" nonsense.

[monkeyshine]

BTTT

The Israeli Government doesn't do PR very well any more. I don't know why that is so. They should be out in force meeting with editors, journalists, politicians.

Not only these statistics are important. There is a lot Israel does that most people don't know about. For example, during the battle of Jenin, IDF troops were actively working to fix broken water mains, they brought a generator to the Palestinian hospital so that the sick and wounded could be treated, they worked to repair damaged power lines, etc ... all while the battle was raging. Yet all we heard was "massacre".

Israel needs to get more information out, and at a faster pace.

[ppaul]