Peace Research and Conflict Management
1.Introduction: Reading
Organised violence 1989-2024, and the challenges of identifying civilian victims, Davies et al. (2025)
Fatalities from organized violence remain extremely high, with 2024 among the most violent years since the
end of the Cold War, even though deaths fell slightly compared with 2023.
Big picture: levels and types of violence
• Around 160,000 people were killed in organized violence in 2024, only slightly fewer than in 2023,
and far above levels in 2010.
• UCDP distinguishes three mutually exclusive types of organized violence: state-based conflict,
non-state conflict, and one-sided violence, and all have increased substantially since around 2011.
State-based conflict: many conflicts, concentrated deaths
• The number of state-based armed conflicts reached a record 61 in 2024, and the number of wars
(≥1,000 battle-related deaths per year) rose from 9 to 11, the highest since 2016.
• Fatalities in state-based conflicts stayed very high at about 128,400, with only the three previous
years being deadlier since 1989.
• Russia–Ukraine remained by far the deadliest conflict, causing almost 75,700 deaths (about 59% of
all state-based fatalities), followed by Israel–Hamas and related regional fighting.
• Regionally, Africa saw a notable 17% drop in state-based fatalities (still historically high), while Asia,
the Americas, and the Middle East saw modest increases tied to specific conflicts (e.g. Taliban–
Pakistan, Arakan in Myanmar, Haiti).
• Interstate conflicts doubled from 2 to 4 in 2024, the highest number since 1987, continuing a trend
of more frequent interstate clashes and internationalized civil wars.
Non-state conflict: slightly declining but still intense
• Non-state conflicts (organized armed groups fighting each other without direct state involvement)
fell from 80 to 74 in 2024, the second consecutive yearly decline.
• Fatalities from non-state conflict decreased to about 17,500, the lowest since 2016, mainly due to
reductions in Africa and the Americas, though Asia and the Middle East saw upticks.
• The Americas remain the epicentre of non-state conflict, especially Mexico and Brazil where cartel
and gang warfare produces very high death tolls and several non-state conflicts reach “war-level”
intensity (>1,000 deaths).
• In 2024, violence in Mexico involved fragmented and competing cartels (especially CJNG and
Sinaloa factions), with several conflicts each causing hundreds or thousands of deaths.
One-sided violence: sharp rise in killings of civilians
• One-sided violence (deliberate killings of civilians by states or organized groups) increased markedly
in 2024: 49 actors killed over 13,900 civilians, the highest one-sided death toll since 2014.
• Non-state actors accounted for about 74% of one-sided fatalities, with Africa contributing 82% of
global one-sided deaths, especially in DRC, Sudan, Burkina Faso, and Haiti.
• Islamic State (various branches) was the single deadliest one-sided actor (c. 3,800 civilian killings),
particularly in DRC and via IS-Khorasan’s international attacks (Iran, Russia, Europe, Turkey).
• State and quasi-state actors such as Sudan’s RSF, the Sudanese government, Burkina Faso’s security
forces, and Haitian gangs also committed extensive killings, often tied to ongoing wars and political
turmoil.
,Key regional patterns you should remember
• Europe: Violence is dominated by the Russia–Ukraine war, which generates huge combatant
fatalities and large numbers of “unknown” victims in intense urban warfare (e.g. Mariupol earlier in
the war)
• Middle East: Israel–Hamas and associated conflicts (Hezbollah in Lebanon, Iran–Israel strikes, Gaza)
produce very high battle deaths and many casualties whose status (civilian/combatant) is hard to
classify.
• Africa: High-intensity state-based conflicts (Ethiopia, DRC, Sudan, Burkina Faso, Somalia, Mali) and
one-sided killings (IS, JNIM, RSF, state forces) drive the bulk of global civilian victimization.
• Americas: Non-state and criminal violence dominates; cartels in Mexico and gangs in Brazil and Haiti
cause large numbers of deaths, but victims are often hard to classify as civilians or combatants.
Conceptual core: how UCDP classifies violence
• UCDP aggregates three categories: 1) state-based armed conflict, 2) non-state conflict, and 3)
one-sided violence, under the umbrella of “organized violence,” each requiring at least 25 fatalities
per year.
• “War” within state-based conflicts means ≥1,000 battle-related deaths in a year; only a subset of
conflicts reach this intensity, but they account for most fatalities.
• For each fatality, UCDP tries to classify the victim as combatant, civilian, or “unknown,” using
event-level information from multiple sources when possible.
• The “unknown” category is used when roles or affiliations cannot be determined, which is very
common in irregular warfare and organized crime settings.
Challenges of identifying civilian victims (exam-relevant)
• Identifying civilians is difficult when combatants lack uniforms, shift roles, or when fighting occurs in
densely populated urban areas with heavy use of artillery and airstrikes.
• Access restrictions, propaganda, contested casualty claims, and the destruction of media and state
institutions reduce data quality and create uncertainty about who was killed.
• The information environment has improved since the 1990s (internet, smartphones, OSINT,
specialized casualty-recording NGOs), leading to fewer “unknown” cases overall, but not
everywhere.
• Still, many major past conflicts (e.g. Gulf War, Chechnya, DRC wars, Iraq invasion) have large shares
of unknown victims because they rely on rough aggregate figures with little victim detail.
Regional contrasts in classification of victims
• In the Americas, especially Mexico and Brazil, the vast majority of deaths from organized violence
are classified as “unknown” because criminal groups blend with communities and violence is used
both against rivals and civilians.
• In Mexico in 2024, about 97% of roughly 11,000 recorded deaths in organized violence were
“unknown,” illustrating how hard it is to separate civilians from criminal combatants.
• In Europe, especially Ukraine, the share of unknowns is lower for many events because frontlines
stabilized and civilians largely left combat zones, making it easier to identify civilian victims from
attacks on civilian areas.
• However, large, intense urban campaigns (Mariupol; current Gaza campaign in the Middle East) can
still generate tens of thousands of “unknown” deaths due to rubble, missing persons, and lack of
independent reporting.
,Why this matters for research and policy
• Civilian fatality ratios are central for detecting ethnic targeting, violations of international
humanitarian law, and patterns of indiscriminate or strategic violence against civilians.
• When many fatalities are “unknown,” it becomes harder to test theories about civilian victimization
and ethnic targeting, especially using event-level quantitative data.
• UCDP’s one-sided data captures targeted killings of civilians but does not measure all civilian
casualties from indiscriminate bombardment, so it must be combined with other data and
assumptions to study overall civilian harm.
• Improving classification requires better access, independent monitoring, triangulation of sources,
and post-war efforts such as tribunals, truth commissions, and demographic reconstructions
1.Introduction: Notes
What is a Conflict?
Conflict has usually a negative connotation, but not always: in some cases that conflicts lead to something
better and progress. Political disagreement is when better communication, creative and innovative solutions
are found (think about European history). However, violent conflicts are normally negative.
Conflict is an important field of study as it touches different aspects of people’s life (e.g. psychology,
economics, politics, laws), it is multidisciplinary. Additionally, conflicts are everywhere, they are
omnipresent.
Definition is more than purely descriptive (Cramer) as it shapes the interpretation and how a concept is
viewed. Yet, there is no agreement on an exact definition of conflict. According to Sambanis, very few
variables are robust to different definitions→most robust variable to explain conflict: GDP pro capita, so the
risk of having a conflict is related to GDP per capita (Figure 2.1 ppt): the lower the GDP, the higher the
conflict risk.
Definition is important also to provide exact data, e.g. Uppsala Conflict Data Program (UCDP) dataset
provides data on countries in conflict. According to UCDP:
• Armed conflict: state-based armed conflict is a contested incompatibility that concerns
government and/or territory where the use of armed force between two parties, of which at least
one is the government of a state, results in at least 25 battle-related deaths in one calendar year.
• War: state-based conflict which reaches at least 1000 battle-related deaths in a specific calendar
year
• Non-state conflict: use of armed force between two organised armed groups, neither of which is
the government of a state, which results in at least 25 battle-related deaths in a year.
• Conflict, intrastate with foreign involvement/internationalised: armed conflict between a
government and a non-government party where the government side, the opposing side, or both
sides, receive troop support from other governments that actively participate in the conflict, i.e.
conflict between government and non-government party supported from a foreign government
• International or interstate conflict: conflict between two or more governments
• One-sided violence: the deliberate use of armed force by the government of a state or by a
formally organised group against civilians which results in at least 25 deaths in a year, e.g. genocide
, Conflict definition:
→ Casualty threshold is very arbitrary
→ Absolute casualty numbers lead to skewed conflict datasets
→ Is battle-related deaths an appropriate measure? Substantial damage on civilians: direct and
indirect (e.g. famine, bad healthcare), so should civilian deaths be considered too?
In this sense, 2 elements are important:
1. Difficulty of defining conflict involving damage on civilians
2. Complexity of finding any type of data and reports on conflict zones (there is a war going on, so
availability of information is limited)
Conflict characteristics
Incidence of conflict by region from 1960-90:
• Low-income countries: corruption, governance institutions and polity might be rotten, so it is not
only GDP pro capita but there might be other features to take into account (spurious relationship)
• Reversed causality: rich countries may be rich because they did not have conflicts→different
explanations.
Indeed, wealthier states are better able to protect assets, i.e. rebellion has less chance of success. In poor
states the opportunity costs for joining a rebellion are lower and therefore more feasible, so more
rebellions and conflict. Poverty causes violence→scarcity leads to migration and fight over scarce.
1.Introduction: Reading
Organised violence 1989-2024, and the challenges of identifying civilian victims, Davies et al. (2025)
Fatalities from organized violence remain extremely high, with 2024 among the most violent years since the
end of the Cold War, even though deaths fell slightly compared with 2023.
Big picture: levels and types of violence
• Around 160,000 people were killed in organized violence in 2024, only slightly fewer than in 2023,
and far above levels in 2010.
• UCDP distinguishes three mutually exclusive types of organized violence: state-based conflict,
non-state conflict, and one-sided violence, and all have increased substantially since around 2011.
State-based conflict: many conflicts, concentrated deaths
• The number of state-based armed conflicts reached a record 61 in 2024, and the number of wars
(≥1,000 battle-related deaths per year) rose from 9 to 11, the highest since 2016.
• Fatalities in state-based conflicts stayed very high at about 128,400, with only the three previous
years being deadlier since 1989.
• Russia–Ukraine remained by far the deadliest conflict, causing almost 75,700 deaths (about 59% of
all state-based fatalities), followed by Israel–Hamas and related regional fighting.
• Regionally, Africa saw a notable 17% drop in state-based fatalities (still historically high), while Asia,
the Americas, and the Middle East saw modest increases tied to specific conflicts (e.g. Taliban–
Pakistan, Arakan in Myanmar, Haiti).
• Interstate conflicts doubled from 2 to 4 in 2024, the highest number since 1987, continuing a trend
of more frequent interstate clashes and internationalized civil wars.
Non-state conflict: slightly declining but still intense
• Non-state conflicts (organized armed groups fighting each other without direct state involvement)
fell from 80 to 74 in 2024, the second consecutive yearly decline.
• Fatalities from non-state conflict decreased to about 17,500, the lowest since 2016, mainly due to
reductions in Africa and the Americas, though Asia and the Middle East saw upticks.
• The Americas remain the epicentre of non-state conflict, especially Mexico and Brazil where cartel
and gang warfare produces very high death tolls and several non-state conflicts reach “war-level”
intensity (>1,000 deaths).
• In 2024, violence in Mexico involved fragmented and competing cartels (especially CJNG and
Sinaloa factions), with several conflicts each causing hundreds or thousands of deaths.
One-sided violence: sharp rise in killings of civilians
• One-sided violence (deliberate killings of civilians by states or organized groups) increased markedly
in 2024: 49 actors killed over 13,900 civilians, the highest one-sided death toll since 2014.
• Non-state actors accounted for about 74% of one-sided fatalities, with Africa contributing 82% of
global one-sided deaths, especially in DRC, Sudan, Burkina Faso, and Haiti.
• Islamic State (various branches) was the single deadliest one-sided actor (c. 3,800 civilian killings),
particularly in DRC and via IS-Khorasan’s international attacks (Iran, Russia, Europe, Turkey).
• State and quasi-state actors such as Sudan’s RSF, the Sudanese government, Burkina Faso’s security
forces, and Haitian gangs also committed extensive killings, often tied to ongoing wars and political
turmoil.
,Key regional patterns you should remember
• Europe: Violence is dominated by the Russia–Ukraine war, which generates huge combatant
fatalities and large numbers of “unknown” victims in intense urban warfare (e.g. Mariupol earlier in
the war)
• Middle East: Israel–Hamas and associated conflicts (Hezbollah in Lebanon, Iran–Israel strikes, Gaza)
produce very high battle deaths and many casualties whose status (civilian/combatant) is hard to
classify.
• Africa: High-intensity state-based conflicts (Ethiopia, DRC, Sudan, Burkina Faso, Somalia, Mali) and
one-sided killings (IS, JNIM, RSF, state forces) drive the bulk of global civilian victimization.
• Americas: Non-state and criminal violence dominates; cartels in Mexico and gangs in Brazil and Haiti
cause large numbers of deaths, but victims are often hard to classify as civilians or combatants.
Conceptual core: how UCDP classifies violence
• UCDP aggregates three categories: 1) state-based armed conflict, 2) non-state conflict, and 3)
one-sided violence, under the umbrella of “organized violence,” each requiring at least 25 fatalities
per year.
• “War” within state-based conflicts means ≥1,000 battle-related deaths in a year; only a subset of
conflicts reach this intensity, but they account for most fatalities.
• For each fatality, UCDP tries to classify the victim as combatant, civilian, or “unknown,” using
event-level information from multiple sources when possible.
• The “unknown” category is used when roles or affiliations cannot be determined, which is very
common in irregular warfare and organized crime settings.
Challenges of identifying civilian victims (exam-relevant)
• Identifying civilians is difficult when combatants lack uniforms, shift roles, or when fighting occurs in
densely populated urban areas with heavy use of artillery and airstrikes.
• Access restrictions, propaganda, contested casualty claims, and the destruction of media and state
institutions reduce data quality and create uncertainty about who was killed.
• The information environment has improved since the 1990s (internet, smartphones, OSINT,
specialized casualty-recording NGOs), leading to fewer “unknown” cases overall, but not
everywhere.
• Still, many major past conflicts (e.g. Gulf War, Chechnya, DRC wars, Iraq invasion) have large shares
of unknown victims because they rely on rough aggregate figures with little victim detail.
Regional contrasts in classification of victims
• In the Americas, especially Mexico and Brazil, the vast majority of deaths from organized violence
are classified as “unknown” because criminal groups blend with communities and violence is used
both against rivals and civilians.
• In Mexico in 2024, about 97% of roughly 11,000 recorded deaths in organized violence were
“unknown,” illustrating how hard it is to separate civilians from criminal combatants.
• In Europe, especially Ukraine, the share of unknowns is lower for many events because frontlines
stabilized and civilians largely left combat zones, making it easier to identify civilian victims from
attacks on civilian areas.
• However, large, intense urban campaigns (Mariupol; current Gaza campaign in the Middle East) can
still generate tens of thousands of “unknown” deaths due to rubble, missing persons, and lack of
independent reporting.
,Why this matters for research and policy
• Civilian fatality ratios are central for detecting ethnic targeting, violations of international
humanitarian law, and patterns of indiscriminate or strategic violence against civilians.
• When many fatalities are “unknown,” it becomes harder to test theories about civilian victimization
and ethnic targeting, especially using event-level quantitative data.
• UCDP’s one-sided data captures targeted killings of civilians but does not measure all civilian
casualties from indiscriminate bombardment, so it must be combined with other data and
assumptions to study overall civilian harm.
• Improving classification requires better access, independent monitoring, triangulation of sources,
and post-war efforts such as tribunals, truth commissions, and demographic reconstructions
1.Introduction: Notes
What is a Conflict?
Conflict has usually a negative connotation, but not always: in some cases that conflicts lead to something
better and progress. Political disagreement is when better communication, creative and innovative solutions
are found (think about European history). However, violent conflicts are normally negative.
Conflict is an important field of study as it touches different aspects of people’s life (e.g. psychology,
economics, politics, laws), it is multidisciplinary. Additionally, conflicts are everywhere, they are
omnipresent.
Definition is more than purely descriptive (Cramer) as it shapes the interpretation and how a concept is
viewed. Yet, there is no agreement on an exact definition of conflict. According to Sambanis, very few
variables are robust to different definitions→most robust variable to explain conflict: GDP pro capita, so the
risk of having a conflict is related to GDP per capita (Figure 2.1 ppt): the lower the GDP, the higher the
conflict risk.
Definition is important also to provide exact data, e.g. Uppsala Conflict Data Program (UCDP) dataset
provides data on countries in conflict. According to UCDP:
• Armed conflict: state-based armed conflict is a contested incompatibility that concerns
government and/or territory where the use of armed force between two parties, of which at least
one is the government of a state, results in at least 25 battle-related deaths in one calendar year.
• War: state-based conflict which reaches at least 1000 battle-related deaths in a specific calendar
year
• Non-state conflict: use of armed force between two organised armed groups, neither of which is
the government of a state, which results in at least 25 battle-related deaths in a year.
• Conflict, intrastate with foreign involvement/internationalised: armed conflict between a
government and a non-government party where the government side, the opposing side, or both
sides, receive troop support from other governments that actively participate in the conflict, i.e.
conflict between government and non-government party supported from a foreign government
• International or interstate conflict: conflict between two or more governments
• One-sided violence: the deliberate use of armed force by the government of a state or by a
formally organised group against civilians which results in at least 25 deaths in a year, e.g. genocide
, Conflict definition:
→ Casualty threshold is very arbitrary
→ Absolute casualty numbers lead to skewed conflict datasets
→ Is battle-related deaths an appropriate measure? Substantial damage on civilians: direct and
indirect (e.g. famine, bad healthcare), so should civilian deaths be considered too?
In this sense, 2 elements are important:
1. Difficulty of defining conflict involving damage on civilians
2. Complexity of finding any type of data and reports on conflict zones (there is a war going on, so
availability of information is limited)
Conflict characteristics
Incidence of conflict by region from 1960-90:
• Low-income countries: corruption, governance institutions and polity might be rotten, so it is not
only GDP pro capita but there might be other features to take into account (spurious relationship)
• Reversed causality: rich countries may be rich because they did not have conflicts→different
explanations.
Indeed, wealthier states are better able to protect assets, i.e. rebellion has less chance of success. In poor
states the opportunity costs for joining a rebellion are lower and therefore more feasible, so more
rebellions and conflict. Poverty causes violence→scarcity leads to migration and fight over scarce.
