Skip to content

Geopolitics & Security

Global Extremism Monitor: Methodology

Explainer13th September 2018

The Global Extremism Monitor (GEM) has been designed to track:

  • attacks and attempted attacks;

  • victims (killed and wounded);

  • perpetrators (killed and wounded);

  • types of violent offensive (large-scale coordinated attacks on armed actors, attacks targeting the public space);

  • nature of attacks (including raids, assassinations, sectarian violence and engagement between groups);

  • method of attacks;

  • number and gender of suicide bombers;

  • nature of executions; and

  • demographic information of victim(s).

The GEM also looks at the counter-extremism efforts and responses being implemented by governments, including airstrikes, ground operations, ground and air operations, and arrests.

The monitor recorded steps taken by nonstate actors against extremism. These include, but are not limited to:

  • Kurdish peshmerga in Iraq and Syria;

  • vigilante groups, for example the Civilian Joint Task Force fighting Boko Haram in northeastern Nigeria; and

  • tribal and clan factions, working with or alongside multistate coalitions.

Chapter 1


The GEM draws information from over 400 English-language news sources that reported on incidents of Islamist extremism throughout 2017. Multiple news sources were used to verify accounts and information for each incident recorded. Over 700 additional materials, including reports, briefings and official group material including statements and propaganda, as well as existing data sets, were used to corroborate facts and reporting of incidents.

The GEM’s investigation of stories attempted to eliminate as much media bias as possible. The monitor employs multiple news sources to corroborate events and limit the amount of media bias in the events recorded. Consequently, the GEM data are a blend of local and international news sources, which are required to capture local events that often go unreported by the international media.

When verifying accounts, researchers did not take claims of responsibility to be a reliable account of an event. The GEM attempted to ascertain the true story of an event. For example:

  • In the reporting of incidents and numbers killed, political agendas or propaganda may obscure the facts, for instance when local security forces wish to inflate death counts of militants killed in operations, or when extremist groups downplay losses to maintain an image of success and strength.

  • Attacks that are not Islamist related may be claimed by groups for media attention, for example ISIS claiming October’s Las Vegas shooting or June’s casino shooting in the Philippines.

  • In so-called lone-wolf attacks, there is sometimes insufficient corroborated information on the exact direct relationship with a group to verify the extent of lone-wolf activity.  

  • Numbers of people killed by violent Islamists are often inflated in reports of claims.

  • Violent Islamists often fail to report the extent of any damage they have received.

When existing databases were used, all incidents were recoded and categorised according to the methodology designed for primary-source data gathering. When multiple sources were used, all events recorded by the GEM were verified for duplicates.

Chapter 2

Coding Criteria

After the verification process, each incident and event was recorded singularly in a spreadsheet for each country. Events were categorised for:

  • date of incident or attack, including for attempted attacks;

  • location of incident or attack (including geolocation at provincial and neighbourhood levels);

  • scene of incident or attack (for example, religious institution, school, tourist place, bank, refugee camp, hospital, or government building or embassy);

  • numbers and details of victims (killed and wounded);

  • numbers and details of perpetrators (killed and wounded) at group and individual levels;

  • target of violent attack or incident (for example, civilians, military, government, police, nonstate militias or vigilantes);

  • method of attack (including raids, offensives, assassinations, intergroup engagement, suicide mission and airstrikes);

  • tactics for attack (weaponry used, for example landmines, improvised explosive devices, vehicle ramming, stabbing, chemical weapons or drones);

  • motive for attack (based on official claims where available, as well as scene and location of attack, details of victims and details of perpetrators), for example religious sectarian (including inter- and intra-sectarian), political and social (based on government or military targets, targeting of local leaders and undermining national security operations), competition and territorial disputes;

  • hostage takings and arrests (including state-government arrests of terrorist suspects);

  • number and gender of suicide bombers, including intercepted bombers; and

  • reasons for extremist punishments (for example capital punishments, including executions, stoning and burnings, as well as amputations and lashings), including blasphemy, espionage, fleeing group territory, adultery, sorcery and drug usage.   

Over 50,000 events were documented in 2017. The GEM translated these events into thematic data strands and recorded them in a bespoke data set that captures the themes that characterise extremism. The data set was updated every quarter and verified according to the standardised verification process (see below).

Chapter 3

Data Capture on Groups

Tracking and monitoring groups has allowed the GEM to attribute an instigator to an event when such information is missing in news reports and the GEM could establish an instigator from other data recorded, including location analysis. In such cases, instigators are recorded only if there is a high probability of the group’s involvement.

For Syria, the GEM designed a specific data set to track and monitor the activities of groups in that country. The monitor captured data on all actors involved in the Syrian conflict, including rebel groups, nonstate militias and proxy actors. This was required to better understand the role violent Islamists play and their interaction with an array of actors.

Chapter 4


Geocoding for the GEM was conducted at the provincial and neighbourhood levels. The latitude and longitude of each incident was recorded to allow for GIS mapping of activity. Occasionally, the exact geolocation of an incident could not be verified. These typically occurred in conflicts such as that in Syria, and in countries with unclear coordinate data. In such cases, coordinates were attributed to incidents that were recorded in a province matching the activity common to a group’s insurgency. If there was a discrepancy in reporting where an incident took place, or if reporting suggested the incident took place along a border, towns were attributed near the disputed area where known activity was taking place, based on the likelihood that the incident was in line with overwhelming trends.

Chapter 5

Data Verification

A multiphase cleaning and verification process was essential. Inaccurate or contradictory information was flagged in the first phase of data collection. After the initial news coverage of an event ended, researchers corroborated information from media, government statements and academic sources to confirm details and address discrepancies. Existing data sets that capture armed conflict at the national and global levels were introduced to corroborate GEM capture, address discrepancies in accounts and confirm the absence of gaps in the final data set.

Chapter 6

Civilians as Intended Targets

When it can be determined, the GEM records the intended target(s) of an attack. In most cases, this is strongly related to the scene and location of an incident or attack and the type of victims (e.g. worshippers, military personnel, government figures, local leaders or humanitarian workers). In some cases, supplementary indicators are used. For example, the GEM has corroborated media reporting with group statements following the claim of an incident.

In addition to recording the numbers of victims per incident, the GEM also has ascribed characteristics when this is indicated in the reporting. For instance, while gathering details to determine types of victims, the GEM data set is designed to be disaggregated according to motive for attack (e.g. sectarian) and method of attack (e.g. suicide mission, offensive). Depending on the event, researchers allocate illustrative and contextual detail to supplement the primary fields of coding. These secondary data points and categories allow for wider analysis after coding, where multiple elements in the data set can be referenced to build a wider picture per incident.

Chapter 7

Suicide Bombings

The GEM captures information on suicide attacks and attempted suicide attacks globally. Beyond recording suicide attacks, the GEM was designed to accommodate the capture of:

  • methods of operation: whether a suicide vest or a suicide car bomb was detonated in an attack;

  • gender dynamics: whether female or male assailants were deployed on suicide missions; and

  • efficacy: whether the assailant deployed on a suicide mission managed to detonate his or her explosive device. If the assailant was intercepted before self-detonation, the mission was recorded as unsuccessful.

Researchers gathered evidence to determine the existence of any of the above attributes in a suicide attack. Media sources and extremist statements were cross-referenced to better understand how events unfolded and who was involved in the missions.

Chapter 8


The GEM captures information on punishments and executions by violent Islamist groups. The fields relating to executions and punishments in the data set were qualitatively rich and drawn from media reporting as well as group self-reporting. Details recorded, when possible, include the accusations made against a person or persons, and the legal verdict(s) applied.

Chapter 9


Sharia: Islamic law or norms as revealed in the Quran and the practices of the Prophet Mohammad, interpreted and applied by Islamic jurists through multiple schools of thought. Islamist groups often claim to be implementing their interpretation of sharia through their actions.

Caliphate: A form of Islamic governance that emerged after the death of the Prophet Mohammad. The leader, known as the caliph, must be appointed by consultation according to Sunni thought and should represent the interests of the entire Muslim community.

Hadith: A collection of traditions containing sayings of the Prophet Mohammad, which constitute the major source of guidance for Muslims apart from the Quran.

Haram: Forbidden by Islamic law.

Hudud: A set of punishments derived from a literalist interpretation of Islam.

Islamism: A modern religious-political ideology that requires a dominant role for an interpretation of Islam as state law.

Istishhad: The Islamic concept of martyrdom.

Shirk: Idolatry or polytheism, literally associating others with God.

Salaf: The first three generations of Muslims after the Prophet Mohammad.

Salafi-jihadism: A transnational religious-political ideology based on a belief in violent jihadism and return to the perceived Islam of the Prophet’s followers. This is the worldview espoused by groups such as ISIS and al-Qaeda.

Takfir: The act of declaring that a fellow Muslim is guilty of apostasy and therefore no longer a Muslim.

Ummah: The global Muslim community.

Chapter 10

The Full Report

Download the full Global Extremism Monitor 2017 or browse individual chapters:

Article Tags


Practical Solutions
Radical Ideas
Practical Solutions
Radical Ideas
Practical Solutions
Radical Ideas
Practical Solutions
Radical Ideas
Radical Ideas
Practical Solutions
Radical Ideas
Practical Solutions
Radical Ideas
Practical Solutions
Radical Ideas
Practical Solutions