Armed Conflict Location and Event Dataset (ACLED) Codebook

Similar documents
Armed Conflict Location and Event Data Project (ACLED) Codebook 3 Released in Clionadh Raleigh Andrew Linke Caitriona Dowd. University of Sussex

Armed Conflict Location & Event Data Project (ACLED) Codebook Version 8 (2017)

ACLED Actors and Interactions

Armed Conflict Location & Event Data Project (ACLED)

February 2013 USER GUIDE TO THE CCAPS MAPPING TOOL

Tables and Graphs. Figure 1: a) distribution violence per month - total; b) distribution Kenema/Kailahun (orange) vs. all other districts (blue)

Select Publically Available Conflict and Violence Datasets- Regional Typology Overview (October, 2015)

Department of Peace and Conflict Research, Uppsala University. This version compiled and updated by Marie Allansson and Mihai Croicu (2017)

African Conflict Baselines and Trends

Department of Peace and Conflict Research, Uppsala University. This version compiled and updated by Marie Allansson and Mihai Croicu (2017)

Data Comparison 2: All Country Years and a Focus on Nigeria and South Africa

UCDP Non-state Actor Dataset Codebook

CONFLICT TRENDS (NO. 18) REAL-TIME ANALYSIS OF AFRICAN POLITICAL VIOLENCE, SEPTEMBER 2013

UCDP Non-State Conflict Codebook

UCDP One-sided Violence Codebook

UCDP Conflict Termination Dataset Codebook v

MAP METHODOLOGY & DATABASE CODEBOOK v2

Definitions, sources and methods for Uppsala Conflict Data Program Battle-Death estimates

Social Conflict in Africa Database Version 3.0. Codebook and coding procedures. Idean Salehyan University of North Texas

Omega Psi Phi Fraternity, Inc. MyPage End-User Help Guide

Former Rwandan Tutsi-led rebel militia group, and later political party created in 1998

UCDP Battle-Related Deaths Dataset Codebook:

Conflict in the Democratic Republic of the Congo

Instructions for (Monthly) Workforce Utilization/Gross Wages Reporting CONSTRUCTION- CONTRACTORS/SUBCONTRACTORS

1) Read the article on American involvement in Vietnam

ecourts Attorney User Guide

LRA CRISIS TRACKER Mid-Year 2014 Security Brief January June 2014 PRINT VERSION

State of Texas Jury Management System. User Manual

Central African Republic

SOE Handbook on Certifying Candidate Petitions

User Guide. City Officials Historical Database. By Susan J. Burnett

AMNESTY INTERNATIONAL MEDIA BRIEFING

Online Remittance User Guide

Earmark Database 101

SOE Handbook on Certifying Candidate Petitions

Introduction to Announcements

I. Summary Human Rights Watch August 2007

Service Request Manual

Internal Displacement Monitoring Centre (IDMC) SOUTH SUDAN

TERANET CONNECT USER S GUIDE Version 1.4 August 2013

Summary This guide explains the general concepts regarding the use of the e- Nominations website Version 3.1 Date 07/02/ e-nominations...

NATIONAL CITY & REGIONAL MAGAZINE AWARDS

JoMUN XV INTRODUCTION

Mojdeh Nikdel Patty George

UCDP Actor Dataset Codebook

South Sudan JANUARY 2018

Darfur. end in sight. There are numerous aspects that lead up to the eruption of conflict in the area

eacademic Foundations Release 4.12

ACS Fellows Program Online Nomination System. Step-by-Step Instructions

ADVANCED SCHEDULING - OVERVIEW OF CHANGES COMING AUGUST 2014

The Digital Appellate Court Introduction to the edca Electronic Portal

EDITORIAL, DAILY RULES, CATEGORIES & ENTRY INSTRUCTIONS

The wealth of nations

Session 5: Violence and Conflict Trends in Africa

Nepal. Implementing the Comprehensive Peace Agreement

Class Action Registry. Handbook for lawyers. Direction générale des services de justice. Version 1.00

UCDP Actor Dataset Codebook

JoMUN XV INTRODUCTION DEFINITION OF KEY TERMS

South Sudan. Political and Legislative Developments JANUARY 2012

When should I use the Voting and Elections Collection?

The Seniority Info report window combines three seniority reports with an employee selection screen.

Introduction. The Security Council. The situation in South Sudan. Student Officer: Mila Escajadillo. Deputy President of the Security Council

Sudan - Researched and compiled by the Refugee Documentation Centre of Ireland on 13 July 2011

Clause Logic Service User Interface User Manual

Guernsey Chamber of Commerce. Website User Guide

DECISIONS. Having regard to the proposal of the High Representative of the Union for Foreign Affairs and Security Policy,

One View Watchlists Implementation Guide Release 9.2

Frequently Asked Questions

Ethiopian govt values concerns of protesters and is working to meet their demands envoy

SOMALIA. Abuses in Government Controlled Areas JANUARY 2013

Internal displacement due to conflict in Senegal. August Table of Contents. I. Internal displacement due to conflict in Senegal 1

They Shot at Us as We Fled. Government Attacks on Civilians in West Darfur H U M A N R I G H T S W A T C H

Service Request Manual

OCHA Regional Office for Central and East Africa Displaced Populations Report January June 2008, ISSUE 3

Intra Syrian Talks - Round Four: Grappling with Peace

DR Congo: waves of displacement follow heightened hopes of peace

January Authorization Log Guide

Atlanta Bar Association Website User s Guide

January 2011 country summary Chad

SCHOOLMASTER. Appointment Scheduling. Student Information Systems. Revised - August Schoolmaster is SIF Certified

JoMUN XV INTRODUCTION

Post-Elections Report Post-election: 31 July 19 August, 2018 (20 days post elections) Report Date: 21 August, 2018

Dealing with the fast-changing environment in the eastern DRC. The split in the CNDP

Navigating the South Dakota Legislature website

A Major Challenge to the Sustainable Development Goals. Andrew Mack and Robert Muggah

CONFLICT TRENDS (NO. 9): REAL-TIME ANALYSIS OF AFRICAN POLITICAL VIOLENCE, DECEMBER 2012

INTERNET SCHEDULING USING BENCHMARK

STATEMENT OF AMBASSADOR SMAIL CHERGUI, COMMSSIONER FOR PEACE AND SECURITY, AT THE HIGH LEVEL MEETING ON THE LORD S RESISTANCE ARMY

Office of the Special Representative for Children and Armed Conflict

CONFLICT TRENDS (NO. 14): REAL-TIME ANALYSIS OF AFRICAN POLITICAL VIOLENCE, MAY 2013

FREQUESTLY ASKED QUESTIONS Updated 1/26/11. User IDs and Passwords

Central African Republic crisis ECHO CRISIS REPORT N 9

Creating and Managing Clauses. Selectica, Inc. Selectica Contract Performance Management System

Security Council The question of Somalia and the spread of terrorism into Africa. Sarp Çelikel

World History (Survey) Restructuring the Postwar World, 1945 Present

Charter Renewal Guidebook

Manage Subpoenas. DA IT Video Library. Supporting Documentation Facilitator: Teresa Radermacher Recorded: November 2008 Duration: 1 hour, 16 minutes

1 e-nominations Parts of e-nominations Basic Principles The e-nominations homepage Global Position...

BOND CREATION AND MANAGEMENT SURETY COMPANY

Violence Affecting Women and Girls in the Eight NSRP Target States

Transcription:

Armed Conflict Location and Event Dataset (ACLED) Codebook Clionadh Raleigh, Andrew Linke & Håvard Hegre Centre for the Study of Civil War, International Peace Research Institute, Oslo (PRIO) 1

Table of Contents 1 Introduction and Brief Description...3 2 Definitions of ACLED events...3 2.1 Conflict Actors...3 2.2 Conflict Events...5 2.2.1 Battle Definition...7 2.2.2 Headquarter or Base Establishment...9 2.2.3 Rebel presence...9 2.2.4 Rioting... Error! Bookmark not defined. 2.2.5 One sided violence...9 2.2.6 Non violent transfer of location control...11 2.2.7 Important notes regarding event type codes...11 3 Coding Forms...11 3.1 Armed Conflict Form...12 3.2 Actor Forms...13 3.2.1 Predecessor...14 3.2.2 Succession Group...15 3.2.3 Startdate...15 3.2.4 Enddate...15 3.2.5 Ethnic Affiliation...15 3.2.6 Notes...15 3.2.7 Editing Actors...15 4.3 Event Forms...17 4.3.1 Adding Actors to Individual Events...17 4.3.2 The Location of the Event...18 4.3.3 The Type of Event...20 4.3.4 The Date of the Event...23 4.3.5 Territorial Transfer...23 4.3.6 Publication...23 4.3.7 Notes...24 5 Conflict Event Reports...24 6 Relationships to other datasets:...24 Important notes: 1) This codebook describes a publicly available beta version of ACLED. 2) This codebook concerns a) data entry in the ACLED web portal and b) understanding the output from the ACLED web database. 2

1 Introduction and Brief Description The ACLED dataset codes reported information on the exact location, date, and additional characteristics of individual battle events in unstable and warring states. There is a specific focus on: Tracking rebel activity Locating rebel group bases, headquarters, strongholds and presence Distinguishing between territorial transfers of military control from governments to rebel groups and vice versa Recording communal violence between ethnic groups and militias Collecting information on rioting and political protests 2 Definitions of ACLED events ACLED definitions mainly concern actors and events. ACLED collects and codes reported information on civil and communal conflicts, violence against civilians and rioting in the developing world. For the most part, ACLED covers activity that occurs within the context of a civil war. However, much violent activity can occur outside of civil wars, including violence against civilians, communal conflict and rioting. 2.1 Conflict Actors In ACLED, conflict actors include rebels, militias, and organized political groups who are involved in events over issues of political authority (i.e. territorial control, government control, access to resources, etc); the specific types of political events collected are discussed below. A range of actors is recognized including: governments, rebels, militias, ethnic groups, active political organizations, and civilians. Governments are defined as internationally recognized regimes in assumed control of the territory of a state. In ACLED, government actors are defined as a series of regimes (e.g. Congo/Zaire 1965-1997 (Mobutu), Democratic Republic of Congo 1997-2001(Kabila, L) and Democratic Republic of Congo 2001-2008 (Kabila, J) as opposed to Congo/Zaire (1962-present). The strength of different governments can change significantly, and ACLED means to capture the differences in government s violent actions. As militaries are the direct arm of the government, these actors are noted as Military Forces of State, 19xx-19xx within the ACLED dataset. Mutinies of militaries are coded as Mutiny of Military Forces of State, 19xx-19xx. Any action that involves the military directly must note the military actor instead of government actor. 3

Rebel groups are identified as political organizations designed to counter an established governing regime within a state through engaging in violent acts. Rebel groups have a stated political agenda, are acknowledged outside of immediate members, and use violence as the primary means to pursue political goals. Rebel groups often have predecessors and successors due to splintering and diverging goals. These relationships are tracked in ACLED. Rebel actors are added to ACLED with both their name and acronym. If a splinter group emerges, the rebel name is repeated with the splinter leader (or other identification) of the splinter group added. Militias are groups of armed participants who engage in military activity but are not professional soldiers. ACLED includes such groups if they are involved in violent events due to local/national politics. Militias are more difficult to assess as they can be created for a specific purpose or during a specific time period (i.e. Janjaweed) and can be associated with an ethnic group, but not entirely represent it (i.e. Kenyan Luo ethnic militias). Alternatively, ethnic militias can be long term policing units, such as those common across Somali clans. ACLED coders distinguish between active ethnic militias when recording actors involved in communal violence (militias against each other) or violence with governing forces that is outside of a civil war context (i.e. Karamonjong violence in Uganda). If an ethnic community has formed a militia to engage in intra or inter communal violence, such groups are designated ethnic group x militia. Other militia groups, such as the Mayi-Mayi, are not associated with a particular ethnic community and may be recorded by their acknowledged name. ACLED s definition of organized political groups includes militias operating in conjunction or in alliance with a recognized government. These groups are either supported, armed, or allied to the government forces and perpetrate acts towards a similar goal. Such groups include the Sudanese Janjaweed or Serbian Tigers. They are not subsumed into the category of government, but must be added to ACLED actors and noted for their alliance to government troops. When these groups and governments operate together, their alliance is noted for each event. ACLED records reported information on both spontaneous and organized rioting and protests. Rioters and protestors are actors involved, respectively, in violent rioting or peaceful protests. Riots are defined as a violent disturbance of the public peace by three or more persons assembled for a common purpose. Rioters belong to a group spontaneously assembled to violently protest a government action. Protestors belong to a group assembled to peacefully protest a government action. Either group may interact 4

with military/government forces, but interaction is not necessary for inclusion. As discussed further in the event description section, if rioters violently interact with military forces such an event is a battle in ACLED terminology. If military forces attack protestors, it is considered violence against unarmed civilians. Based on news reports, an action may escalate from a protest to a riot. An organized riot/protest involves a previously recognized political group. The organized group rioting can be a political party (i.e. ZANU-PF in Zimbabwe) or another recongized entity such as a section of the military. Spontaneous riots/protests are without direct reference to an organized political group. Spontaneous protesters and rioters are a general category and each are noted as Protestors (Country X) or Rioters (Country x) in the dataset. Civilians are unarmed victims of violence in the ACLED dataset. Civilians are, by definition, never involved in participation or perpetration of a violence act and therefore cannot be included in any category but violence against civilians. The civilian actor is a general term unless there is evidence that a particular defined group is targeted. In such cases, the civilian actor can be entered as e.g. Civilians (Banyamulenge-DRC) or Civilians (Fur, Sudan) or Civilians (MDC Party Supporters, Zimbabwe). If there is not clear evidence of targeting a specific group, then the general category Civilians (Country X) is entered, and suggested target information recorded in the notes category. In ACLED, all actors have an official name, a stated political purpose and use violence or protest as a means to a political end. Organizations are cohesive and are not designed for single events. Further, the events in which they are involved are connected. This necessary and sufficient definition of actors allows us to track a range of movements from rebel organizations to various types of political rioting. 2.2 Conflict Events The fundamental unit of observation in ACLED is the interaction of actors in an individual event. Events occur between designated actors a rebel group, a rebel group faction, or a government 1 and are coded to occur at a specific point location (name, coordinates, etc) and on a specific day. Efforts are taken to ensure that the most specific location and time possible are coded. ACLED s currently codes for eight types of events, both violent and non-violent, that generally occur within the context of a civil war and state failure. Figure 1 displays the eight ACLED events. 5

Table 1: ACLED Event Types Event Number Event Type Event Description 1 Battle-No Change of A battle where the control of the Location Control fighting location is not exchanged. This is the most frequent event. If the government controls an area and fight with rebels but win, this is the correct code. If rebels control a location and fight with government forces, this is the correct code. If two rebels are fighting and the group in control of the location still has authority in that location, this is the correct code. 2 Battle- Rebel Control A battle where rebels win control of Location location. This is the correct code if, after fighting with another force, a rebel group acquires control of a location. If two rebel groups fight and the group which did not begin with control acquires it, this is the correct code. 3 Battle- Government Regains Control A battle where the government regains control of a location previously lost in a government-rebel battle. This event type is solely used for government reacquisition of control. 4 Headquarters or Base A rebel group establishes a base or Establishment headquarters. This event can be nonviolent, but must be coded when a semi-permanent base is established by a group. 5 Non-Violent Rebel Presence This event means to record activity by rebel groups that does not involve active fighting but are within the 6

context of the war/dispute. Examples include recruitment drives, incursions or rallies. 6 Rioting/Protesting Protest involves a group involved in non-violent public meeting against a government institution. Rioting is a violent form of protest. The actors for this group are noted as protestors (country) or rioters (country). Interaction with government forces is not mandatory. 7 Violence Against Civilians Violence against civilians occurs when any armed groups attacks unarmed civilians within the context of a larger conflict. Rebels, governments, and militias can all perpetrate violence against unarmed civilians. This is the only event that can involve civilians. 8 Non-Violent Transfer of This event is for situations where Location Control rebels or government s acquire control of a location without engaging in a violent act. 2.2.1 Battle Definition ACLED defines a battle as a violent interaction between two politically organized armed groups at a particular time and location. Typically these interactions will occur between government militaries/militias and rebel groups/factions within the context of a civil war. However, these interactions also include rebel on rebel violence and military on military violence. There is no causality minimum necessary for inclusion. The specific elements of that definition are as follows: (1) A violent interaction is the use of armed force, including guns or military hardware, machetes, knives or any tool to inflict harm upon the opposing side. 7

groups. (2) Organized armed groups including but not limited to rebel and government Battle events are coded in one of three ways depending on the result of the armed clash. (1) No transfer of territory ACLED assumes that the government is in control and holds all territory under its internationally recognized mandate. When rebels and governments engage in battle and there is no report of rebels winning control or governments losing control of the location in which the battle was fought, the event is coded as battle with no transfer of territory. In situations where there is rebel control at night but no during the day, these areas are assumed to be tacitly under government control. (2) Rebels gain territory If, at the end of a battle, rebels have won control and/ or subdued government forces in the area of contestation, this event designation is used. Rebels have won territory when they are capable of acting with impunity and are regarded as having a monopoly of force within that territory. Although rebels and government forces may fight numerous times in that location after the rebels gain hold of the territory, only in initial interactions, or when rebels are regaining hold of a territory after losing it, will this event type be used. For continuous battles between rebels and a government after the rebels acquire the territory, event type 1 (battle-no transfer of territory) is used. This event can also be used to note the transfer of control from one rebel group to another as a result of inter-rebel violence. Side 1 will be the rebel group that currently controls the area (if the territory has been lost by the government). Rebel on rebel violence uses this and the previous code (battle with no transfer of territory), and no others. (3) Government gains territory This event is coded if after a rebel group has gained hold of an area, further battles between the government and rebels result in the government re-establishing control over that location. This code is only used for government re-establishment of control and not for events concerning rebel on rebel violence. In the course of the civil war, control over locations may change hands multiple times. Depending on the eventual outcome of the war, rebel losses and rebel gains must match. If the rebel gains control of the government, then all rebel acquisitions may stay 8

in rebel hands; however, if the rebel loses to the government, all territory must be returned to government control at recorded points in the data. 2.2.2 Headquarter or Base Establishment Rebels will establish a number of bases and headquarters during their active periods. This type of event records the date, location, and actor involved in base or HQ establishment. Rebels can have a number of active bases at one time. These bases can be inside or outside the country/countries in which the insurgents are operating. In the notes field, please distinguish whether the position is the main headquarter or a regional base (or forest base, mountain base, border base, etc). Please be careful to note whether bases or HQ have been abandoned in the course of the conflict. If a base or HQ is no longer in use because of government or other rebels encroachment, note the estimated final date of use in the dataset by using code 2 if other rebels overtake the area or code 3 if the government resumes control of the area. If a rebel group abandons a base, it can be assumed that the government has retaken that area. If this occurs through a battle, it is noted as event type 3 (battlegovernment regains territory). If it occurs without a violent interaction between rebel and government forces, it should be coded 8 (location control transfer without violence). 2.2.3 Rebel presence The rebel presence category records all actions wherein rebels were present in an area, not engaged in fighting a battle nor setting up bases, and not harming civilians. This event is necessary as the literature will mention instances when rebels are active without military engagement with a military or other rebel force. However, unlike the event type for one-sided violence, rebel presence notes that rebels may be present or active in a town without reference to violence between government and rebels, rebels and rebels, or rebels and civilians or without controlling this area. Such nonviolent rebel actions include recruitment, rallies, gathering foods, speeches, looting, destruction of property etc. 2.2.4 Protesting and Rioting Riots are defined as a violent disturbance of the public peace by three or more persons assembled for a common purpose. Protests are a nonviolent gathering to ACLED records reported information on both spontaneous and organized rioting. Either group 9

may interact with military/government forces, but interaction is not necessary for inclusion. If the rioters and protestors engage in their respective activities without interaction with military forces or a violence group, then such an action is recorded under the riot/protest event. If rioters violently interact with military forces such an event is a battle in ACLED terminology. If military forces attack protestors, it is considered violence against unarmed civilians. Based on news reports, an action may escalate from a protest to a riot. Should a news report note that a protest became a riot, both actors (protestors and rioters) can be included as actors. Escalation or de-escalation over the course of a day is not covered in ACLED directly, but can be mentioned in the notes. However, escalation from protests and riots in consecutive days should be recorded as separate events (note the connection to a previous day s events). 2.2.5 One-sided violence One sided violence is defined as deliberate violent acts perpetrated by an organized political group, typically either a rebel or government force, on an unarmed noncombatant. These acts are of a political nature, and result in the harming or death of civilians. There is no minimum number of people who need to be affected by this violence to qualify as an ACLED event. One-sided violence includes killing either of an individual or group. Although the individual can be a combatant in a different context, here they are UNARMED and NOT ABLE to defend themselves against directed force. This definition also includes inflicting significant harm (e.g. bombing, shooting, torture, rape, mutilation etc) and accosting person/persons (e.g. kidnapping and disappearances). It does not include actions where people are not physically harmed, such as looting or burning of villages, destruction of sacred spaces, and forced displacement. Such events are covered in ACLED by event 5 for rebel presence. 2 ACLED includes one-sided violent events that occur outside of an armed conflict. These violence in peace episodes are difficult to ascertain. However, when found in the literature they will be coded as a separate episode, outside of the typically defined war times. 2 Government actions of this type will be able to be coded under this category in the near future 10

2.2.6 Non violent transfer of location control When control of a location changes between groups without any violence (battle), record this as the event type. This may occur in situations where rebels acquire control due to the absence of any government group ir in cases were government retake an area previously used as a rebel base. 2.2.7 Important notes regarding event type codes One or more events can occur within the same location on the same day. If two similar events between the same actors in the same locations are reported, this can be noted by one single event. However, if an additional event type occurs, this should be coded. In most cases, an event will require at two actors. However there are two caveats to this general rule: if groups are allied to either of the main actors, this should be noted. Event types 4, 5, 6, and 8 can be single actor events, as headquarter establishment, nonviolent territorial transfer, rebel presence, and protesting can occur without response from the government. Civilian actors can ONLY be included in event type 7. If civilians engage in organized violence, they are not operating as unarmed civilians. If a group is clearly organized but unknown or unnamed, 'unidentified armed group' will suffice. 3 Coding Forms 3 Individual coders are required to register personal information including name, email and institution, within the ACLED dataset (see Figure 1). Each coder will have access to edit only those conflicts associated with their individual coder profile (see Figure 2). 3 The structure of the coding portal changes to accommodate new specifications, and new sections will be included in this codebook to explain those changes. 11

Figure 1 Login Form 3.1 Armed Conflict Form After the coder has logged in, the first screen to appear contains the conflicts accessible to that coder (see Figure 2). When first beginning, a coder will find that the area designated your coded actors will be empty. Coders have the choice to either code New Actor or Add Actor. To add an actor that is already coded for another conflict- for example a cross border rebel group or a supportive regime in engaged in a separate conflict, click on Add Coded which will bring you to a list of your previously coded actors for all conflicts assigned to you. To add a new actor, click on New Actor (see figure 3). Coders will begin all coding by choosing to either add or edit actors or to Code Conflict Events which link to the conflict event coding (see Figure 6). Figure 2 Armed Conflict Form 12

3.2 Actor Forms All rebel groups, and all their active factions, are actors in ACLED. In addition, different regimes of state government are separate actors. However, not all successive governments are actors. Only those governments which have a substantially different leadership profile from the previous government should be designated as separate actors. Hence, separate governments within the same state should be designated as new actors if the change of government was a violent or intensely difficult one. For example, leadership passing from a dictator to a hand picked successor will not constitute a major change in government, but a rebel leader acquiring power is a different regime to the overthrown, previous ruling power. There should not be more than one government actor active at one time. Governments actors should be designated as: Military Forces of Country x, 19xx-19xx and/or Police Forces of County x, 19xx-19xx. Additional government forces should be added if they are substantially different from those above. Examples of such actors include Presidential Guard of Country x, 19xx- 19xx or Mutiny of Military Forces of Country x, 19xx-19xx. 13

Figure 3 Adding Actors From the pull-down menu Get Actor, all previously noted actors as part of a designated conflict will be available. If an actor is not recorded here, skip this pull down menu and proceed to filling in the name etc of the new actor. Coders will designate the different governments active in the conflict in this manner. Once an actor is chosen, the coder can add descriptive information regarding active periods, ethnic affiliation (if any), home territory, whether this actor is the descendant of another actor (e.g., a government that was previously a rebel group). 3.2.1 Predecessor The predecessor is the actor (name/acronym) this actor was previously known as. For example, the UPDM (Uganda People's Democratic Movement) rebel group in Uganda was primarily made up of the previous government s military; therefore in the predecessor field of the rebel group, the Military Forces of Uganda, 1980-1985 government/military designation is chosen. 14

3.2.2 Succession Group The succession group is the actor (if any) this actor evolved into. For example, the Tigray People s Liberation Front of Ethiopia (TPLF) became the Ethiopian People s Democratic Revolutionary Movement (EPDRM) when it allied with other groups to finally topple the Meginstu regime and established a new government in Addis Ababa. 3.2.3 Startdate This field records the date of the first event in which this actor was involved, unless there is evidence that the group formed well before any events took place. These dates can be general, due to the possibility of a predecessor or early group formations. Dates are recorded as year-month-day, separated by dashes. 3.2.4 Enddate This field records the last day of activity for the group in question. For governments, it should be the last official day holding office, for rebel groups it should be the last day of recorded activity, regardless of ceasefire or related information. Dates are recorded as year-month-day, separated by dashes. 3.2.5 Ethnic Affiliation If the actor is associated with one ethnic group more than any other, or if ethnic identity is an important component of group mobilization, note the ethnicity here. For example, the exfar/interahaweme militias are extremist Hutus, so this should be noted here. 3.2.6 Notes Coders may record any information of interest regarding this actor in the notes field. 3.2.7 Editing Actors If coders want to edit the coded actor information or delete an actor (from the database or from the conflict) they can click on the actor name in the conflict form, (See Figure 7) and get the Edit Actor form. (See Figure 6) Figure 4 Editing Actors 15

3.2.7.1 Save Actor/Delete Actor These buttons allow the coder to save or delete additions or edits to the actor form (see figure 7). All coded actors will appear in the your coded actors field of the conflict form. Any necessary edits to be made to actor information from this page will return the coder to the Armed Conflict form. Once actors are chosen, the coder can move onto Code Conflict Events. Figure 5 Coded Actors 16

4.3 Event Forms In the coded events form, coded events can be searched, browsed, sorted, selected and deleted. (see Figure 6). All events for the current conflict will appear on this page. The search action narrows down the list of events. To start with all events for a new search push the reset button underneath the search field. Figure 6 Coded Events Form To add a new event, a coder selects Add New and is transported to the coding page (see Figure 7). To delete or edit an event, first click the event (s) on the select column, and proceed to edit or delete. To sort the events within this conflict, click on the column headings. 4.3.1 Adding Actors to Individual Events To begin coding a conflict event, you can code from the left to right side. Actors can be selected last from the actor column. All actors the coder has added to the conflict will be displayed here, with their active dates also noted. All actors involved in an event should be chosen by checking the box beside the actor name. At least one actor must be chosen. An Actor s role is then selected. The options for role include Actor 1 and 17

Actor 2. Choosing a particular designation does not confer any particular significance, but government, if involved, are often designated Actor 1. It is possible that multiple groups can be designated as allies of Actor 1 or Actor 2. If the government or the rebel group is involved in an active alliance with a group that participated in this event, that additional group should be designated a role based on whom they are allied with. In episodes of one sided violence, the perpetrator (either government or rebel group) should be designated Actor 1, and the civilian population as Side 2. Figure 7 Coding Event Information 4.3.2 The Location of the Event To add a location, the coder must click select location in the location box. Most locations are already recorded in an ACLED Gazetteer. A new page will appear where the coder can search the ACLED gazetteers according to country, general region or the name of the village, town, or area (see Figure 8). A number of possible matches will appear in a new window (see Figure 9). The more information is specified before, the smaller the range of options for a location will be. 18

Figure 10 Finding a Location Figure 11 Choosing a Location 19

A location is chosen by clicking on the location name. All associated information, including the coordinates, regions and country will be automatically recorded in the previous event page. Coders will automatically be returned to that page where they will need to designate a precision level. This will be based on the source material. If the source notes a particular area, and coordinates are available for that area, then the highest precision level is chosen. If the source material notes that activity has taken place in a small part of a region, and notes the general area, coders must decide on a town with georeferenced coordinates to represent that area, and must note part of region in the precision level drop down menu. If a larger region is mentioned, the provincial capital is chosen to represent the region and region is chosen for the precision level. With few exceptions, coders should choose the provincial capital to represent the larger regions. Coders can deviate from this choice if there is an indication the activity happened in an area other than the provincial capital 4. If a location does not appear, coders have the option of recording this information through the Add Location button. The coordinates of these possible locations can be found through searching online gazetteers, including www.fallingrain.com. This website collects coordinates of villages, towns, cities, and airports by country alphabetically. It also will note alternate spellings and names of locations. If you do add a location, please add proper name, administrative zone it sits in, country, and type of location (city/airport etc). 4.3.3 The Type of Event Event type is the most critical designation in this larger form. It records the kind of event that occurred between actors. Figure 10 Selecting an Event Type 4 The location table includes georeferenced coordinates and geographical characteristics of each location. ACLED has an internal gazetteer that is derived from a larger gazetteer project available for viewing and downloads at: http://earth-info.nga.mil/gns/html/index.html. Each location is associated with a feature designation, which indicates the type of location it is. For a complete list of feature designations, visit the website: http://gnswww.nga.mil/geonames/desig_code/desig_code_search.jsp 20

As noted in section 2.2, event type is determined by a series of decisions related to action, actor, and consequence where a coder chooses whether the action was violent or nonviolent, which actors (s) were involved in the action, and what was the consequence of the action. A drop down menu will note the following choices of event activity: A battle resulting in no change of territory. This event describes a battle in which the literature notes no transfer of territory. By default, all territory belongs to and is controlled by the government. Unless specifically noted, all battles result in no territory transfer. If there is a battle in rebel controlled territory and it does not result in the government regaining the territory, this battle event type should be chosen. Note that although the term battle is used here to describe different kinds of encounters/events involving the parties, e.g. the ceasefire is broken, it must be a violent event containing at least two actors. Communal violence may be coded as a series of battles with no transfer of territory, as communal actors are not officially considered owners of towns and larger localities. 21

A battle resulting in a transfer of territory to the rebel actor. This event is recorded when the literature notes that the rebel is in control of an area after a battle with the military. It only is for rebel holdings, since by default all territory is controlled by the state. A battle resulting in government forces recapturing rebel held territory. This event can only be coded when government forces resume control of a location. A rebel base or headquarters established. This event is coded when the rebels take over a location and establish a base or headquarters (with or without a battle). If a headquarters is established, note this in the notes field as headquarters are of more importance than bases. Rebel activity that is not battle related/ Rebel presence. This event is necessary as the literature will note a lot of instances were rebels were active without military presence to counteract the rebels. However, unlike the event type one-sided violence, rebel presence notes that rebels may be active in a town- hiding, recruiting, organizing etc- without reference to violence between government and rebels, rebels and rebels, or rebels and civilians. Rioting/Protests. This event type records protests and riots initiated by a political organization. This organization can vary: the group can have existed prior to the riot and continue to exist after the riot. The members of the political organization do not have to be the sole participants, but must be the main organizing force of the riot. In such cases the name of the rioting or protest group can be entered as an actor. In most cases, events will involve riots and protests initiated by unorganized political groupings (i.e. disgruntled young men or unemployed laborers), in such cases a general Rioters (Country) or Protestors (Country) designated in sufficent. One Sided Violence. This event type records all activity where either rebel or government actors perpetrated violence upon unarmed people/civilians. This event must involve violence and is the only event where civilians can be an actor. Location Transfers Without Violence. Use this event type if control of a location (i.e. village) is transferred to a rebel actor without violence. In this case, the government group is Side A and the rebel organization is Side B. If locations are transferred between rebel groups, the recipient rebel group if Side B and the former location authority is Side A. 22

4.3.4 The Date of the Event The event date should be the most exact date possible from the literature source. Year, month and day should be recorded in that order, separated by dashes. If a date is chosen without exact knowledge, note this in the precision drop down menu. Similar to location precision levels, if sources note an actual date day should be chosen from the drop down menu. If sources note a week, choose the first day of the week and note week in precision drop down menu. If sources note an activity took place within a particular month, without reference to the general time of month, coders should choose the mid-point; if early month is noted, choose the beginning of the month, if late month is noted, choose the end of the month. For example- if the literature notes a battle in the beginning of the month, choose the 1 st of that month; if middle-15 th ; if end, choose 30 th. Note month on the drop down precision level. It will not be helpful to estimate past month levels. 4.3.5 Territorial Transfer At some point, the literature may note entire zones or parts of a country as being in rebel hands. To demonstrate this through the use of multiple rebel gains location events can be arduous. Since each country is broken into administrative zones, a coder can note that an entire zone is a territorial transfer to a particular group. If the rebels are in control of many zones (e.g. all those that create the western part of a country), a number of territory transfers can be made on the same day. Even though a number of held positions may be present within the zones, this function of territory transfer gives a map viewer a better impression of rebel holdings. Transfer Actor. This notes the actor (rebel or government) that has gained or regained entire regional areas. 4.3.6 Publication Publication notes the name, acronym, website, or letter of the primary information source. One is necessary. There is a pull down menu of some of the more common sources. For example, BBC Monitoring can be chosen in the pull down menu, and the second open box should be filled with specifics of the BBC Monitoring report (date, page, web address). The details on the publication must be sufficiently complete so that a data user can find the reference with ease. Coders are required to keep a separate document where source details are well specified, if not a direct copy of reports. 23

4.3.7 Notes Any important notes can be included at this point. Long notes are not necessary and instead should be recorded in monthly conflict update reports written by individual coders on each country. 5 Conflict Event Reports Conflict reports can be downloaded directly from the ACLED dataset by clicking create report. A conflict report will give details for every conflict on your user screen. In ACLED s guest coder version, the conflict report has all available data separated by conflict ID number. Alternatively, clicking on the Actor Events can generate an actorbased conflict report. An Actor Event conflict report will present all events by individual actors, hence repeating individual events for both Side A and Side B etc. By default, all reports are in Excel CSV format. To use these data in a GIS interface, the excel sheet will need to be converted to a.dbf file. 6 Relationships to other datasets: The conflicts coded in the dataset are generally compatible with other conflict data collections, including the UCDP/PRIO data on armed conflicts. ACLED information between many, if not all, other conflict datasets can be joined either by country, year, actor or specific locality. ACLED contains information about extrasystemic armed conflict, internal armed conflict, and internationalized armed conflict. Generally, the ACLED dataset disaggregates civil wars into their constituent events. However, since the threshold for inclusion in an event dataset is lower than that for the conflict dataset, ACLED occasionally codes information for actors that are not in other datasets. It should be noted that the threshold for inclusion as an ACLED event is lower than most civil war datasets. As a result, ACLED occasionally records events that do not fall within the dates or the set of actors recorded in other datasets. In ACLED, not every event will include the government as a combatant. Information on actors are coded independently of the fixed government/rebel dyad combinations in the UCDP/PRIO dataset and each warring party is recorded as an independent actor as the changing roles dictate (i.e. a rebel group can become a governing regime and vice versa). 24

Below are links to websites providing an overview of GIS, information about importing data and map-making. 1. http://ltgis.launchpad.onenw.org/technology/gis-data 2. http://www.colorado.edu/geography/gcraft/notes/intro/intro_f.html 3. http://gislounge.com/ll/basics.shtml 4. http://www.gis.com/ 5. http://www.gisdevelopment.net/tutorials/ 6. http://www.colorado.edu/geography/gcraft/notes/datacon/datacon_f.html 7. http://www.colorado.edu/geography/gcraft/notes/mapproj/mapproj.html 8. http://www.colorado.edu/geography/babs/geog_4103_f06/ 25