Principal Investigators:

David R. Davis, Department of Political Science, Emory University

& Will H. Moore, Department of Political Science, University of California, Riverside

Coding Rules Authors:

Brett Ashley Leeds, David R. Davis, and Will H. Moore, with Christopher McHorney

20 September 1995

Version 7.3


1. Introduction

1.1 The Goals of the Project p. 4

1.2 Comparison with Other Event Datasets p. 4

1.3 Conceptualizing Conflict p. 6

1.4 Conceptualizing Cooperation p. 6

2. Coding Events

2.1 Defining a Political Event p. 7

2.2 Actor and Target Assignment p. 8

2.3 Multiple Actors and Targets p. 10

2.4 Coding Multiple Events p. 10

2.5 Dating the Events p. 13

2.6 Controversial Reports p. 14

2.7 Distinguishing Domestic Events from International Events p.

2.8 Sources for Events Data p. 15

2.9 Ambiguous Values p. 15

2.10 The ‘Deaths’ Variable p. 15

2.11 The Coding Sheet p. 17

2.12 Managing Computer Files p. 19

3. Actor and Target Codes

3.1 Selecting Actor and Target Codes p. 20

3.2 The State p. 21

3.3 Mass/Elite Distinction p. 22

3.4 Unspecified Actors and Targets p. 22

3.5 Business and Labor p. 23

3.6 Multinational Groups and Foreign Nationals p. 23

3.7 Leftists and Rightists p. 24

3.8 Religious Groups p. 24

3.9 Socially Influential Individuals p. 25

3.10 Supranational Organizations p. 25

3.11 Case Specific Codes p. 25

3.12 Distinguishing Among Ethnic Groups, Social Groups, and Dissident
Groups p. 26

3.13 Coding Factions and Coalitions p. 26

4. The Conflict Scale

4.1 Introduction to the Conflict Scale p. 27

4.2 Press Criticism of the Government p. 27

4.3 Rumors of Threatening Action p. 27

4.4 Strikes, Protests, and Riots p. 28

4.5 Executive Adjustment, Political Resignations p. 29

4.6 Censorship p. 29

4.7 Mobilizing Resources p. 30

4.8 Violent Political Acts p. 30

4.9 Mass Arrests and Political Convictions p. 30

4.10 Bombings p. 31

4.11 Assassinations p. 32

4.12 Reports of Disappearances p. 32

4.13 Breaking a Truce, Breaking a Peace Treaty p. 32

4.14 Attacks, Clashes, and Battles p. 32

4.15 Occupation of Territory p. 33

4.16 Coups d’Etat p. 33

5. The Cooperation Scale

5.1 Introduction to the Cooperation Scale p. 34

5.2 Promises p. 34

5.3 Relaxation of Repression p. 35

5.4 Political Reform p. 35

5.5 Talks, Agreements to Talk p. 36

5.6 Surrenders, Releases of Prisoners or Hostages p. 36

5.7 Cessation of hostilities p. 37

5.8 Elections p. 37

6. Using the Conflict and Cooperation Scales Together

6.1 Using the Scales Together p. 38


1.1 The Goals of the Project

The Intranational Political Interactions (IPI) project is designed to measure political conflict and cooperation within societies through the coding of political event reports from international, regional, and local sources. These events are coded on two ten point scales which reflect the severity of various cooperative and conflictual statements and actions. We use this scaled events data to calculate the volume and intensity of political conflict and cooperation within the domestic polity. In addition to facilitating the calculation of general levels of political conflict, the IPI coding scheme allows the examination of the dynamics of interaction among specific groups within the society. While scholars have focused a great deal of attention on the causes of political violence, internal war, and revolutions, their explanations have typically been based on attributes of the parties involved and attributes of the socieconopolitical environment in which these conflicts occur. As a consequence, the literature has failed to address the impact of the behavior of these groups and, more importantly, the interactions between groups. IPI gives scholars the ability to track interactions among social groups and between the state and social groups.

1.2 Comparison with Other Events Datasets

Two data collection efforts begun in the 1960s [COPDAB (Azar, 1982) and WEIS (McClelland, 1972)] have given scholars of international relations a useful bank of information with which to examine issues of conflict and cooperation among nation states in the international system. Referred to as “events data”, these data record the behavior of nations interacting with one another. Burgess and Lawton define events data as “words and deed–i.e. verbal and physical actions and reactions–that international actors (such as statesmen, national elites, intergovernmental organizations (IGOs), and nongovernmental international organizations (NGOs) direct toward their domestic or external environments” (1972, p. 6). The availability of this data has spawned a large literature examining peace and conflict in the international system. Unfortunately, similar data have not been available for the study of conflict and cooperation within nation states. The IPI project is an attempt to fill that void.The only existing dataset which attempts to measure the behavior of domestic actors is the “Daily Political Events Data” from The World Handbook of Political and Social Indicators (Taylor & Jodice, 1983). The World Handbook provides information on several classifications of conflict relevant events, including four types of demonstrations, six types of armed attacks, two types of strikes, seven types of leadership changes, five types of elections, nine types of state action with respect to civil liberties, and finally the number of riots and the number of deaths from domestic political violence. While this is an impressive collection of variables, it is only a typology. No scaling system for the events analogous to those used in the international events data is developed, and, more importantly, only a small number of types of events are coded, thus limiting the utility of the data for the comprehensive study of political conflict in a society. The World Handbook scheme ignores, for example, statements, speeches, and petitions. Clearly, these events are relevant to the study of internal interactions (i.e., the opposition makes demands against the state, the state denounces the opposition, etc.). The World Handbook also fails to track cooperative interactions between groups embroiled in conflict. If we are interested in modeling conflict processes and testing those models, we need data that account for the full spectrum of relevant events.In addition to the type of event, the Handbook provides information on the actor, the target, and the issue that motivated the event. Taylor and Jodice clearly intended to make their data useful for testing hypotheses regarding the behavior of actors and their interactions, but their choice of actor and target designations leads one to believe that they were envisioning crossnational studies rather than timeseries case studies. For example, there are eight “actor groups” variables: government, political party, political group, military, clergy, intellectuals, workers, students, minorities, revolutionaries, and general population. While these categories are useful for broad comparisons across many cases, they are less useful for the analysis of specific countries. In most states, for instance, there are several political parties pursuing very different policies, but the Handbook coding scheme fails to distinguish among them. The IPI project is designed to overcome some of the limitations of the World Handbook scheme. First, IPI data are scaled according to intensity. Second, IPI provides codes for a broader range of political events, both verbal and substantive, both conflictual and cooperative. Third, IPI assigns unique codes for actors and targets which may be aggregated into broad categories for cross-national studies, but also can be used for detailed time series analysis of the interactions among particular groups. Fourth, IPI data serves to update and increase the temporal domain of the World Handbook data. Many of the World Handbook variables cover only the period 1968-1982. Because the IPI coding scheme remains compatible with World Handbook data, the two datasets together provide a consistent longer term time series.We note that the Handbook is not the only source for data on internal conflicts. Those familiar with the COPDAB project (Azar, 1982) recall that it includes data on conflict and cooperation within nation states. Unfortunately, the utility of the domestic COPDAB data is limited by the fact that it is aggregated at the national level. As a result, the data indicate an overall level of cooperation and conflict in a given nation state, allow no indication of the
extent to which the state or opposition groups are instigating the violence. It is the ability to distinguish among actors that drives the IPI project.The IPI coding scheme represents our attempt to mirror the successful data collection efforts at the international level using domestic actors as the unit of analysis. In the process, we have overcome most of the limitations of existing domestic political event datasets by 1) increasing the number of actor/target codes to allow users flexibility in choosing appropriate levels of aggregation, and 2) coding events with respect to intensity on ten point scales for cooperative and conflictual interactions.

1.3 Conceptualizing Conflict

Conflict has two crucial elements. First, conflict involves perceptions of incompatible interests. Groups in conflict believe that the realization of one party’s interests inhibits the realization of the other party’s interests. Second, conflict is reflected in actual behavior; we are not interested in “latent” or “potential” conflict. One or both parties take purposive action against other parties in an effort either to advance their own interests or inhibit the realization of the interests of others. In other words, conflict is visible in actions and events. As Tilly states, conflict tends to flow directly out of a population’s central political processes, instead of expressing diffuse strains and discontents within the population; . . . the specific claims and counterclaims being made on the existing government by various mobilized groups are more important than the general satisfaction or discontent of those groups, and . . . claims for established places within the structure of power are crucial (Tilly, 1985, p. 1).Thus, the actions undertaken by groups to advance their own interests at the expense of others or to inhibit the realization of the interests of another group are a valid measure of the level of political conflict within a society.

1.4 Conceptualizing Cooperation

Cooperation has historically been thought of as the absence of conflict. Our concept goes beyond this simple definition. We believe that cooperation, like conflict, involves intentional, purposeful action in pursuit of goals. Keohane’s distinction between cooperation and harmony is especially useful. Harmony refers to a situation in which actors’ policies (pursued in their own self-interest without regard for others) automatically facilitate the attainment of others’ goals. Cooperation, on the other hand, requires that the actions of separate individuals or organizations– which are not in preexistent harmony– be brought into conformity with one another through a process of negotiation which is often referred to as ‘policy coordination’ . . . . Cooperation occurs when actors adjust their behavior to the actual or anticipated preferences of others, through a process of policy coordination (Keohane, 1984, p. 51). Groups pursue their goals through both conflictual and cooperative actions. Understanding the dynamics of political interactions thus requires measuring both the conflictual and cooperative behavior of domestic actors. The identification of actors and the categorization of their behaviors lies at the heart of the IPI coding scheme.


2.1 Defining a Political Event

With respect to the conceptual definition of an event, this project adopts a modified version of the Azar (1975) definition. We are interested in cooperative and conflictual political events: A political event is an action taken by an actor at a given point in time to advance its political interests. Thus a political event involves (1) an actor, (2) a target, (3) a time period, (4) an action, and (5) a political interaction. By political, we mean
issues that involve the authority to make decisions concerning the extraction and distribution of social resources or values. We replace Azar’s concern about issues with the requirement that these events involve a political interaction. Operationally, we modify Gerner et al. (1992) and define a political event report as: A natural human language statement describing a political event, where the (sometimes implied) subject and object(s)/direct object(s) denote an actor/target and the verb denotes an action taken by the actor.Azar additionally specifies that events must be “distinct enough from the constant flow of ‘transactions’ . . . to stand out against this background as ‘reportable’ or ‘newsworthy'” (Azar, 1982, p. 3). Howell interprets this to mean that “Transactions are the normal activities between nation-states, while interactions are the extraordinary events that occur in irregular patterns” (1983, p. 150). We, too, are interested in extraordinary events (interactions), rather than normal activities. Thus, opinion polls, the imposition of austerity measures, standard labor-business negotiations, and other similar actions which represent the normal functioning of government or society are not codable under the IPI scheme. Reactions to any of these events, however, may be codable. Protests in response to austerity measures, for instance, are coded.In some instances, the line between normal activity and an extraordinary event is fine. Democratic elections are important and unusual events in some societies, but in established democracies they represent no more than the normal functioning of government. We code all activities surrounding the first two elections held in any state undergoing a transition to democracy with the exception of personal attacks against candidates. Grandstanding and content free statements (e.g., criticisms of an official’s family life, nonspecific campaign promises) are never codable in the IPI scheme. In established democracies, we do not code campaign rallies for individual candidates unless they become explicitly pro- or anti-government demonstrations, and we do not code the elections themselves. We do, however, continue to code all statements regarding policy. Similarly, it is sometimes difficult to isolate the circumstances under which conflict between governmental branches is extraordinary and should be coded and when disputes represent the normal policy making process. We code all conflict between governmental branches which relates to institutional questions. In other words, when conflict between the legislature and the executive relates to the definition of areas of jurisdiction and responsibility, we code the conflict. When the legislature and the executive dispute policy, however (for instance, the legislature passes laws which the executive does not favor), we do not code the activity. Policy statements by political actors, however, are codable. If the leader of a political party criticizes the executive’s anti-inflation plan and vows to defeat it in Congress, we code a conflictual statement from that party to the executive.

2.2 Actor and Target Assignment

The term actor carries with it a connotation of initiator. We wish to note explicitly that we do not intend that connotation in our use of the term. It is frequently impossible to ascertain who started a set of given interactions, and no attempt is made to do so. Rather, we begin by assuming that groups face options to act or not to act at any given point in time. If they act, they are coded as actors. If they are acted upon (i.e., the recipient or target of another party’s action), they are coded as targets. We have created (and attached) a list of generic actor-target codes to facilitate defining actors and targets. For each country we examine, the generic sheets are modified to reflect the specific circumstances and political context of that case (see section 3.11).

Consider some examples from The New York Times Index’s reporting of events in Colombia, 1965: Bandits attack peasant farms, Paime, Cundiamarca Dept., kill 5, wound others, Je 2, 5:5. Here, the actor is bandits and the target is peasants. Since there is no account of the peasant farmers responding to the attack with violence, they are not coded as actors with the bandits as targets. If such a report were noted, however, each party would be coded as an actor and as a target. That is, when two parties interact, both are assumed to have the option of acting: in this case, the peasants can respond to the attack, or they can run and take cover. Since the report does not explain that they attacked, we assume that they took cover. However, had the words “peasant farms” been “police station”, the same judgment would not have been made. Different parties to a conflict have different ‘base line’ responses. We adopt the simplifying rule that if a party to a conflict event is known to be armed, then we assume that, unless reported otherwise, the party will respond to an attack or ambush by firing upon its assailants. Since peasants are not generally known to be armed in Colombia, their `base line’ response is assumed to be flight. However, guerrillas, bandits, police, and the armed forces are assumed to have a ‘base line’ response of attack.

Thus, the following report is coded as explained below: Police kill bandit chief J B Ruiz Castro, Jl 7, 14:3. Here, the police and the bandits are coded as both actor and target (i.e., two events are coded whereas in the attack on the peasant farms only one event was coded). Since a bandit surrender would be unusual (and would likely not lead to the chief’s death, but rather to his capture), we invoke the assumption that the reporters would have noted it had the bandits surrendered. Certainly, this assumption will be poor in some cases. However, we contend that failure to invoke this assumption would lead to one-sided coding of two-sided events much more frequently than the assumption leads to two-sided coding of one-sided events (i.e., invoking the assumption reduces the noise in the data).

The first report above raises another question about the actors and the targets. We are interested in political conflict events. Are ‘bandits’ political actors? Unfortunately, this is a thorny issue because government actors who provide reporters with the information they use to file their reports have a significant interest in minimizing the political claims of their opponents. Thus, they sometimes refer to political opponents who have taken up arms as bandits. As coders, we face three choices: 1) we can code them as guerrillas, 2) we can ignore bandits as actors, 3) we can code them as bandits and allow users of the dataset to determine whether they want to include bandits in their study. We find the third option vastly superior to the alternatives.Sometimes, coding the target can represent a problem, as in the following example: Govt reports 11 dead in guerrilla attack; repts 14 others dead in Dec 31 attacks. From this report it is unclear who is the target and, as a consequence, how many events should be coded.  Unfortunately, this record is incomplete. We know that at least two events took place, that the actor in both cases was unspecified guerrillas, and that eleven and fourteen people died in either armed attacks or military clashes, depending on who the target was. Since it is better to err on the conservative side, we code it as an armed attack (604 on the conflict scale) on an unknown target (99) and indicate in our event record that we need more information. We hope that as we code additional sources, more facts will be reported which will allow us to fill in the missing information, although there will be a substantial number of events for which this is not the case. It is important to note that victims are not always targets. This distinction becomes important in coding terrorist or repressive events. The state assumes responsibility for maintaining public order. Terrorist events like bombings and hijackings challenge the ability of the state to fulfill this responsibility. While bystanders may be injured or killed and suffer as victims of such events, in most cases, we consider the target of the event to be the state. Thus, if a foreign embassy, or an electrical pylon, or a hotel is bombed, the target is the state. If stores are robbed during a riot, the target of the event remains the state, not business elites. The exception to this rule is terrorist events which are clearly aimed at disadvantaging other groups, for instance a bombing of a dissident group’s headquarters, or of the home of a trade union leader, or of a group of businesses owned by members of a particular ethnic or religious group. In these cases, the target of the event is the disadvantaged group.

The argument can be extended to repressive events. In many cases, the individual victim of repression is not the target of state repression. Rather, repression is used as a tool for intimidating the entire society rather than the individual. If hundreds are disappeared by the state, therefore, we code the target as ‘general population’ (09), regardless of the affiliations of the victims. Once again, however, if repressive actions are aimed at disadvantaging particular groups, those groups are considered targets. Thus, the murder by death squads of a single trade union leader is an action targeting labor unions. Occasionally, an action will be aimed at a social group, but will additionally challenge the ability of the state to maintain order. In this case, the event is coded with multiple targets. Consider the following paraphrased report from Reuters News Wire, November 21, 1986, on events in Mexico: During the parades, supporters of the PRI violently clashed with members of the PAN. Seventy were wounded as police used teargas to break up the rioting. In this case, the PRI is acting violently toward the PAN and the PAN is acting violently toward the PRI. Both the PAN and the PRI are also challenging state order. Thus, this report would result in six event records: two records representing the violence from each of the parties to the other (401 on the conflict scale), two representing the violence to the state from each of the parties (401 on the conflict scale), and two representing state violence to the parties (402 on the conflict scale).

2.3 Multiple Actors and Targets

There are a number of instances in which multiple actors or targets are identified in a given report. Consider the following fictitious example: In a show of solidarity, government employees, teachers, laborers and peasants marched on the capital today to protest recent government cut-backs in wages and prices.Here we have a single target (the state–10) but several actors: government employees (53), teachers (48), laborers (56) and peasants (50). Thus, the coder codes four conflictual events: a protest (301 on the conflict scale) for each of the four social actors against the state. In addition, the coder accounts for the cooperation between the four groups acting together. In this case, each of these four social groups is cooperating with other social groups, so the coder codes four 304’s on the cooperation scale, one for each social group, with the target of the cooperative action listed as ‘unspecified social actors’ (75).

2.4 Coding Multiple Events

Often multiple events appear in a single news report. Coders need to pay strict attention to this issue, as it accounted for the vast majority of the unmatched coding in the first round of coding for Moore’s “Violent Intranational Conflict Data Project: Phase I” (Moore and Lindstrom, 1994).

One example was covered in the second event noted above (section 2.2): a report of one group attacking another should be coded as two events if the group that is attacked is generally known to be armed, unless it is expressly reported that the other group did not respond in kind. A second example appears in section 2.3. When several groups act together, we code a separate event for each actor/target combination. A third issue concerns references to past events in a report that is primarily concerned with a present event. A previous report (section 2.2) provides an obvious example; it is noted that there were “25 others dead in Dec 31 attacks”. The coder must be sure to enter this event as well.

A less obvious example is this report from Colombia, 1956 (TNYTI): dispatch from Quito repts 80-200 villagers slain by Army troops in May near Chaparal, Tolima Prov, in reprisal for death of a few soldiers ambushed by guerrillas;…. Je 12,1:5. In this example there are three events: 1) the army’s slaying of peasants in May, 2) guerrillas ambushing the army at an unspecified date, and 3) the army attacking guerrillas on that same unspecified date. We discuss problems associated with dating events below (section 2.5).Occasionally, journalists offer summary reports of events which have occurred during a recent time period. For instance, a report might state, “fifty bombings have occurred in the last month in Chile”. This report is coded as fifty separate events, albeit with much missing information. The coder checks the ‘Needs More Information’ field for each of these fifty event reports with the hope that more information can be located in other sources.A fourth issue concerns events that endure for more than twenty-four hours. General strikes, riots, etc. often take place for several days and even months. In this project we note that all human beings live on a twenty-four hour schedule because our bodies simply require that we get some sleep, nourish ourselves, and attend to our bodily functions. Thus, no human beings sustain strike activity, rioting, or other events for days on end without a break. Since the rising and setting of the sun provide a convenient marker, we set the rule that no single event can last more than twenty-four hours. Thus a twelve day strike is coded as twelve separate events.

As an example, consider the following paraphrased report from Lebanon, 1974 (TNYTI): Police storm campus, kill 2, terminate 12 day strike. It was the first mention of the student strike, and hence the coder must code twelve strike events by the students against the state (302, conflict scale). In addition, the coder must code one ‘Police violence in response to demonstration with deaths’ event (504, conflict scale) with students as the target. Finally, the coder will note two deaths from political violence (see section 2.10). Any time a long term political standoff results in daily interactions, an event is coded for each day of its duration. Thus, if a group occupies a building and holds hostages for five days while negotiating with the government, we code a violent political act (505 on the conflict scale) for each of the five days. Similarly, battles, strikes, protests, and riots are coded for each day of their duration. There are other situations which can not be assumed to result in daily interactions, however, and these events are coded only as they begin, end, and experience reported changes.

Examples include most repressive actions (for instance, the imposition of states of emergency, the closing of universities, and the imposition of censorship), kidnappings, and occupation of territory by the opposition. A difficult situation arises, however, from the following type of report (Philippines, 1975, TNYTI), which is typical of statements where body counts of deaths from violent political conflict are announced by state officials: announcement coincides with new fighting with 15 Govt soldiers reptd killed in last 10 days with guerrillas in provinces of Sulu, Cotabalto and Zamboanga; at least 5 rebels are known dead and mil announced that some 500 civilians were rescued from Zamboanga del Sur where they had been held hostage; map (M), Ja 11, 3:1.

We know that more than one clash occurred between the state and guerrillas during this ten day period, but we do not know how many. In the absence of better information, we assume that fighting occurred on half of the ten days in each of the three locations (total thirty military clashes, fifteen with the state as actor and fifteen with the guerrillas as actor). The coder codes military clashes for five days and checks ‘Needs More Information’ on the coding sheet in the hope that information from another source might clarify the report. The twenty-four hour rule is useful because it enables us to build duration into the coding without creating a separate scale (e.g., Taylor and Jodice use a scale). Some might argue that this scheme is biased toward long events, that because it counts each day as a new event, events with long duration overwhelm other events of shorter duration. This argument is predicated on the idea that there is a decreasing marginal impact by day of an event as it persists over time.

Our coding scheme does not preclude end users from taking such an argument into account. Anyone using the data may transform the data on an event-by-event basis by identifying those events that occurred over multiple days, taking the natural logarithm of all events that persist beyond some number of days, and using that log-transformed figure in his or her analysis. This scheme enables the end-user to make those decisions. Finally, political conflict events have a spatial as well as a temporal dimension. Just as we use twenty-four hours as the natural temporal unit, we use the village/town/city as the natural spatial dimension. If the same event occurs simultaneously in one village, four towns and two cities, then we code seven events (for that day). To summarize, events are identified with regard to actors and targets, time, and space. If two groups protest in three cities for five days, a total of thirty (two times three times five) events are coded. Occasionally, however, a report will seem to indicate that several codable events have occurred when really one encompasses the others.

Consider this example from the Latin America Weekly Report, March 26, 1982: At least two soldiers and six M-19 guerrillas were killed in a series of clashes in the southern departments of Cacqueta, Huila and Cauca last week. Guerrillas blew up a bridge and high-tension pylons on the Florencia-Neiva road, and ambushed military vehicles. Fighting was continuing this week, as troops of the Juanambu battalion moved in (p. 12).In this case, we code twenty-four military clashes (704) because we have reports of two actors fighting in three locations over four days (We invoke the assumption that fighting occurred on half of the days mentioned in the report–last week continuing into this week = eight). We do not, however, code bombings and ambushes as well. The military clash encompasses these activities. Similarly, if a group protests in front of the seat of government shouting anti-government slogans, we code the protest (301), but we do not code additional verbal conflict events (100 and 200 level) (see section 4.4).

2.5 Dating the Events

The events are assigned the date on which they are reported to have occurred, not the day on which they are reported. In some cases, a precise date is impossible to assign. When this happens, we record as much information about the date as we can. For instance, if we do not know exactly which day the event occurred, but we do know which week, we record the week. In this case, the entry in the “Event Date” field appears in the following form: JAN w1 1989. We follow the GEDS (Davies, 1993) rules for assigning weeks. They are as follows: week 1: 1-7 week 2: 8-14 week 3: 15-22 week 4: 23-end.

When we cannot even determine the week of the event, we simply assign the missing value score ’99’ for the day variable and then seek to obtain a precise date from a different source. (The coder should check ‘Needs More Information’ on the coding sheet.) In most cases, the month in which the event occurred will be apparent. If the month can not be reliably isolated, then we record the quarter using the following format: Q1 99 1989. quarter 1: January-March quarter 2: April-June quarter 3: July-September quarter 4: October December

While missing data on the ‘day’ variable do not present a problem for people who aggregate the data at monthly, quarterly, or annual units of time, we wish to avoid using the missing data value ’99’ in the ‘month’ variable whenever possible.The “dispatch from Quito” example above (section 2.4) highlights a dating problem where it is necessary to invoke some assumptions. We assume that parties respond quickly to wrong-doing so that, given no other information, when a report notes that a party is responding to a previous event, we will assume that event occurred within the past five days. Thus, if the reported event occurred on or after the fifth day of the month, the coder dates the event to which it is a response as occurring during the same month, with the day coded as missing (i.e., 99). If the reported event occurred prior to the 5th day of the month (i.e., 1st-4th), the coder dates the event to which it is a response as occurring during the previous month, again with the day coded as missing (i.e., 99). In these cases, the ‘Needs More Information’ box is checked in the hope that the missing date will be clarified by secondary sources.

2.6 Controversial Reports

Sometimes it is difficult to tell whether an event actually happened. Consider the following report from Latin America Weekly Report, February 26, 1982: The government claims that six more members of M-19 gave themselves up to the southern command of the army, in Florencia, Cacqueta, on 22 February, three days after the publication of the latest official amnesty offer (p. 11). By using the word “claims” the source suggests some doubt as to whether these surrenders actually occurred. We elect to code all such reports as actual events except where the source explicitly notes controversy over the event’s occurrence. We code this report as a surrender (see section 5.6 for further information about coding surrenders). Had the report continued with the statement, “Leaders of the M-19 deny that these surrenders occurred and continue to reject the amnesty offer”, however, we would not code a surrender. Instead we would code a verbal conflict event (101) from the state to the M-19, and another from the M-19 to the state.

2.7 Distinguishing Domestic Events from International Events

Some events which may initially appear to be international events are codable as domestic political conflict or cooperation in the IPI coding scheme. One example was given above in section 2.2. When a foreign embassy is bombed, the event can be coded as domestic political conflict between the actor and the state, which is responsible for maintaining order in society. In addition, when domestic groups protest against the influence of a foreign power with whom the state is allied or against alliance with a foreign power, the event can be recorded as conflict between the domestic group and the national government.

Two examples of such activities are given in the GEDS (Global Events Data System) Coders’ Manual (Davies, 1993): A militant Islamic group demonstrated in Cairo for an end to the alliance with US forces in Saudi Arabia (p. 27). Hundreds of Filipino students demonstrate outside the US Embassy in Manila for the withdrawal of American bases (p. 28).In both of these instances, the events can be coded as protests (301, conflict scale) with a social group as actor and the state (10) as target. Similarly, when members of a polity storm a foreign embassy requesting asylum, the event can be coded as a protest (301 on the conflict scale) against the national government. Events which happen outside the borders of the state in question are occasionally codable in the IPI system if they are explicitly related to a political conflict within the state. When a dissident organization is working to unseat the government of a state from the territory of a neighboring state, we do code interactions between the dissident group and the government it is opposing.

For example, if it is reported that Colombian drug lords have assassinated the Colombian ambassador to Hungary in Budapest, the assassination is codable as a domestic political event in Colombia. When a Pakistani dissident group hijacks a plane in Pakistan and flies to Syria where a ten day standoff with the Pakistani government ensues, we code domestic political conflict for each day of the standoff. When the Indonesian government raids an Indonesian dissident group’s settlement in Papua New Guinea in response to the group’s attacks in Indonesia, this too is codable under the IPI system.There are also cases, however, when political conflict events which occur in the territory of the state being coded are not domestic political events relevant for that state. For example, if Afghan students studying in India protest in front of the Afghan embassy in India to show their objection to the government of Afghanistan, this is not codable as an IPI event for India. It is codable as an IPI event for Afghanistan.

2.8 Sources for Events Data

In order to improve the density and quality of our data collection effort, numerous local, regional and international news sources will be coded during the project. In the early stages of the project we are concentrating on Reuters News Wire. As the scope of the project expands, we anticipate adding additional local and regional sources. Current source codes are as follows: 1. The New York Times Index 2. Reuters News Wire 3. Latin America Weekly Report

2.9 Ambiguous Values

Consider the following fictitious report: State authorities launched a nationwide sweep yesterday, arresting 90 members of a well known dissident group. The operation required police to coordinate their efforts in the cities of X, Y, and Z as well as several municipalities.We know in this instance that arrests should be coded for three cities and for “several” municipalities. News reports are full of terms like ‘several’, ‘scores’, ‘hundreds’, and other vague terms. In this project we arbitrarily (but consistently) assign the following values to these terms: couple 2 few; some 3 several 5 many; numerous 15 dozens 36 scores 60 hundreds 250 thousands 2500. Thus, in the above example, eight events are coded because we have assigned the word ‘several’ a value of five.

2.10 The Deaths Variable

The variable “Deaths from Political Violence” is included as a measure of the intensity of political violence within societies. Gurr (1970) argues that ‘Deaths’ is the best estimate of the magnitude of political violence, and numerous studies of political stability and violence have employed this measure (see, Muller and Seligson 1987; Muller, Seligson, Fu, and Midlarsky 1989).We adopt the definition developed in the World Handbook of Political and Social Indicators which points out that “Unlike the other indicators, the number of deaths from political violence is not an event variable, but a body count.” (Taylor and Hudson, 1975, p. 18). Many of the events which are coded in the IPI system may involve deaths. Certainly this would include armed attacks, military clashes, assassinations, executions, and death squad activity. We will also keep track of victims of riots, bombings, disappearances, and others whose deaths are directly attributable to a domestic political event.One of the problems with the ‘deaths’ variable is illustrated by TNYTI report from the Philippines reproduced above in section 2.4: summary scores will be given about the number of dead between two given dates. A favorite approach of states (and, hence, journalists) is to note the number of casualties since some defining event which marks a renewed round of violence (e.g., 211 people have been killed since violence erupted six weeks ago).

To address this problem we have included a ‘Since When’ field in our coding sheet which coders use to indicate the length of time over which the death count is aggregated. This information will be used at a later date to estimate values for smaller time periods. The ‘Since When’ field will be left blank when the number of deaths reported are not a cumulative score since a given date. Cumulative death scores, however, should not be reported as part of an event record. Instead, a report of a cumulative death count should employ only the following four fields on the coding sheet: ‘Deaths’, ‘Since When’, ‘Reference’, and ‘Description of Event’, with the possible inclusion of ‘Needs More Information’ and ‘Comment’.

Another difficulty in keeping track of the number of deaths from political violence is the fact that we often receive conflicting reports of the number of deaths. New information becomes available as time passes. The number of deaths reported two or three days after a riot or bombing may be significantly higher than the count which appeared in the first report of the event immediately following its occurrence. In this case, we update the death count for the event to the final figure after all casualties are reported. In some instances the updating of the number of deaths may alter the appropriate code for the event itself. A riot with ten deaths or fewer, for instance, is coded as a 504 on the IPI conflict scale. A riot with eleven to ninety-nine deaths, however, is a 602. Were the death count for a riot to rise from an initial report of eight to a final report of fifteen, the coder should not only correct the entry in the ‘deaths’ field, but change the coding of the original event as well.

Finally, it is also frequently the case that parties to a conflict disagree as to the number of deaths attributable to a conflictual event. We hope that in these instances our sources will provide independent counts, or at least suggest which claim is more reliable. If we have no way of determining from the event report which casualty figure is more reliable, we record the more conservative number in the ‘Deaths’ field, report the discrepancy in the ‘Description of Event’ field, and check the ‘Needs More Information’ field in the hope that we will be able to clarify the report from other sources.

2.11 The Coding Sheet

A copy of the IPI Coding Sheet can be can be obtained from the authors. It is simply a hard-copy of a blank spreadsheet created with Quattro Pro. Many of the fields are self-explanatory. The exceptions are discussed in more detail below.

First, there is a ‘Multi’ field. As discussed in section 2.4, the IPI coding scheme considers a number of episodes to contain multiple events. Long term strikes, protests which spread to several cities or involve many groups, and mass arrests all result in multiple IPI event records, as we code a separate event for each day, each group, and each location involved in a particular action. Because some end users of the data may choose to manipulate these episodes to lessen their impact, we identify them with an indication in the ‘Multi’ field. The following are the appropriate entries in the ‘Multi’ field: 0 Not a Multiple Event 1 Multiple Days (e.g., Teachers strike for 100 days) 2 Multiple Groups (e.g., Students, workers, and peasants marched) 3 Multiple Locations (e.g., PRD led protests took place in five cities) 4 Multiple Numbers (e.g., Over 500 were arrested) 5 More than one of the above codes applies (e.g., Students, workers, and peasants marched for five days)

Second, there is a ‘Since When’ field. As discussed in section 2.10, ‘Since When’ is used primarily in conjunction with the ‘Deaths’ variable. Journalists frequently provide summary reports of the number of people killed since a given date. To keep a running tally, it is important to keep track of the time period covered by the summary report. A date in YRMODA format (i.e., year, month, day) should appear in the ‘Since When’ field. In general, the ‘Since When’ field is not employed as part of an actual event record. Instead, it is used to keep track of useful and necessary information reported in our sources but not connected to identifiable events.

Third, there is a ‘Reference’ field. This field enables the coder to document the Month, Day, Year, and Page Number of the report. When working with Reuters news reports online, coders should copy the proper Reuters article reference (e.g., rl040586.138) into the ‘Reference’ field on the IPI coding sheet.

Fourth, there is a ‘Description of Event’ field which is employed to summarize the event briefly. Because we employ multiple sources, we must be able to distinguish among multiple similar events. The ‘Description of Event’ category is used for this purpose. This field also facilitates editing of the IPI data. Should recoding of a particular class of events be desirable or necessary in the future, the ‘Description of Event’ field makes this possible. Coders should provide the most specific information given in the news source regarding the actor, target, date, and location of the event in this field. Careful use of the ‘Description of Event’ field will allow end users to create additional distinctions in the data to further examine the political dynamics of a particular case.

Fifth, we employ a ‘Needs More Information’ field. This field is particularly useful for vague reports like “sustained guerrilla activity was reported yesterday” or “fifty bombings have occurred in the last month”. It should be checked any time the coder thinks that there is missing information because the report is incomplete. There is no need to check ‘Needs More Information’, however, if the nature of the event implies that the information does not exist. For instance, if the ‘actor’ is missing in a bombing report, then there is no reason to check ‘Needs More Information’ because no source will authoritatively provide us with the missing information. The utility of having this field is that it draws our attention to search other sources for information on a given event and should help us to minimize doublecoding. Thus, ‘Needs More Information’ should be checked any time that the coder thinks a record is incomplete due to problems with the source rather than the nature of the event. ‘Needs More Information’ should be left blank when it is not used, and have a Y in it when it is needed.

The ‘Needs More Information’ field is also useful when events are reported as likely to happen, but the source is mute as to whether they actually happened. For example, a source may report that military officers have been sentenced to death and never report that the execution actually occurred. In this case, we code the conviction and check the ‘Needs More Information’ field, making a note in the ‘Comment’ field to look for a report of the execution in another source. The ‘Comment’ field should be used in conjunction with the ‘Needs More Information’ field. Whenever ‘Needs More Information’ is checked, an entry in the ‘Comment’ field should describe what information is needed. In addition, coders may use the ‘Comment’ field to direct concerns, questions, or further explanations to the editor. Any uncertainty about the coding of an event report should be indicated in the ‘Comment’ field for future review.On the coding sheet, coders should leave blank all fields which are not relevant for a particular event. The one exception to this rule is the case of the deaths variable. When no deaths occur, a zero should appear in the field.

2.12 Managing Computer Files

Data management is a crucial aspect of the IPI project. Each coder should maintain at least two copies of all of his or her work, one on a computer hard drive or server, and at least one on floppy disk. Each country/year should occupy a separate quattro pro file. The protocol for naming IPI computer files is as follows: (1) three alphabetical digits for the country name; (2) two numeric digits for the year; (3) three alphabetical digits for domestic.

The following are the three digit country codes for our initial cases:
























Thus, the IPI coding for Nigeria 1985 would be located in a file called NIG85DOM.WB1. The coding for Argentina 1982 would be named ARG82DOM.WB1. Proper naming of files is essential to allow us to distinguish between cases, and also between coded and edited data. Edited data is saved with a different nomenclature.


3.1 Selecting Actor and Target CodesOne of the primary goals of the IPI project is to create data which allows analysts to examine the behavior of particular societal groups and their interactions with other social groups and the state. As a result, the IPI coding scheme provides a large number of actor and target codes so that groups can be identified and tracked with a greater degree of precision than has been previously possible. Because of this increased precision and the plethora of available actor and target codes, it is often incumbent upon coders to make difficult decisions. Many actors and targets belong to more than one category. In this case, the coder must determine which group the actor or target is representing in the event in question. For instance, if a person is arrested who is both a labor union leader and a member of the communist party, the coder must determine if this individual is being arrested because he or she is a union leader or because he or she is a communist party member. We hope that this information can be gleaned either from the report itself or from the context of other reports. Similarly, when a former President makes political statements or undertakes political actions, the coder must determine if the ex-President is acting as a former leader or as a member of his or her political party.

It is also the responsibility of coders to translate journalistically assigned group identities to the appropriate IPI actor or target codes. The word “farmers”, for instance, could conceivably be most accurately captured by three different IPI codes. Farmers may be peasants (50); they may be landed elites (65); or they may be a business sector (52). The context of the report should suggest which code is most appropriate in a given instance. In all cases, coders must report the precise description of the actor and the target as given in the news source in the ‘Description of Event’ field, including proper names when appropriate. This allows end users of the data to alter the coding of specific groups with relative ease, or to create new actor and target groups.

The IPI Actor and Target Codes are listed in the appendix. The scheme is adopted from the system developed for identifying sub-national actors in Ted Robert Gurr’s “Minorities at Risk Project” (1993) and John Davies’ GEDS project (1993). We use the standard ICPSR country code to identify the state and append a two or three digit code to identify specific actors. Coders then assign specific codes for major ethnic populations, dissident organizations, and major political parties that appear in that case (e.g., the FARC in Colombia; the Radicals in Argentina) working within the framework established below (see section 3.11). In the following sections we provide a more detailed discussion of some of the generic and case specific actor and target codes. 3.2 The State (10-24)

In the IPI system, the state is disaggregated into the following governmental actors: (10.1) National Executive, (10.2) Head of State (if distinct from the national executive) (11) High level officials (Cabinet members), (12) Elected representatives (legislature), (13) the Judicial Branch, (14) Regional/state governments, (15) City/village governments, (16) Military and armed forces, (17) Paramilitary forces (i.e., death squads), (18) Police forces, (19) Deposed leaders, (20) Former leaders, (21) State enterprises, (22) Dissident military faction, (23) Former military leaders, (24) Other. Government workers are treated separately as social actors. This coding scheme allows us the flexibility to record conflict and cooperation within the state apparatus. We can code, for instance, criticism of the executive in a democratic state by the ruling party, or a purge of the army by a ruling military junta. If there is a coup attempt by junior officers, the event can be coded with actor as military/armed forces (16) and the target as national executive (10.1). Further if the executive purges his or her cabinet or junta, the actor is 10.1 and the target is 11. When students clash with the police, the actors and targets are 49 and 18.Actor and Target Code 10 represents the State as a whole. When dissident groups, social actors, or ethnic populations conflict or cooperate with the government, 10 is usually the appropriate actor/target code to represent the state. Only when the actions of a societal group are focused specifically on particular governmental actors (e.g., a city government (15), the military (16), or the legislature (12)), should coders employ the more specific codes. In addition, when actions of the state toward the population or particular groups within it are statements or acts of overall government policy, 10 is the appropriate actor code. For instance, if troops are placed in an area of heavy guerrilla activity, this is an act representing mobilization of resources by the state against the guerrillas (actor code 10, target code 45, conflict code 304). If the President signs a truce with a guerrilla group, he or she is representing the State, and the cooperation occurs between 10 and the guerrillas. If, however, the Minister of the Interior expresses an opinion that increased action should be taken against guerrillas, this would be a threat from 11 (high level officials) to 45, as it is unclear that this is the policy of the government as a whole.

In states with coalition governments, we additionally seek to distinguish between intra-coalitional conflict and cooperation and extra-coalitional conflict and cooperation. In these states (for instance, Belgium), we use actor and target code 11 (high level officials, cabinet members) to represent the ruling coalition. Thus, if leaders of the coalition parties threaten the national executive, the actor code is 11 and the target code is 10.1.

3.3 Mass/Elite Distinction

Just as we track conflict and cooperation within the government, we also make it possible to track interactions within certain societal groups by distinguishing between mass and elite actors and targets. We define elites as those individuals who control significant social resources. Whenever actions are either directed at or undertaken by only a portion of a group (i.e., only the leaders or only the masses), we append a decimal to the group code to reflect the more specific nature of the actor or target group as indicated on the Actor/Target Scale. This distinction between mass and elite actors allows users to investigate differences in the dynamics of intra-elite and mass interactions and to capture conflict between masses and elites of the same group.

3.4 Unspecified Actors and Targets (9, 45, 75, 89, 99)

There are a number of different codes for unspecified actors and targets. The most general code is 09, which represents the general population. Only in very rare cases should 09 serve as an actor. Most of the time, the code for general population is only used as a target; it is a group which is acted upon, usually by the state as it imposes or relaxes general repressive measures. The more common codes for unspecified actors and targets are unspecified guerrillas (45), unspecified social actors (75), and unspecified political parties (89). The code for unspecified guerrillas should be employed either when the report defines the actor or target as guerrillas but is not more specific, or when the action undertaken suggests that the actor was clearly a dissident group which has adopted armed struggle as its mode of operation. For instance, if unknown assailants ambush soldiers and a group of guerrillas is active in the case being coded, we can make the assumption that a violent dissident group was the actor and we can code the actor as 45.

The code for unspecified social actors (75) is commonly used in cases of reports of demonstrations, mass arrests, and other such actions which affect large numbers of people who may not share identifiable characteristics other than their common action. It is also used to represent conglomerations of groups. For instance, as discussed in section 2.3, when several groups coordinate their activities we code cooperation from each group to all of the other groups using the unspecified social actors code (75).The code for unspecified political parties (89) should be employed when neither the party nor its political leanings can be identified. For instance, if a known political party is reported to have formed an alliance with five other parties which are not named in the report, we code cooperation between the known party and unspecified political parties (89). Coders should note, however, that we do provide codes for unspecified leftists (87), and unspecified rightists (88), so when the political leanings of a party can be determined, these are more appropriate codes than is 89. In addition, we employ the various unspecified codes when a party, social group, or dissident group is identified in a report, but does not qualify for an IPI case specific code (see section 3.11). Dissident groups which exist for less than a year, for instance, should be coded as unspecified guerrillas (45). Similarly, social actors which do not appear on the generic list should be coded as 75.

Actor and Target code 99 indicates missing information. This code should not be used to represent unspecified actors and targets, but rather unknown actors and targets. It will be common, for instance, for the actor in a bombing to be coded as 99.

3.5 Business and Labor (51, 52, 64, 54, 55, 56)

There are three codes in the IPI coding scheme to represent business, and three to represent labor. The business codes are distinguished from one another as follows. Actor/Target code 51 represents organized business groups. The code for business sector (52) is used to represent a group of business people employed in the same sector working together. If a report states, for instance, that “bankers declared support for the opposition economic program and criticized government handling of the economy”, the bankers would be coded as a business sector (52). 64 is the default business code. If a report simply states, “labor, business, and the government met today to discuss inflation”, we code business as 64. The three labor/worker codes are distinguished as follows. Labor confederations are coded as 54. Labor confederations are large organizations which bring together workers from many regions and many occupations into one umbrella organization. An example is the CGT in Argentina. Code 55 is used to represent individual labor unions. Code 56 is employed to represent unspecified or unaffiliated workers. If it is unclear than an event involving workers was organized by a labor union (e.g., “workers and students marched on the capital to protest price increases”), we use the code 56. Code 53, however, represents government workers. Thus, when the workers are exclusively state employees, 53 is a more appropriate code for unorganized workers than is 56.

3.6 Multinational Groups and Foreign Nationals

When multinational groups and foreigners become involved in the political activities of a state from bases within the state, they are codable as actors and targets under the IPI scheme. It is important to recognize, however, that only the local actions of these groups are domestic political events. Foreign media (47), for instance, can act and be acted upon in a domestic political context. The Roman Catholic Church (57), human rights groups (62), and public interest groups following such issues as environmental policy (63), are active in many states and can be coded as domestic actors in all of those states. The international actions of these groups, however, are not codable as domestic events. A pronouncement by the Pope can not be a domestic political event, but a pronouncement by a local Bishop can. Reports by Amnesty International issued in Washington, DC, are not domestic political events, but statements by the local Amnesty office are.The code for foreign nationals (70) should be employed only when an individual is acting or being acted upon because he or she is a foreign national. For instance, if Chileans living in Argentina protest Argentine policy toward the Beagle Islands, we code a domestic political event with foreign nationals (70) as the actor. But if a member of a labor union who happens to be a foreign national launches a protest, we code the actor as labor union (55), not foreign nationals.3.7 Leftists and Rightists (73, 74, 87, 88)

The IPI system provides Actor and Target codes to represent leftists and rightists both in the social actors category and in the political parties category. As mentioned above in section 3.4, codes 87 and 88 represent unspecified political parties with left and right leanings respectively. Often, however, actors and targets will be identified in news reports as “leftists” or “rightists” when they are not representing organized political parties. In these cases, we employ the codes 73 and 74. Consider the following statement: The President blamed leftists for inciting the student riots. In this case, we code the actor as 10.1 and the target of the negative statement (101, conflict scale) as 73. Similarly, a report like “Right wing sympathizers protested in front of the Ministry of Defense to express their anger at proposed defense cuts,” would be coded with 74 as the actor and 10 as the target. 3.8 Religious Groups (57, 58, 59, 60, 71)

There are a number of IPI actor and target codes which represent various religious groups. A religious group code can only be employed as the actor in an event when the action is taken by official leaders or members of the faith acting in a manner commensurate with a sanctioned position of the religion. Thus, if members of the Catholic Church act violently toward abortion providers, we can not code 57 as the actor because the Catholic Church does not sanction violent action. These actors are not a Catholic social group, but a political dissident group. Similarly, if Islamic extremists blow up a building or take hostages, we code their actions as the actions of a political dissident group.

Any time members of a faith are targeted because of their religious faith, religious group codes can be employed as targets. If Christians are targeted for repression because they are Christian, we can code 58 as the target. In many cases, however, religion and ethnicity will be correlated with one another. When a group which shares a religion is truly targeted for its ethnicity, the ethnic group code is more appropriate. Whenever there is a question as to whether a group should be defined by its ethnicity or its religion, the default should be ethnicity. In addition, conflict between factions within a given religious faith (e.g., Sunni v. Shiite) is captured through the use of ethnic group codes. Thus, when a particular sect takes action, the appropriate code is often an ethnic group code.

3.9 Socially Influential Individuals (66)

Actor and target code 66 captures the actions of socially influential individuals, people who command attention due to the respect, fame, and popularity they have achieved in their chosen fields. Certainly, if leading economists comment on an economic plan, the actor code might be 66. Also codable in this field, however, would be famous musicians, novelists, actors, and sports players whose political activities receive special attention. For instance, in Argentina in 1986, immediately before the World Cup, the manager of the Argentine national soccer team makes political statements. Because we believe his statements are socially influential because of his high stature in his field, the actor is coded as 66. Similarly, we often code the actions of family members of high level government officials (e.g., the wife or brother of the President) with actor code 66.3.10 Supranational Organizations (95)

IPI Actor and Target Code 95 represents Supranational Organizations. The only cases for which this code is currently relevant are those which are members of the European Community. Interactions between the European states and the European Community are international events and are not codable under the IPI system. Actions of domestic political groups toward the EC, however, are codable as domestic political events. Thus, if Belgian citizens protest against EC nuclear policy in Brussels, the actor is 75 and the target is 95.3.11 Case Specific CodesIn addition to the generic actor and target codes which are applicable to all states, the IPI coding scheme tracks the most important political groups specific to each of our cases. First, we assign specific codes to major ethnic populations within each of our cases (codes 01-08). We begin with the groups identified by Ted Robert Gurr’s “Minorities at Risk Project” (1993) and add additional groups as necessary. Second, we assign specific codes to dissident groups (codes 25-44) which remain politically active in a state for a minimum of one year. Third, we designate codes for the major political parties (76-85) in each state. We define major parties as those which consistently receive more than five percent of the vote nationally, or which frequently receive a winning plurality of the vote regionally. Currently assigned case specific codes can be located in the appendix. In addition, brief descriptions of coding decisions specific to each case are available from the Principal Investigators upon request.

3.12 Distinguishing Among Ethnic Groups, Social Groups, and Dissident

It is important for coders to distinguish between political dissident groups and social or ethnic groups. Some of these distinctions (between religious groups and ethnic groups and between religious groups and dissident groups) were discussed above in section 3.8. Similar issues arise in distinguishing the instances in which it is appropriate to employ an ethnic group code and when it is appropriate to employ a dissident group code. Ethnic groups are not organized entities; they are merely collectivities which share common characteristics. While members of such groups may be targeted due to these characteristics, it is rare that action erupts spontaneously from these groups. As a result, the IPI ethnic group codes are more often used to identify targets than actors. When groups or organizations emerge to pursue the interests of ethnic groups
politically, these groups are functioning as political dissident groups. Separatist movements, for instance, are political dissident groups, not ethnic groups. Thus, in the IPI system, we code the group or organization which instigates action as the actor, not the broader social grouping into which members of the group may fall.

3.13 Coding Factions and Coalitions

As factions and splinter groups begin to act independently from their parent organization, we create new codes for them by appending a decimal to the code of their parent organization. For example, the FARC is actor/target code 26 in Colombia. FARC factions are coded as 26.1, 26.2, etc. If a faction or splinter group survives as an independent group for a year or more, it receives its own actor/target code. For example, in Colombia, in 1984, a faction of the FARC rejected a cease fire agreement and chose to continue the armed struggle. This group continued to act independently for more than a year under a new name: the Ricardo Franco Front. The Ricardo Franco Front is thus assigned its own actor/target code in Colombia (28). Similarly, when coalitions form which act consistently for their member groups for a year or more, they too receive independent actor/target codes. For instance, the Argentine multipartidaria, formed in 1981 as an alliance of five political parties (the Peronists, the Radicals, the Christian Democrats, the Desarrollistas, sand the Partido Intransigente) is actor/target 80. When questions arise, coders should err in the direction of proliferating the number of actors. Users of the data can always aggregate actors as they see fit, but they cannot disaggregate beyond what the data provide. It is important to note as well that when members of any dissident group, political party, or social group renounce their ties to that group, they can no longer be coded as speaking for the group in question. For example, if a few members of a dissident group surrender, renounce their ties to the group, and inform the state about the group’s activities, we code these former members as actors in cooperative events with the government and conflictual events with the dissident group. They are no longer coded as members of the dissident group, however. Instead, we code their actions as the actions of unspecified guerrillas (45). Similarly, if a few members of a political party resign from the party and criticize its activities, we do not code their actions as emanating from that party itself. Instead, the appropriate actor code is unspecified parties (89) or unspecified leftists or rightists (87 or 88).


4.1 Introduction to the Conflict Scale

The IPI conflict scale is designed to facilitate the measurement of the volume and intensity of political conflict within a domestic polity. Events are assigned values on a one to ten scale according to their severity. In accordance with accepted scholarly practice, weights capturing the relative intensity of each category of political events will be developed by polling a panel of experts (see, for example, Goldstein 1991, and Moore and Lindstrom 1994). These weights can then be used to develop indices of the intensity of conflict and cooperation.In general, 100 and 200 level events represent verbal conflict and 300-1000 level events are conflictual actions. 100 level verbal conflict implies no threat of action, whereas 200 level verbal conflict involves an explicit or an implicit threat to the target. 300 level events are always nonviolent, whereas actions starting at the 400 level may involve violence. In the following sections we discuss further rules for coding particular conflictual events. Coders should additionally refer to the scale itself in the appendix for helpful examples.

4.2 Press Criticism of Government (101)

We code press criticism of the government only when it is reported by a secondary source, not when it appears in the source we are coding. In other words, if we are coding a local publication, we do not code a verbal conflict event from the media to the state for every critical editorial which appears. If another source writes a news report about a critical local editorial, however, it is appropriate to code a verbal conflict event (101) with the media (46) as actor and the state as target. A similar rule applies to letters to the editor. They may be coded as verbal conflict, but only when they are discussed in a secondary source.

4.3 Rumors of Threatening Action (201)

When our sources report rumors of threatening action, we code them as actual threats (201). We are interested in measuring the perceived threat by the target, and we believe that reports of rumors of threatening action are perceived by the target as threatening. When a rumor of a potential coup d’etat surfaces, for instance, we consider it to be as threatening to the executive as an explicit threat would be.

4.4 Strikes, Protests & Riots (301, 302, 401, 404, 504, 602, or 701)

These three types of events raise a number of coding ambiguities that must be addressed. One difficulty concerns long-term strikes or protests (i.e., those that persist for substantial periods of time such as the anti Marcos demonstrations in the Philippines prior to the `People Power Revolution’ or the pro-Democracy demonstrations that preceded the Tianenmen Square crackdown in China). The problem is that once these events become commonplace, they are no longer treated as newsworthy by the press unless something unusual (like violence) occurs. This can lead to reports like the following fictitious report: Violence broke out in the 2 week old stand-off between low-paid bureaucrats in the capital and the armed forces. Three people were reported injured in the clash.Sometimes, this will be the first and only report of the event. In these cases, we do not want to code fourteen days of violent strike activity (401s and 402s on the conflict scale), but rather fourteen 302s and one 402 (note that since the report does not indicate that the bureaucrats were violent, we assume that only the police were). It is also important to note that in cases where a peaceful demonstration or strike erupts in violence, we do not code both a demonstration and a riot for the same group on the same day. We count the entire day at the highest level of conflict reached.

A second issue with these types of events concerns the fact that we distinguish among major and minor types. These are generally distinguished by the damage–typically to human beings–done (e.g., riots) or the number of cities that are involved (e.g., strikes). These distinctions are noted in the scales below. The distinction between local and national strikes (i.e. 302 v. 404, conflict scale), for instance, is whether a majority of major urban areas are involved. Thus, a strike that occurs in three of the ten cities in country X, but involves only salaried government employees, is three 302s (conflict scale). If the employees are joined by the armed forces in the capital, but not elsewhere, we code four 302s. A national strike involving salaried government employees and the armed forces would instead be coded as two 404’s. One could make the case that protests and demonstrations should also be coded with respect to the verbal conflict that they entail. As discussed above in section 2.4, however, we do not code these events as verbal conflict (i.e., the verbal conflict is implied by coding it as a 302 or 402).

Lastly, the coder must be able to differentiate between those strike demands that are economic in nature and those that are political. Strikes for higher wages are economic and not political events. As such, they are not codable under our scheme. We consider strikes against the minimum wage (which is a government policy) or the length of the work week, however, to involve political demands and to be codable as political events. We also consider all strikes by state employees to be codable IPI events.When a long term strike or demonstration ends, we stop coding conflictual events. In some cases, however, it is appropriate to code cooperative events as well which correspond with the cessation of hostilities. If active negotiations and/or a specific agreement precedes or corresponds with the end of the strike or demonstration, we code cooperative events between the parties. If, however, the strike or demonstration ends with no active resolution, we do not code cooperative events.

4.5 Executive Adjustment, Political Resignations (306, 301)

The key difficulty in coding these events is deciding how many events to code. A single executive adjustment can involve the replacement of several cabinet ministers. We adopt the rule that only one executive adjustment can occur per day. Thus, if the entire cabinet is asked to resign at once, we code one executive adjustment (306), just as we do if only the finance minister is asked to resign. If one minister is replaced on Monday and another on Thursday, however, two executive adjustments are coded. Similarly, if ten legislators resign together in protest, we code one political resignation (301). It is also important to note that a single executive adjustment is often a two step process. Many times, a cabinet minister is asked to resign on one date, and his or her replacement is named at a later date. The firing of the original minister and the hiring of a new one constitute a single executive adjustment; only one event (306) should be coded. We recognize the date of the firing of the original minister as the date of the executive adjustment in our coding. Finally, we are interested only in political conflict events. As a result, changes in personnel attributable to illness, natural death, retirement, pursuit of other political office, or other non-political explanations are not coded under the IPI rules.

Occasionally it is unclear when a member or members of a Cabinetn resign whether it is appropriate to code political resignations or an executive adjustment. In other words, there is sometimes ambiguity as to whether the politicians in question are resigning of their own accord due to policy differences or have been asked by the national executive to leave. When it is impossible from our sources to determine which of these is the case, we code an executive adjustment with the executive (10.1) as actor and cabinet members (11) as the target.

4.6 Censorship (303, 406, 501)

There are three different codes for censorship on the IPI conflict scale at the three hundred, four hundred, and five hundred levels. Individual acts of censorship are coded at 303. At this level, each individual act of censorship is a separate event. Closing two newspapers, for example, is two 303s. When a larger number of publications of a given type (e.g., Jewish books in Nazi Germany; all leftist newspapers) are closed, however, we consider this “widespread media censorship” and record a single 406. Complete censorship, involving a total lack of uncontrolled media access, is a single 501. 4.7 Mobilizing Resources (304, 502)

The conflict scale offers two codes for mobilizing resources and increasing power: 304 and 502. The difference between the two codes is whether the actions taken are legal or illegal. When groups use legal means to increase their power in relation to their adversaries, we code the events at the three hundred level. Using illegal means to increase power, however, displays a higher degree of conflict. As a result, illegal means of mobilizing resources are coded at the five hundred level. Coders should consult the conflict scale for examples of activities which fall into each of these categories.

4.8 Violent Political Acts (405, 505, 604)

Events 405, 505, and 604 on the conflict scale all represent violent political acts of differing intensities. Code 405 should be used for acts of intimidation which last less than twenty-four hours. Some examples are brief abductions or beatings. In general, these events are scare tactics or mild shows of force. Code 505 represents acts which involve a higher level of violence, but which do not involve deaths. Bombings, hijackings, kidnappings, and shootings which do not result in injury or deaths belong at this level. Violent acts which result in casualties should be coded at the 600 level (604). Whenever these distinctions are unclear in a report, we assume the less intense action.

For instance, if it is unclear whether an abduction lasted more or less than twenty-four hours, we code a 405. When there are no indications of injuries as a result of a bombing, we code a 505. Finally, when the same person is kidnapped and killed on the same day, this constitutes only one violent political act (a 604). If the two actions occur on separate days, however, we record both a kidnapping (505) and an armed attack (604) with one death from political violence.

4.9 Mass Arrests and Political Convictions (503, 606)

Event 503 on the conflict scale represents the arrest (among other things) of dissidents. We do not, however, code a separate event for the arrest of each individual. Rather, we code a separate event for every group of 100 people who are arrested, and a separate event for each leader arrested. This decision reflects our belief that it constitutes a higher level of conflict to arrest the leader of an organization than to arrest rank and file members. The following scale applies: 1-100 arrests 1 503 101-200 arrests 2 503’s 201-300 arrests 3 503’s 301-400 arrests 4 503’s 401-500 arrests 5 503’sand so on. For example, if seven members of a dissident group are arrested, it is a single 503. However, if 750 ‘rank and file’ and ten labor union leaders are arrested, it is coded as eight 503s (target=55.2) and ten 503s (target=55.1).

Consider the following report from Latin America Weekly Report, April 2, 1982: The CGT-organized trade union demonstrations against the government on 30 March were met with police repression. In Buenos Aires running battles between demonstrators and police went on for four hours; over 400 arrests were reported. Saul Ubaldini, the CGT general secretary, was arrested as he tried to lead the march on the Casa Rosada in Buenos Aires (page 12). In this case, we code five 503s (target 54) and one 503 (target 54.1). We use mass and elite subcodes only when the report specifically makes that distinction. It is important to keep in mind as well that if several groups are targeted, or if arrests occur in many cities or on many days, each of these is a separate event regardless of the total number of people arrested.

When those arrested are actually convicted of political crimes, we code a separate event for the conviction (606 on the conflict scale). We use a similar rule for calculating the number of events. Each leader convicted of a political crime constitutes a separate event. When a mass of people are convicted at once, we count one conviction for every 100 people affected. Occasionally, our sources will fail to report a political conviction, but will report a sentencing. In this instance, we code a political conviction on the basis of the sentencing report and check ‘Needs More Information’ in the hope of clarifying the conviction date from other sources.In some cases, the very act of bringing certain people to trial is politically significant. Particularly when former leaders or high ranking opposition members are tried, the trial itself is newsworthy and politically significant. When the beginning of a political trial is reported in our sources, we code the act of bringing an opponent to trial as a 304 on the conflict scale, a legal action to mobilize resources.

4.10 Bombings (505, 604)

There are two bomb codes (both on the conflict scale): 505 and 604. The distinction between these two codes is whether human beings are injured or killed as a consequence of the bombing. Attempted or failed bombings (i.e., defused bombs), sabotage bombings (economic targets such as electric pylons), etc., should be coded 505. Bombs (land mines, car bombs, etc.) that kill or injure people should be coded 604. Because the state claims a monopoly over coercive behavior, most bombings should list the state (10) as the target (see section 2.2). The exception is a well documented bombing of a political, social, ethnic, or religious group, in which case that group is coded as the target. A bombing of the headquarters of a political party or labor union, of a place of worship, or of a group of businesses owned by members of a particular ethnic group targets the group in question, not the state.

4.11 Assassinations (605, 702)

Codes 605 and 702 on the conflict scale are used to record unsuccessful and successful assassinations. We follow Taylor and Jodice’s World Handbook of Political Events in defining potential assassination targets. Taylor and Jodice state: Political Assassination… is a politically motivated murder or attempted murder … of a national leader, a high government official, or a politician. Among the national leaders included are chiefs of state, heads of government, ministers, legislators, judges and high-ranking civil servants and military officers. In addition to national leaders, state, province or district governors, mayors of large cities, and newspaper editors are included. Nationally prominent politicians who are not holding office (e.g., Dr. Martin Luther King and Robert F. Kennedy) were also included (1975).

Political murders and attempted murders of citizens who do not fall into these target categories (e.g., newspaper reporters, local trade union leaders, etc.) should instead be coded as 604 (if the target dies) or 505 (if the attempted murder is unsuccessful).

4.12 Reports of Disappearances (608)

Disappearances are coded as deaths from political violence in this project. A report which states, “Twelve human rights activists disappeared on Tuesday. Death squad activity is suspected.” is coded as twelve deaths from political violence and one report of disappearances (608 on the conflict scale). We treat summary reports of disappearances much like we do those of deaths (see section 2.10); we record the number of disappeared in the ‘Deaths’ field and complete the ‘Since When’ field to indicate the period over which the count is aggregated. For summary reports, however, it is not necessary to code a 608, as summary reports are not in themselves events.

4.13 Breaking a Truce, Breaking a Peace Treaty (603, 707)

These two events represent the abrogation of an agreement and the renewal of hostilities. The distinction between a peace treaty and a truce is discussed further in section 5.7. When a party to a truce or a peace treaty breaks the treaty by committing acts of violence, we code a 603 or 707 on the conflict scale for the act of breaking the agreement, and we additionally code the acts of violence at their appropriate level on the conflict scale.

4.14 Attacks, Clashes, and Battles (604, 704, 708, 806, 1002)

There are several different codes on the IPI conflict scale representing armed action. The least severe is the armed attack (604; 708). Armed attacks occur when one armed group or individual fires on an unarmed group or individual or when two armed groups exchange fire in an unorganized fashion (i.e., gunmen fire at police station and police fire back). An armed attack can involve no more than fifteen deaths at the 604 level, and no more than thirty deaths at the 708 level. If more than thirty deaths result from an armed action, the event should be coded either as a military clash (704) or as a massacre of civilians (806). A military clash (704) is an organized armed action involving groups which have adopted armed struggle as their accepted course of action. Common examples include battles between guerrilla groups and the army or between rival ethnic groups.

It is important to recognize that the distinction between an armed attack and a military clash is the organized nature of a military clash, not the number of deaths. A military clash need not involve an extensive number of deaths to be coded as such, but military clashes are always reciprocal events involving at least two actors.A massacre of civilians (806) is an attack by an armed group (either the state or another actor) on an unarmed group which results in more than thirty deaths. If fewer than thirty are killed as a result of the attack, the event should be coded as a 708 or a 604. Finally, event 1002 represents a major battle in which both sides defend territory with conventional military tactics and weaponry. We expect that events falling in this category will involve extensive casualties.

4.15 Occupation of Territory (808, 901)

In order for an event to qualify as “rebel occupation of territory” (808) , both sides must recognize that the rebels control the territory. The rebels must display widely recognized symbols of control and take up defensive positions. Simply hiding successfully from the government in an area without being dislodged does not qualify as occupation of territory. The requirements for 901, “rebels setting up rival government”, are even more strict. In this case, the rebels must be able to administer the territory (i.e., to provide services and to monopolize the use of coercion). For both events 808 and 901 we code only the first day. If the rebels lose control of the occupied territory, we code the last day and complete the ‘Since When’ field.

4.16 Coups d’Etat (608, 705, 706, 803, 904, 1003)

We have several codes for unsuccessful attempted coups d’etat and successful coups d’etat. The appropriate code depends on whether the coup is successful, the degree of violence involved, and the implications of the coup. More violent coups are more conflictual than less violent coups, and coups which actually alter the structure of government constitute a higher level of political conflict than those which simply serve to rotate rulers. Event 904 (successful violent coup which alters the structure of government) and event 1003 (successful violent coup and purge) are distinguished by the systematic, violent elimination of the former leadership which occurs in the latter case.


5.1 Introduction to the Cooperation Scale

The IPI cooperation scale is designed to be used in conjunction with the conflict scale. The two scales are parallel in many respects. Together they allow us to measure the underlying level of conflict within a society. The key to measuring cooperation on our scale is understanding the historic relationship among groups, individuals, and the state. Cooperative actions among parties to longstanding, violent disputes are unusual and important events. Routine cooperation among actors with established patterns of cooperative behavior, on the other hand, may not reflect changes in the underlying level of political conflict. Thus similar events between parties with different historic relationships have a different impact on their own societies and are coded at different levels on our cooperation scale. We do not code routine cooperative behavior.

In the normal process of government, there are groups which have a tendency to work together, often because they pursue similar interests. For instance, cooperation between students and educators, or leftist political parties and labor unions, generally has no impact on political behavior in the society. We assume that routine cooperative behavior represents a harmony of interests rather than policy coordination, and as such is not what we intend to capture with the IPI cooperation scale. We do, however, code cooperative actions between groups which have not consistently cooperated in the past, as this type of policy coordination is a measure of the underlying level of conflict in a domestic polity. The largest impact on societal relations occurs when groups which are involved in a protracted social conflict cooperate with one another.

Protracted social conflict (see Azar 1990 for a detailed discussion of the nature of the concept) refers to ongoing and seemingly unresolvable conflict within societies over both material and identity issues. At the broadest level, Azar (1990) argues that a combination of concerns over individual and community identity and security, distributive social justice and material interests drive most incidents of protracted social conflict. The existence of significant cleavages along identity or ideological lines and the willingness of groups to resort to violence in pursuit of their interests complicates conflict management. For parties involved in longstanding, often violent, disputes with interests in strong opposition to one another, minor actions may represent major breakthroughs. This intensity is reflected in our scales.

Similar events appear at different levels on the IPI cooperation scale depending on the historic relationship between the actor and target. For example, a meeting for negotiations between groups which are not involved in a protracted conflict is a 203 on the IPI cooperation scale, whereas a meeting for negotiations between parties to an ongoing conflict is a 505.

5.2 Promises (201, 202)

There are two important issues to keep in mind in determining if an event report is codable as a promise (201, 202). First, promises are verbal events only. Substantive actions carrying out promises are coded at a higher level on the scale (e.g., 303, 603). Second, to be codable, promises must be concrete. Vague references to possible future action do not constitute promises in this coding system. Consider, for example, the following report from Latin America Weekly Report, January 8, 1982, on a major speech by General Leopoldo Galtieri, President of Argentina, to the nation: A traditional theme of military speeches over the past six years was also present– the nod in the direction of an eventual return to civilian rule (p. 1). Such a vague reference is not codable as a promise. But the following report, from Latin America Weekly Report, June 4, 1982, is codable as a promise: Argentine interior minister, General Alfredo Saint Jean, says draft law legalizing activities of political parties will be released as promised, by the end of June (p. 12).

5.3 Relaxation of Repression (301, 401, 501, 601)

Relaxation of repression appears on the cooperation scale at four different levels: 301, 401, 501, and 601, depending on its scope (who is affected), and the severity of the sanctions which are ending. Coders should refer to the cooperation scale for examples of events which should be coded at each level. Frequently, in reporting the removal of government sanctions, journalists will also note how long they have been in effect. In this case, coders should be careful to code not only the removal or reduction in sanctions but also their institution.

Sometimes it is unclear whom to code as the target of a reduction in government sanctions. A reduction in censorship, for instance, impacts the media most directly, but clearly has an effect on the general population as a whole. For the purposes of this project, we code as target the group most directly affected by a reduction in sanctions. When it is unclear who is most directly affected, we code either unspecified social actors (75) or general population (09) as the target of the action. Thus, a reduction in censorship is coded with the state (10) as the actor and the media (46 or 47) as the target. Reopening universities is coded as an action by the state (10) directed at students (49). Legalizing political parties is coded as a cooperative action toward unspecified parties (89). We code the release of hundreds of unspecified political prisoners with the state (10) as actor and unspecified social actors (75) as target. The target of an event such as “ending a regional state of emergency” is (09), general population. As usual, each actor/target combination represents a distinct event.

5.4 Political Reform or Accommodation (302, 603)

There are two codes on the IPI cooperation scale for political reform, 302 and 603. The difference between the two is the scope and intensity of the reform undertaken. A 300 level reform action is a minor concession designed to lessen conflict, for instance, a minor increase in wages following a devaluation. Reforms which change the political environment, however, are coded at the 600 level. These include land reforms, the institution of a comprehensive system of new social welfare programs, and other such similar actions. Coders should consult the cooperation scale for further examples. These two codes (302 and 603) represent the implementation of these reforms. Promising reforms is coded at a lower level, either 201 or 202.

5.5 Talks, Agreements to Talk (203, 402, 505)

As mentioned above, there are two codes for holding talks or negotiations: 203 for talks between parties not involved in a protracted social conflict, and 505 for groups which are engaged in protracted conflict. In either case, we code one event per day. We have an additional code, however, which is relevant for parties engaged in a protracted social conflict. 402 on the cooperation scale represents either an agreement to talk between these groups, or reports of secret talks. Because we believe that public negotiations represent a higher degree of accommodation/cooperation than do secret negotiations, we code them at a higher level, but reports of secret negotiations, especially when they are persistent, are also significant. Each report of secret negotiations or dialogue through third parties is coded as a 402 with each party to the negotiations as actor and as target (i.e., there are two events in a two party negotiation). If after a report of such negotiations, one or both groups deny that the negotiations occurred, we code the denial(s) as verbal conflict (101). The original report, however, remains a 402.

5.6 Surrenders, Releases of Prisoners or Hostages (501, 503, 507, 601)

The main issue to be aware of in coding these events is the number of cooperative events to record. Just as we do not code a separate event for the arrest of each individual (see section 4.9), we do not record a separate event for the release of each individual. Again, we code one event for each 100 rank and file member of a group and one event for each group leader who is released or who surrenders. Once again, however, it is important to keep in mind that each day, each location, and each target represent individual events. There is an additional distinction made between releasing arrested or detained individuals and releasing political prisoners who have been convicted of political crimes. Releasing arrested or detained individuals is a 501, whereas releasing political prisoners convicted of political crimes is an event reflecting a higher degree of cooperation or accommodation. As a result, it is coded at a higher level on the cooperation scale; it is a 601. In coding reports of surrenders, coders must be careful to appropriately identify the actor. We can not assume that groups are responsible for the actions of individual members. As a result, in the example listed in section 2.6 which discusses the surrender of six members of the M-19, we code the actor as unspecified guerrillas (45) and the target as the state (10).

5.7 Cessation of hostilities (502, 602, 703)

Events 502, 602, and 703 on the IPI cooperation scale all represent the termination of violence, but with different degrees of scope and permanence. Event 502, a cease fire, represents the cessation of hostilities temporarily or in an isolated area. Event 602, a truce, is an agreement between groups to end hostilities indefinitely. The intention is to allow time for negotiations to see if the conflict can be resolved. During this period, the armed struggle is renounced. The signing of a peace treaty (event 703) implies that the conflict has been permanently managed or resolved and is often followed by disarmament.

5.8 Elections (504, 603, 704, 801, 902)

When a society has not been democratic and begins to institute or reinstitute a democratic process into its political affairs, the holding of elections represents a crucial cooperative event. In societies in which democracy is institutionalized, elections represent no more than the normal process of governing. Thus, the IPI cooperation scale includes numerous codes for electoral activity, but these are intended to apply only to societies without
institutionalized democratic systems, as anything constituting the normal process of government is not codable under our system (see section 2.1). The process of planning for and implementing national elections involves many separate events. We code a concrete agreement to allow elections to be held as a 504 on the cooperation scale. When a date for national elections is set, or when minor elections are held, we code a 603, “implementing policy reform”.

The election itself is codable as a 704, an 801, or a 902 depending on the conditions. Any election held without full participation (i.e., an election which any of the parties boycotts) is a 704. The boycott would additionally be coded as a 404 on the conflict scale with the party refusing to participate as actor and the state as target. An election with full participation is coded as either an 801 or a 902 depending on whether the election was preceded by the development of a new constitution. In a society where a democratic process is not institutionalized, the assumption of power by the winners and the acceptance of the results by the losers does not always follow elections without incident. As a result, we recognize the act of allowing the opposition to take power following an election as a separate cooperative event, coded as a 603 on the cooperation scale with the state (10) as actor and the general population (09) as target. Similarly, voiding the results of an election is coded as an 807 on the conflict scale, again with the general population (09) as target. If an election is not deemed free and fair by international observers, we still consider the holding of the election to be a cooperative event and code it as such. We additionally code one or more conflictual events, however, to represent any violations of the electoral process.


In some instances, the same action will represent cooperation toward one target and conflict toward another. One common example is a judicial decision, which generally represents cooperation from the judicial branch toward the winner and conflict from the judicial branch toward the loser. Consider this report from Latin America Weekly Report, April 2, 1982: The labour courts judged that the Cimetal management had not behaved responsibly and ruled the strike legal, a highly unusual decision (page 7). In this case, we code two events: one cooperative event (307) between the judicial branch (13) and the labor union (55) and one conflictual event (307) between the judicial branch (13) and business elites (64). Another example in which both the conflict and cooperation scales are used to code the same event was highlighted in section 2.3 above. When two or more groups coordinate their actions against a common enemy, we code both conflictual events from the groups to the common enemy, and cooperative events between the groups themselves. It is important to note, however, that it is the coordination of action which represents a cooperative event rather than the action itself. Thus, if two groups protest against the government together for three days, we code six conflictual events (one per group per day), but only two cooperative events. We do not assume a need to recoordinate policies each day, and thus we need not repeat the coding of the cooperative events for additional days of the same protest. The same assumption is invoked for events which occur in several cities. If two groups protest together in two cities, we code four conflictual events and two cooperative events.


Azar, Edward E. 1975. Ten Issues in Events Research. Theory and Practice of Events Research. E. E. Azar and J. D. Ben-Dak, Eds. New York: Gordon and Breech: 1-16.

Azar, Edward E. 1982. The Codebook of the Conflict and Peace Data Bank. Center for International Development, University of Maryland.

Azar, Edward E. 1990. The Management of Protracted Social Conflict: Theory and Cases. Brookfield, Vermont: USA Gower Publishing Company.

Burgess, P. M. and R. W. Lawton. 1972. Indicators of International Behavior: An Assessment of Events Data Research. Beverly Hills: Sage.

Davies, John L. 1993. The Global Event-Data System Manual. Center for International Development and Conflict Management, University of Maryland.

Gerner, Deborah J., Philip A. Schrodt, Ronald A. Francisco, and Judith L. Weddle. 1992. Issues in the Creation of Event Data: Machine Coding of Regional and International Sources. Paper presented at the Annual Meeting of the American Political Science Association, Chicago.

Goldstein, Joshua S. 1991. Reciprocity in Superpower Relations: An Empirical Analysis. International Studies Quarterly 35: 195-209.

Gurr, Ted Robert. 1970. Why Men Rebel. Princeton: Princeton University Press.

Gurr, Ted Robert. 1993. Minorities at Risk: A Global View of Ethnopolitical Conflicts. Washington, DC: United States Institute of Peace Press.

Howell, Llewellyn D. 1983. A Comparative Study of the WEIS and COPDAB Data Sets. International Studies Quarterly 27: 149-159.

Keohane, Robert O. 1984. After Hegemony: Cooperation and Discord in the World Political Economy. Princeton: Princeton University Press.

McClelland, Charles. 1972. World Events Interaction Survey. Los Angeles: University of Southern California, ICPR Edition.

Moore, Will H. and Ronny Lindstrom. 1994. The Violent Intranational Conflict Data Project (VICDP) Codebook. Department of Political Science, University of California- Riverside.

Muller, Edward N. and Mitchell A. Seligson. 1987. Inequality and insurgency. American Political Science Review 81: 425-451.

Muller, Seligson, Fu, and Midlarsky. 1989. Land Equality and Political Violence. American Political Science Review 83: 577-595.

Taylor, Charles Lewis and Michael C. Hudson. 1975. The World Handbook of Political and Social Indicators, Second Edition. New Haven: Yale University Press.

Taylor, Charles Lewis and David A. Jodice. 1983. World Handbook of Political and Social Indicators. New Haven: Yale University Press.

Taylor, Charles Lewis and David A. Jodice. 1985. The World Handbook of Political and Social Indicators III, 1948-1982: Daily Political Events Data. ICPSR Study #7761. Ann Arbor: Inter-University Consortium for Political and Social Research.

Go to the IPI Home Page.

Go to Will H. Moore’s Home Page.

I can be reached by email at


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s