Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, INTERNATIONAL STUDIES (internationalstudies.oxfordre.com). (c) Oxford University Press USA, 2017. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 20 February 2018

Review of Available Data Sets

Summary and Keywords

The International Studies Association’s (ISA) Scientific Study of International Processes (SSIP) section is dedicated to the systematic analysis of empirical data covering the entire range of international political questions. Drawing on the canons of scientific inquiry, SSIP seeks to support and promote replicable research in terms of the clarity of a theoretical argument and/or the testing of hypotheses. Journals that have been most likely to publish SSIP-related research include the top three general journals in the field of political science: the American Political Science Review, American Journal of Political Science, and Journal of Politics. A number of more specialized journals frequently publish research of interest to the SSIP community, such as Conflict Management and Peace Science, International Interactions, International Organization, International Studies Quarterly, Journal of Conflict Resolution, and Journal of Peace Research. Together, these journals published a total of 1,024 qualifying articles between 2003 and 2010. These articles cover a wide range of topics, from armed conflict and conflict management to terrorism, international political economy, economic development or growth, monetary policy, foreign aid, sanctions, human rights and repression, international law, international organizations/institutions, and foreign policy attitudes and beliefs. Data users who are interested in conducting their own research must: choose the most appropriate data set(s), become familiar with what the data set includes and how its central concepts are measured, multipurpose data sources, investigate missing data, and assess robustness across multiple data sets.

Keywords: International Studies Association, journals, research, conflict management, terrorism, international political economy, foreign policy, data sets, empirical data, data analysis

Introduction

Article 3 of the charter of ISA's Scientific Study of International Processes (SSIP) section, which is available on the section's website at www.isanet.org/ssip/proclaims the following purpose for the section:

SSIP is dedicated to bringing together researchers who, at all levels of analysis and with respect to the entire range of international political questions, pursue these issues using (1) formally stated arguments and/or (2) systematically collected and analyzed empirical data. Following the canons of scientific inquiry, the section seeks to support and promote replicable research in terms of the clarity of a theoretical argument and/or the testing of hypotheses.

The systematic analysis of empirical data thus comprises a central goal of the section. This chapter reviews the data sets that have been used in recent research, in order to help researchers pursue this goal.

This chapter begins with a brief review of recent quantitative research of interest to SSIP section members. These articles are classified by their substantive focus, in order to identify the research interests that have attracted the most scholarly attention in recent years. The most common data sets used for each topic are then discussed, in order to assess the many options that are available to scholars working in these areas. The chapter concludes with recommendations for data users to consider in their research.

Survey of Published Articles

An important starting point for this review of data sets is the purposes for which these data sets are used. This survey identifies all relevant articles from the journals that have been most likely to publish SSIP-related research. This includes the top three general journals in the field of political science: the American Political Science Review, American Journal of Political Science, and Journal of Politics; while SSIP membership is not limited to political scientists, many fewer articles have been published by section members in the general journals of such fields as economics, geography, psychology, or sociology. This survey also includes a number of more specialized journals that frequently publish research of interest to the SSIP community: Conflict Management and Peace Science, International Interactions, International Organization, International Studies Quarterly, Journal of Conflict Resolution, and Journal of Peace Research.

To be included in this survey, an article must use quantitative data to run empirical analyses of some international phenomenon. This excludes articles that develop but do not test theoretical arguments, whether or not they use formal mathematical modeling techniques, as well as articles that employ simulations and experiments. While such articles often make important scholarly contributions, the lack of quantitative data leaves them beyond the scope of any survey of data sets. The topic of the article must also involve some sort of international phenomenon, broadly defined, in a way that is consistent with the SSIP section's charter.

Table 3.1 summarizes the articles meeting these criteria over the eight years before this chapter was written. Together, these nine journals published 1024 qualifying articles between 2003 and 2010. The three general journals published 158 of these articles, led by the Journal of Politics (JOP) with 67 articles; the American Journal of Political Science (AJPS) published 58, while the American Political Science Review (APSR) published 33. There does not appear to be any systematic trend in the number of SSIP articles appearing in these journals over time; while 2007 and 2010 saw the most articles of any year across these three journals, 2006 and 2009 saw the fewest articles. The six specialized international relations journals published a further 866 articles, led by the Journal of Peace Research with 198 articles and the Journal of Conflict Resolution with 195. International Studies Quarterly published 167 articles over these eight years, followed by International Organization with 126, and Conflict Management and Peace Science with 92, and International Interactions with 88. Together, these six journals have published substantially more qualifying articles over time, ranging from 90 to 95 between 2003 and 2005, 103 to 119 between 2006 and 2009, and 142 in 2010.

Table 3.1 Journal Articles with Quantitative Analyses of SSIP-Related Topics, 2003–2010

Number of Articles

Journal

2003

2004

2005

2006

2007

2008

2009

2010

Total

AJPS

4

9

8

5

10

7

5

10

58

APSR

8

3

3

3

6

4

1

5

33

CMPS

6

3

11

8

15

13

19

17

92

II

10

14

10

14

11

13

8

8

88

IO

14

13

10

26

16

16

13

18

126

ISQ

14

17

18

27

15

18

25

33

167

JCR

27

23

23

23

23

26

25

25

195

JPR

19

21

23

16

23

26

29

41

198

JOP

6

5

8

5

12

11

8

12

67

Total

108

108

114

127

131

134

133

169

1024

These articles cover a wide range of topics, as seen in Table 3.2. Each article is categorized by the topic that best summarizes the main focus of the article, generally based on the dependent variable(s) being studied. Where multiple types of dependent variables are used, the article is categorized based on the topic that receives that most space in the tables and the accompanying text.

Well over half of these articles focused on armed conflict in some form. The most common conflict-related topic has been the causes, processes, and impact of interstate conflict, which was the primary focus of 321 of the articles (31.4 percent of the total). The causes, processes, and impact of other forms of conflict that involve non-state actors (including such related topics as civil/intrastate war, ethnic conflict, terrorism, repression, and human rights violations) account for 280 articles (27.2 percent). Conflict management (including such topics as negotiations, mediation, and peacekeeping) accounted for 49 articles (4.8 percent). Another 234 articles (22.9 percent) fall under the broad heading of the international political economy, covering topics ranging from international trade and trade policy (the most common economic topic, with 47 articles) to economic development or growth, monetary policy, foreign aid, sanctions, and debt. A further 80 articles (7.8 percent) addressed topics related to treaties, norms, or international organizations/institutions, ranging from the formation of alliances to voting in the United Nations or the diffusion of specific norms. Finally, 60 articles (5.9 percent) used survey research to examine political attitudes or beliefs on international matters, ranging from minority-state relations to the desirability of free trade, globalization, or anti-terror policies.

Several notable trends in published topics have become apparent over this period. First, while interstate conflict has been covered by more articles than any other category, intrastate conflict and terrorism are getting much more scholarly attention in recent years. During 2003 and 2004, many more published articles addressed interstate conflict than intrastate conflict and terrorism (58:12 and 40:22); the numbers were essentially even by 2006 and 2007 (36:34 and 31:36) and interstate conflict articles were clearly outnumbered in more recent years (most notably 24:45 in 2009). This suggests that researchers' attention is focusing on more current topics. A number of studies have noted that intrastate conflicts now outnumber interstate (e.g., Hensel 2002; Harbom and Wallensteen 2007), and terrorism is an ever-present concern. Research on interstate conflict will not stop any time soon, but research on other forms of conflict is now equally prominent, and a cursory examination of papers presented at recent professional conferences suggests that this trend is not likely to be reversed in coming years.

Table 3.2 also reveals several other trends. Besides the shift in focus from interstate conflict to intrastate conflict and terrorism, these journals have been publishing much more work on more peaceful, cooperative dimensions of international processes. Research on conflict management is becoming more common, increasing from 7 articles in 2003–4 to 15 in 2009–10. Research on the formation and dynamics of international treaties, norms, organizations, and institutions is also becoming much more common, increasing from 11 articles in 2003–4 to 28 in 2009–10, and survey-based research on foreign policy attitudes and beliefs has increased from 11 articles in 2003–4 to 22 in 2009–10.

Table 3.2 Categorization of Articles by Topic Area

Number of Articles

Article Topic

2003

2004

2005

2006

2007

2008

2009

2010

Total (%)

Interstate Conflict

58

40

46

36

31

35

24

51

321 (31.4)

Intrastate Conflict or Terrorism

12

22

32

34

36

41

45

58

280 (27.2)

Conflict Management or Negotiation

4

3

4

13

4

6

7

8

49 (4.8)

International Political Economy

27

28

17

29

36

38

33

26

234 (22.9)

Treaty/Norm/IO Behavior

3

8

7

11

13

10

13

15

80 (7.8)

Foreign Policy Attitudes & Beliefs

4

7

8

4

11

4

11

11

60 (5.9)

We now turn to a survey of the data sets that have been used in these articles, before concluding with several recommendations for the most effective use of these different data sets. The most prominent data sets used to study each of the categories in Table 3.2 are included. This survey also includes a number of other data sets on topics like geography, national capabilities, and political systems, which are rarely used as dependent variables but frequently appear as independent variables or control variables.

Survey of Data Sets

The articles covered in this survey have used dozens of data sets to measure the dependent, independent, and control variables of interest. We now examine the most widely used data sets, categorized by the general purpose for which each is typically used. It should be noted that some data sets may be listed under several different categories, as with data sets that are used to study both armed conflict and peaceful conflict management.

It is important to note that data sets are constantly being updated. Thus, this survey indicates the current state of each data set as of the time of this writing, but interested users should consult the official website or contact the data sets' creators to determine whether there is a more up-to-date version. Listing the URLs where each data set may be downloaded is beyond the scope of this compendium, and the speed with which file locations change on the Internet would render such a printed list obsolete within months of publication, so an online appendix with this information is made available at: www.paulhensel.org/compendium.html. The links in this online appendix allow interested users to download the latest version of most of these data sets, and in many cases codebooks or other supplementary materials.

Interstate Conflict Data Sets

More SSIP-related research has studied armed conflict, whether between or within states, than any other general subject. Most research on armed conflict has relied on a relatively small number of prominent data sets. For interstate conflict, the most widely used data set in this period has been the Correlates of War (COW) project's Militarized Interstate Dispute (MID) data set, which covers the threat, display, or use of military force between two or more states (Ghosn et al. 2004). Zeev Maoz has produced a dyadic version of this data set (DYMID) that attempts to correct the coding of dispute-wide variables for each distinct pair of adversaries in multilateral disputes. Several other conflict data sets use more restrictive thresholds than the MID data. The International Military Intervention (IMI) and Military Intervention by Powerful States (MIPS) data sets require the foreign deployment of military forces. The International Crisis Behavior (ICB) project's crisis data set (Brecher and Wilkenfeld 2000) requires leaders' perception of “a threat to one or more basic values, along with an awareness of finite time for response to the value threat, and a heightened probability of involvement in military hostilities” (Brecher and Wilkenfeld 2000: 3). There are many more MIDs than ICB crises during any given time period because these perceptions are not present in many lower-intensity MIDs, particularly those that remain limited to isolated threats to use force or displays of force. The Uppsala Conflict Data Program (UCDP) and the PRIO Center for the Study of Civil War collaborate on the Armed Conflict data set (Gleditsch et al. 2002), which currently covers conflicts since 1946 and is updated each year to reflect the past year's conflicts. This data set includes all conflicts between two parties, at least one of which is a state government, that produce at least 25 battle deaths in a given year. The highest severity threshold involves the COW project's interstate war data, which requires at least 1000 battle deaths in sustained combat between the regular armed forces of at least two nation-states (Sarkees 2000).

Besides identifying cases of conflict, most of these data sets include additional data on the severity of conflicts. The MID data includes the level of hostility (threat to use force, display of force, use of force, interstate war), the presence or level of fatalities in the dispute, and the duration of the dispute. The ICB data set includes information on a number of dimensions of crisis behavior and severity, ranging from the type of military action taken to start the crisis to the timing, centrality, and severity of violence in the crisis. The Armed Conflict data set measures the severity of the conflict in each year of observation rather than providing only an aggregated total, distinguishing between minor conflicts (which produce 25–999 battle deaths in the year) and wars (which produce at least 1000). The COW war data set includes an estimate of the total battle deaths in each war. Lacina and Gleditsch (2005), Valentino et al.'s “Covenants without the Sword” data set, and the UCDP's One-Sided Violence data set supplement the COW interstate war and Armed Conflict data sets with more detailed data, including both soldiers and civilians killed in combat. Most of these data sets also include some information on the contentious issues or the nature of the incompatibility that is involved in the conflict, as well as its outcome.

Event Data

There is also a long tradition of research on broad patterns of conflict and cooperation, including a number of events below the threshold of militarized conflict. Researchers have been using event data from the Conflict and Peace Databank (COPDAB) and World Events Interaction Survey (WEIS) data sets for decades now. More recently, the Kansas Events Data System (KEDS) project has used machine coding of newsfeeds to produce event data, initially using the WEIS event categorization before adding its own Conflict and Mediation Event Observations (CAMEO) coding scheme. King and Lowe (2003) have introduced a machine-coded data set with over ten million dyadic events, coded using techniques from Virtual Research Associates (VRA). Other machine-coded data sets include the Protocol for the Analysis of Nonviolent Direct Action (PANDA) data (Bond et al. 1997) and the Integrated Data for Events Analysis (IDEA) data (Bond et al. 2003).

Interstate Rivalry

Some scholars are more interested in longer-term conflictual relationships or “rivalries.” The so-called “dispute density” approach identifies rivalries based on the frequency of militarized conflict over a certain period of time. The best known example is the Goertz and Diehl rivalry data set (Goertz and Diehl 1993; Klein et al. 2006) that is based on the MID data, and Hewitt (2005) followed a similar approach with the ICB crisis data. Thompson's (2001) alternative “strategic rivalries” data set used intensive historical research to determine which actors identified and treated each other as rivals during which periods – some of which lasted for only a few years (as opposed to the two decade minimum used in many dispute density approaches) or never produced a single militarized dispute. While the dispute density approach is more useful for identifying the most conflictual adversaries, the strategic rivalry approach may be more appropriate for studies that assume a competitive relationship between adversaries that perceive each other as mutual threats, particularly for studying the conditions under which such competition is likely to become militarized (since the dispute density approach cannot identify any rivalries that never became militarized).

Contentious Issues

Several data sets identify contentious issues, or specific subjects of disagreement between states, and can be used to study both armed conflict and conflict management over these issues. Huth and Allee's (2002) territorial dispute data includes information on disagreements over territory, while the Issue Correlates of War (ICOW) data set (Hensel et al. 2008) includes information on disagreements over territory, cross-border rivers, and maritime zones. One important distinction between these two data sets is their spatial-temporal domain, although this distinction is being reduced as additional data collection is done to extend each data set: the Huth/Allee data is currently limited to events between 1919 and 1995 and includes events across the entire world, while the latest public release of the ICOW data set covers 1816 to 2001 for territorial claims and 1900 to 2001 for river and maritime claims and covers events in the Americas, Europe, and the Middle East. Another distinction is that the Huth and Allee data is organized with a single observation for each directed dyad-year (regardless of how many claims are underway in the dyad that year), while the ICOW data includes a separate observation for each distinct claimed territory, river, or maritime zone that is underway between two states in any given year (with several dyads engaging in multiple claims simultaneously).

A related data set is the COW Territorial Change data set (Tir et al. 1998), which includes all transfers of territory to or from nation-states since 1816. All three data sets include variables to measure the salience or value of the issue at stake, such as the economic or strategic value of the claimed or transferred territory. They also include information on the management of the claims, which has been used to study both armed conflict and conflict management. In particular, the Huth/Allee and ICOW data sets include information on armed conflict over each claim, and the Territorial Change data set includes details on the process by which it was transferred.

Intrastate, Civil, and Ethnic Conflict Data Sets

As noted above, recent years have seen a marked increase in research on armed conflict that involves at least one non-state actor. As with interstate conflict, such forms of conflict have been addressed by a number of data sets, with very different thresholds. Banks' Cross-National Time Series data set – which is only available as a commercial product – includes a variety of events such as assassinations, strikes, riots, revolutions, demonstrations, and guerrilla warfare. The UCDP/PRIO Armed Conflict data set, discussed earlier with respect to interstate conflict, includes all conflicts with at least 25 fatalities that include a state government on at least one side of the conflict; the vast majority of conflicts in this data set are internal rather than interstate in nature. The UCDP has also released a Non-State Conflict data set that includes armed conflicts where neither party is a state, as well as a One-sided Violence data set for attacks on civilians.

Focusing on events with a higher threshold, the COW project also collects data on wars that do not involve states on each side: intrastate wars (formerly known as “civil wars”) that involve a state's forces against another actor within the state, and extrastate wars (formerly “extra-systemic wars”) that involve a state against a non-state actor from beyond its borders. Fearon and Laitin (2003) have produced an Ethnicity, Insurgency, and Civil War data set that includes somewhat different cases than the COW intrastate war data. The UCDP Non-State Conflict includes conflicts that do not involve a state participant on either side. Lacina and Gleditsch (2005) and the UCDP One-Sided Violence data supplement the COW, Armed Conflict, and Fearon/Laitin data with more detailed battle death data. Lyall and Wilson's Correlates of Insurgency data set focuses on insurgencies. Finally, the Political

Instability Task Force (formerly the State Failure Task Force) provides data sets on revolutionary wars, ethnic wars, adverse regime changes, and genocides and politicides.

Other details on civil wars and related conflicts are provided by several sources. The UCDP offers Conflict Termination and Peace Agreements data sets with details on how armed conflicts were terminated. Doyle and Sambanis' (2000) data on international peacebuilding, Fortna's Peacekeeping in Civil Wars data, and the UCDP Managing Low-Intensity Conflict (MILC) data examine the effectiveness of various types of peaceful conflict management techniques. The Minorities at Risk data (MAR) also offers an intrastate equivalent to the interstate data sets on contentious issues discussed above. MAR identifies and tracks nearly 300 politically active ethnic groups, with information on group characteristics (including the nature of their inequalities or other grievances) and group activities (including political organization and armed conflict). Like the interstate issues data sets, this is useful for studying the origins of armed conflict, because it includes data on cases that never escalated to the point of violence and it includes a variety of information that might help to predict which groups are most likely to escalate to this level.

Human Rights and Repression

Topics related to human rights and repression are beginning to see much more systematic research. Two data sets have been most prominent, both of which draw from the US State Department's Country Reports on Human Rights Practices and Amnesty International's Annual Reports. The CIRI Human Rights data set (Cingranelli and Richards 1999) contains information on 15 different dimensions of government respect for human rights, ranging from torture to women's rights and workers' rights, which are provided in both disaggregated fashion for each dimension and several indices that aggregate the data into a multidimensional indicator of support for human rights. The Political Terror Scale (Gibney and Dalton 1996; Poe et al. 1999) categorizes countries more generally using a 5-point scale, ranging from the secure rule of law (where political imprisonment, torture, and murder is rare) to society-wide political terror (where leaders ruthlessly pursue personal or ideological goals and civil and political rights violations are widespread).

Terrorism

The most widely used data source in terrorism-related articles published during this period has been the International Terrorism: Attributes of Terrorist Events (ITERATE) data, which includes transnational terrorist events since 1968. While quite useful, this data set has the disadvantage of being a commercial product with a substantial fee for users whose institutions have not purchased site licenses. Other data sets that have been used, often with more limited temporal domains, include the Global Terrorism Database (GTD) from the National Consortium for the Study of Terrorism and Responses to Terrorism (START); the Institute for Counterterrorism's (ICT) International Terrorism Database; the RAND Database of Worldwide Terrorism and its predecessors, the RAND Terrorism Chronology and RAND-MIPT Terrorism Incident Database; the Terrorism, Insurgencies, and Guerrillas in Education and Research (TIGER) Suicide Attacks Worldwide and Terrorist Groups Worldwide data sets; and the US State Department's annual Country Reports on Terrorism (formerly known as Patterns of Global Terrorism).

Conflict Management Data Sets

During the period covered by this review, scholars have begun to focus on peaceful conflict management efforts, beyond the outbreak or escalation of armed conflict. Several distinct approaches have been used to study peaceful conflict management. One approach examines the management of specific types of contentious issues, using the issue data sets discussed earlier. Both the Huth/Allee and ICOW data sets include peaceful efforts to settle the issues, ranging from bilateral negotiations to non-binding third party activities such as good offices or mediation and binding third party activities such as arbitration or adjudication. Using such data sets allows the scholar to focus on all attempts to manage or settle issues of each type, including attempts made in peacetime as well as attempts made during ongoing crises or wars, although obviously limited to attempts to manage a limited number of issues.

The other approach examines the management of armed conflicts once they have begun. The ICB crisis data set, discussed earlier, includes numerous variables related to the actions of outside actors during the crisis. Separate information is collected for the actions of each superpower as well as global and regional organizations, where relevant, including activities ranging from political or military support for the crisis participants to fact-finding, mediation, and other peaceful settlement techniques. Similarly, Bercovitch's International Conflict Management data set (Bercovitch and Fretter 2004) includes several thousand attempts to manage conflictual international relationships that have been marked by the threat or use of force since 1945. Several other data sets focus on various forms of intervention into conflicts, ranging from peacekeeping operations to military intervention to support one side; examples include Regan's Interventions in Civil Wars (ICW) data, Mullenbach and Dixon's Third Party Interventions in Intrastate Disputes (TPI) data, and the Doyle/Sambanis peacebuilding data discussed earlier. While these data sets do not include attempts to manage conflicts that never led to armed conflict, which the issues data sets do, they have the advantage of including a much broader range of issues being managed.

Economic Data Sets

Economic topics make up about one-fourth of the quantitative articles published during this period, and economic variables are used frequently in studying both interstate and intrastate conflict. The most widely used data sets include information on a wide range of economic factors, and cover a relatively long spatial-temporal domain (although rarely extending to before World War II). A good example is the Penn World Tables data set, which is widely used to measure overall wealth or income levels; it includes a variety of GDP, GNP, openness, and savings variables, measured in both constant and current prices. Other prominent examples include the Bank for International Settlements' International Financial Statistics, IMF Government Finance Statistics and International Financial Statistics, UNCTAD Handbook of Statistics, and World Bank World Development Indicators. Furthermore, Maddison's Historical Statistics database provides population and GDP estimate for some economies as far back as 1 AD, and Gleditsch (2002) has produced an expanded collection of GDP data that attempts to fill in missing values in the Penn World Tables data set.

A number of other economic data sets focus on more specific topics: Chinn and Ito's Capital Account Openness (KAOPEN) data, Cukierman and Webb's (1995) central bank data, the Fraser Institute's Economic Freedom in the World index, Klein and Shambaugh's Exchange Rate Regimes data, the KOF Index of Globalization, and Transparency International's Corruption Perceptions index. Deininger and Squire's “Measuring Income Inequality” data and the University of Texas Inequality Project (UTIP) data set measure inequality. The IMF Balance of Payments data covers each state's balance of payments with the rest of the world, including the total goods, services, factor income, and current transfers an economy receives from or provides to the rest of the world as well as capital transfers and changes in each economy's external financial claims and liabilities. The World Bank's Global Development Finance Data (formerly known as the World Debt Tables) covers the external debt of developing countries. Economic sanctions have been a major research topic, accounting for more than 20 articles in this period, and are covered by Hufbauer, Schott, Elliott, and Oegg's sanctions data, Marinov (2005), and the Threat and Imposition of Sanctions (TIES) data.

Focusing on international trade and investment, the IMF's bilateral Direction of Trade data set is used relatively rarely by itself in SSIP publications, but it forms the basis for several widely used data sets. Three prominent examples are Russett and Oneal (2001), Gleditsch (2002), and the COW trade data set (Keshk et al. 2004; Barbieri et al. 2009). These data sets vary in how the IMF data or other trade data sources are handled, particularly in the treatment of missing data, as well as in the use of alternative sources that are used to supplement the IMF data. Feenstra and Lipsey's NBER-UN World Import and Export Data separates trade flows by SITC category, rather than providing aggregated data on all trade between countries in a given year. Trade barriers and various trade and investment policies are covered by the IMF Annual Exchange Arrangements and Exchange Restrictions data set (formerly the Annual Report on Exchange Restrictions), UNCTAD Trade Analysis and Information (TRAINS) data, UNCTAD Bilateral Investment Treaties (BIT) and Double Taxation Treaties (DTT) data, and the Kee, Nicita, and Olarreaga Overall Trade Restrictiveness indices. Flows of foreign aid and foreign direct investment are covered in the OECD International Development Statistics and International Investment Statistics data sets as well as the UNCTAD FDI data set, and the SIPRI Arms Transfers database covers the flow of arms.

Geographic Data Sets

Geographic variables are used frequently in studies of interstate conflict, dating at least as far back as Bremer's (1992: 336) conclusion that contiguity is the strongest predictor of war and “should be commonly included in all studies of war, at least as a control variable.” The most widely used data set is the COW Contiguity collection (Stinnett et al. 2002), which includes two distinct data sets: Direct Contiguity (for contiguity between the homeland territory of two nation-states) and Colonial Contiguity (for contiguity between a state and a colony or other dependency, or between two such dependencies). Both include entities that share a land border as well as those that are separated by up to 400 miles of sea. The Furlong and Gleditsch (2003) Length of International Boundaries data set also provides an estimate for the length of each land border between states, while Gleditsch and Ward (2001) and the CEPII Distance Dataset measure the distance between states' capital cities or closest points.

A growing number of studies also use geographic data to help understand conflict, as well as environmental phenomena such as environmental degradation. The Armed Conflict Location and Events Data (ACLED) and the COW project's MID Location data set record the exact location of each armed conflict, allowing for more detailed analysis of geographic factors in the outbreak or spread of conflict. Sources such as Fearon and Laitin's Ethnicity, Insurgency, and Civil War data set and Collier and Hoeffler's “greed and grievance” data set include information on terrain and other related geographic factors that might help to explain armed conflict, often collected using Geographic Information Systems (GIS) technology. The World Resource Institute's annual World Resources volumes and the UN Food and Agriculture Organization's Terrastat and State of the World's Forests databases have been used for a variety of resource- and environment-related topics. The Transboundary Freshwater Dispute Database project (Hamner and Wolf 1997–8), PRIO's Shared Rivers and Shared River Basins data sets (Gleditsch et al. 2006), and ICOW's River Claims data set have been used to study a number of dimensions of water scarcity, demands on water, shared rivers, and river treaties. Data on the distribution or production of specific resources in sources such as DIADATA (Gilmore et al. 2005), DRUGDATA (Buhaug and Lujala 2005), GEMDATA (Lujala 2009), and PETRODATA (Lujala et al. 2007), Humphreys (2005), and World Bank data have been used to identify the value of primary commodity exports to each state's economy as well as to distinguish between renewable and non-renewable resource abundance.

Finally, scholars are beginning to use data on natural disasters or other physical processes to help understand the origins, escalation, or settlement of civil war or other forms of internal conflict. The US Geological Survey's Centennial Earthquake Catalog has been used to identify earthquakes around the world. The EM-DAT Emergency Events Database includes data on a variety of natural disasters and related emergency events as far back as 1900, ranging from earthquakes, floods, and droughts to epidemics, insect infestations, and heat waves.

Political Data Sets

One consequence of the well-known democratic peace is an effort to consider the possible impact of regime type in almost any study of international processes. Even where a scholar does not initially seek to include regime type, manuscript reviewers and journal editors often suggest its addition to the model, at least as a control variable. The most widely used source is the Polity IV data set (Jaggers and Gurr 1995; Marshall et al. 2002), which includes data on specific political institutions that are relevant to leader selection and constraints on leaders' actions. The Polity data sets are much better known for their summary indices, though, including separate indices of institutionalized democratic and autocratic characteristics as well as an index that subtracts the institutionalized autocracy score from the institutionalized democracy score and runs from -10 (most autocratic) to +10 (most democratic). Other data sets that have been used to identify democracies or their elections include Vanhanen's Polyarchy data, Freedom House's Freedom in the World data, Golder's (2005) Democratic Electoral Systems around the World data, the Data on International Election Monitoring data set, Johnson and Wallack's Database of Electoral Systems and the Personal Vote, the International IDEA Voter Turnout data, and the Multidimensional Institutional Representation of Political Systems (MIRPS) / Scalar Index of Polities (SIP) data.

Other data sets focus on specific political constraints, rather than summarizing political systems in a single variable. Examples include the ACLP Democracy and Development data (Przeworski et al. 2000); Armingeon et al.'s Comparative Political Data Sets; Freedom House's Freedom of the Press data on press freedom; Henisz’ Political Constraint Index (POLCON); the Logic of Political Survival data (Bueno de Mesquita et al. 2003), which provides a measure of the size of countries' winning coalitions and selectorates; Persson and Tabellini's “Electoral Rules and Corruption” and “Economic Effects of Constitutions” data; Regan and Clark's Institutions and Elections project (IAEP); Woldendorp, Keman, and Budge's data on partisan government; and the World Bank Database of Political Institutions. Several data sets also focus on the rule of law or corruption, such as the International Country Risk Guide (ICRG) from Political Risk Services, La Porta et al.'s Quality of Government data, Transparency International's Corruption Perceptions index, and the World Bank's Worldwide Governance Indicators.

Recent years have also seen an increasing emphasis on individual political leaders, rather than focusing broadly on regime types. The Archigos project provides extensive data on world political leaders since 1875 with such information as the leader's dates of birth and death, the start and end of each term in office, and the leader's fate after leaving office. An earlier effort was Bueno de Mesquita and Siverson's leadership data, now included as part of the Logic of Political Survival data discussed above. Barbara Geddes' “Authoritarian Breakdown” data, Cheibub et al.'s “Democracy and Dictatorship Revisited” data, Hadenius and Teorell's Authoritarian Regimes data, Peceny et al.'s “Dictatorial Peace” data, and Weeks' “Autocratic Audience Costs” data further distinguish between different types of authoritarian leaders, which would be treated interchangeably in a single “non-democratic” category in most political data sets.

Power/Capability Data Sets

Many studies include measures of states' “power” or capabilities as either independent variables or control variables. Measures of economic capabilities were discussed earlier, but overall state capabilities are typically measured using the COW National Material Capabilities data set (Singer 1987). This data set includes information for each state on six indicators of national material capabilities: two demographic (total population and urban population), two military (military personnel and military expenditures), and two industrial (iron/steel production and energy consumption). While one or more of the six indicators are sometimes used individually, they are most often combined in the form of the Composite Index of National Capabilities (CINC), an index that measures the overall proportion of capabilities in the entire interstate system held by a given state in the year of observation. An alternative is Arbetman and Kugler's (1997) Relative Political Capacity measure, which measures a government's ability to extract resources from the population relative to other governments at similar levels of socioeconomic development.

Social and Demographic Data Sets

A number of studies have used socioeconomic development data. Many of the sources for this data have already been discussed earlier with the economic data sets, such as the World Bank's World Development Indicators. Other data sets that focus more on social factors in development than on primarily economic factors include the UN World Urbanization Prospects data (formerly known as the World Urbanization Report) and the WHO World Health Report. International migration is also covered by the OECD International Migration databases and World Bank Migration and Remittances data.

A number of recent studies have sought to study the cultural makeup of each state to help understand civil war or ethnic conflict. Ethnic, linguistic, and/or religious minority groups are often identified using the Minorities at Risk data set, the Soviet-era Atlas Narodov Mira, and even the CIA World Factbook. The Ethnic Power Relations (EPR) data and its geo-coded version GeoEPR, Fearon and Laitin's (2003) data, the Geo-Referencing of Ethnic Groups (GREG) data set (Weidmann et al. 2010), and Roeder's Ethnolinguistic Fractionalization (ELF) data are widely used compilations based on sources such as these. The COW project also has a cultural data set that has been used in several studies, although the project is still working on cleaning and improving the data set before releasing it for general public use.

Survey Data Sets

Table 3.2 indicated that 60 quantitative studies in the period covered by this overview have studied political attitudes or beliefs using survey data. Some of these articles used original surveys that had been collected with a specific purpose in mind, such as surveys of former civil war combatants in a single country. Many of these survey-based articles, though, relied on widely available cross-national survey data sets.

The Chicago Council of Foreign Relations Public Opinion Survey dates to 1975 and is conducted every other year, with additional topical surveys being conducted on a less regular schedule. This survey typically includes a representative sample of US adults, who are asked questions regarding foreign policy beliefs and opinions. Many of the surveys also include samples of adults in selected countries that are relevant to the theme of the particularly survey, and many of the earlier surveys also included samples of Americans in leadership positions in government, business, the media, academia, labor unions, churches, and interest groups.

The Eurobarometer survey dates to 1973 and is conducted in European Union member states. The standard Eurobarometer survey is conducted between two and five times each year, with additional surveys being conducted when circumstances require; there were also several surveys of citizens in prospective new EU member states between 2001 and 2004. Similar surveys are conducted in other regions, notably the Afrobarometer for African states and the Latinobarometer for Latin American states. The World Values Survey has been conducted in a series of waves since 1981, and includes a broader range of countries than the Eurobarometer or Chicago surveys. The initial wave covered 20 mostly European countries, but coverage has expanded over time, and the most recent wave was conducted in more than 50 countries representing every region of the world.

Treaty, Institution, and International Law Data Sets

The discussion of recent trends noted that this five year period witnessed a rapid rise in the number of articles focusing on international treaties, norms, and institutions. This rapid rise has been made possible, at least in part, by the availability of the following data sets. At the most general level, several data sets include information on state membership in a large number of international organizations. The COW International Governmental Organizations data (Pevehouse et al. 2004) includes membership in approximately 500 IGOs dating from the start of the COW system in 1816. The Formal Intergovernmental Organization (FIGO) data set and Koremenos' Continent of International Law data includes more detailed information about selected organizations' institutional design.

Other data sets focus on specific types of treaties or organizations. The COW Alliance data (Gibler and Sarkees 2004) and Alliance Treaty Obligations and Provisions (ATOP) data set (Leeds et al. 2002) both list formal military alliances, and the ATOP data includes information about the specific obligations and provisions in each one. The ICOW Multilateral Treaties of Pacific Settlement (MTOPS) data set includes information on membership in several dozens of multilateral treaties or institutions that obligate members to settle their disputes peacefully or to refrain from challenging each other's territorial integrity. The Transboundary Freshwater Dispute Database project has been used to study the signing, design, and impact of treaties over international rivers. Morrow and Jo's (2006) Laws of War data also addresses state compliance with a number of international laws of war, ranging from treatment of prisoners or civilians to the use of chemical or biological weapons.

Finally, several data sets use treaty or organization information to measure the relationship or foreign policy similarity between states. The Expected Utility Generation and Data Management Program, EUGene, uses alliance information to measure country risk scores and expected utility for conflict following procedures from Bueno de Mesquita's War Trap and Bueno de Mesquita and Lalman's War and Reason. Voeten's “Documenting Votes in the UN General Assembly” data and Gartzke and Jo's “Affinity of Nations” data use the similarity in UN voting for a similar purpose.

Advice for Data Users

This chapter has addressed a number of patterns and trends in the scientific study of international processes. The most prominent data sets used in this research have also been discussed. I conclude with advice for data users regarding the choices that must be made in one's own research.

Choose The Most Appropriate Data Set(s)

Perhaps the most important point is the need to evaluate potential data sets carefully to determine which one is most relevant for one's specific purpose. Many key concepts – such as armed conflict between or within states, democracy or other political characteristics of states, wealth/development or other economic characteristics of states, trade, and military alliances – have been addressed by multiple data sets. Each tends to approach the topic somewhat differently, emphasizing different dimensions of the underlying concept or collecting the data in different ways. The data user should look carefully at the definitions, coding rules, and data sources that are used to compile each data set in order to understand which best suits his or her scholarly needs.

A well-known example is the use of event data to measure interstate conflict and/or cooperation. A 1983 symposium in International Studies Quarterly noted a number of differences between the COPDAB and WEIS data sets, which were the two most widely used event data sets at the time. Howell (1983) began by noting that the data sets were collected somewhat differently, with WEIS using a single source (the New York Times) to identify events in 63 nominal event categories, while COPDAB used numerous sources from around the world to identify events in 15 ranked categories. Whether because of the different sources or other differences in the coding scheme, Howell found substantial differences between the data sets even in such basic questions as whether the level of US-Soviet conflict or cooperation was increasing or decreasing over time. Exploring further, Vincent (1983) noted that WEIS (with its reliance on the New York Times) tended to include relatively more events involving the major powers, Europe, or Asia, while COPDAB (with its more global source list) tended to include relatively more events in the Middle East, Africa, and Latin America. McClelland (1983) agreed that there can be important source effects; the proportion of all world events initiated by the United States is more than twice as high in the standard New York Times-based WEIS data file as in an alternative file collected from the Times of London. McClelland's advice to data users remains relevant more than 25 years later: “Let the user beware.”

Similarly, scholars today have many data sets to choose from in studying armed conflict, not all of which might be appropriate for a particular purpose. If the goal is to examine the impact of interstate conflict on internal political or economic processes, it may be best to choose one of the conflict data sets that uses a higher severity threshold, or at least to exclude cases with the lowest severity levels. The MID data include a number of cases that remained limited to an unreciprocated threat to use military force, which should hardly be expected to affect the outcome of an election or to weaken the economy; it is much more plausible that a full-scale war should have these effects. On the other hand, tests of arguments that domestic political or economic troubles might prompt leaders to divert attention through foreign conflict might be tested best using data with a lower threshold; it seems unlikely that a leader facing domestic challenges would seek to divert attention from those problems through a long, bloody war.

Finally, convenience should not be the main criterion for the choice of data sets to measure an important concept. It was not long ago that most data sets could only be obtained from the Inter-University Consortium for Political and Social Research (ICPSR), by corresponding with the author, or by copying the data from a friend or colleague (with little guarantee that the copied version was the latest official version). Today, most data sets may be downloaded from the collectors' established websites, with high confidence that the downloaded version is the most complete and correct version that is currently available. Yet the Internet has also made it easy for users to access other scholars' replication data sets; a number of the articles reviewed for this chapter obtained most or all of their data from single sources such as the Russett and Oneal (2001) interstate conflict data or the Fearon and Laitin (2003) civil war data. There are circumstances when it is desirable to use such a data set, most notably when one is replicating the earlier study in every respect but one and wants to be sure that nothing else has changed. But for other purposes, users would do well to identify the most appropriate data set(s) for their specific needs, rather than simply using the same data set that other scholars used for what may have been very different purposes. This is particularly true when replication data sets include out-of-date versions of many of the variables that are used, since replication data sets by their nature are not updated every time one of the component data sets is updated.

Know Your Data

Once the needed data sets have been selected, the user must become familiar with what the data set includes and how its central concepts are measured. Returning to the event data symposium discussed earlier, McClelland (1983) argued that users should be more familiar with the data that they are using. For example, many of the events included in WEIS (over one-third of the events in the period studied by Howell) are “comment” or “consult” events that do not directly involve either cooperation or conflict – yet these cases are often included in analyses of cooperation and conflict. Users who are unfamiliar with the data set may well try to include such observations in a category where they do not belong, which may be responsible for problems like the apparent discrepancies in patterns of conflict and cooperation that had been observed between the COPDAB and WEIS data sets.

The widely used CINC capability score is another example of a measure that users need to understand carefully. CINC refers to the percentage of the entire system's capabilities held by a state in a given year, and is widely used to measure state capabilities. This is useful for comparing the relative capabilities of two states in a given year, when both states' capabilities are being compared to the same denominator (the total capabilities in the system that year), but it is problematic for studying change over time in a single state's capabilities. Because CINC represents the state's share of the overall system's capabilities, its value can change substantially if there is a change in the overall level of capabilities in the system, as might happen if a large state joins or leaves the system. Nearly 200 states have joined the COW interstate system since its beginning with 23 states in 1816, and a number have also left the system either briefly or permanently. While many of these changes involve countries that are so small as to have little impact on the system's total population or military personnel, the addition of China and Japan in 1860 or East and West Germany in 1954–5 represented major changes, likely creating the appearance that most states' capabilities declined in those years simply because their proportions of the overall system's capabilities declined.

A related problem concerns the stated purpose behind international agreements. Leeds et al. (2000) investigated a well-known finding that military alliances are only honored in about 25 percent of the cases where one of the allies fights a war. While earlier research concluded that alliances are generally unreliable, Leeds et al. examined the specific security commitments made as part of each alliance, noting that many alliance treaties do not call for members to join their allies in war except under very specific circumstances. When these commitments are considered, alliance obligations have been carried out 74.5 percent of the time – triple the reliability rate found in earlier work.

Another example involves the study of conflict management. The ICOW issue data sets include data on negotiations over territorial, river, and maritime claims, which can be very useful for the study of peaceful conflict management. Three different types of negotiations are included, though: substantive negotiations that address the main substance of the issue (such as the question of territorial sovereignty), procedural negotiations that address future moves to be made in resolving the issue (such as negotiations over submitting the issue to a specific third party, which do not make any effort to address the substantive issue directly), and functional negotiations that concern the usage of the claimed territory, river, or maritime zone but not its ultimate disposition (such as attempts to demilitarize the area or to share any mineral wealth from it, rather than settling the ultimate sovereignty question). All three types of negotiations can and should be studied for some questions, such as the relative likelihood that each type can lead to a treaty or agreement, or the impact of failed negotiations or failed agreements on subsequent relations between the countries. Yet for studying the likelihood that negotiations will end the entire contentious issue, only substantive negotiations are relevant; neither procedural nor functional negotiations are capable of ending the issue, and including all three types in the analysis may produce misleading results.

The central point here is that users should be sure to read each data set's codebook and documentation, and ideally earlier papers written by those who know the data best, before using it. Making this effort will reduce the likelihood of making an embarrassing mistake when using the data. Reading the documentation carefully will also typically give the user a better sense of how the data set is organized and how it was collected, which may be very helpful in planning one's analyses.

Multipurpose Data Sources

The general point of “know your data” is especially relevant for users who obtain their data from sources that include a variety of materials, such as the replication data sets discussed above or the EUGene software program (Bennett and Stam 2000). Such sources greatly simplify the process of producing a single file with most or all of the needed data for one's analyses, without forcing the user to download each data set individually and then merge all of the needed together into a single file. Yet as noted earlier, replication data sets may not have been updated since the original book or article was published, meaning that users employing these data sets may be limited to data that has been extended or improved substantially since that time.

EUGene is updated regularly, so outdated data is unlikely to be a problem with data sets assembled in this way. Yet while EUGene appears to be a simple solution to data problems, it is a very complex piece of software, with many assumptions and decisions that must be made before a data set can be produced. Users are strongly advised to read EUGene's complete documentation file – which currently covers more than 90 pages – and to think carefully about each of the options that are available (such as the treatment of ongoing dispute years or the distinction between initiators and joiners) before using the software to create and use their data files. This will minimize the risk that the resulting data set will include or exclude cases or variables in a way that the user did not expect, providing greater confidence that the generated data set is actually appropriate for the intended usage.

Investigate Missing Data

A related point is that the user should be sure that missing values are handled appropriately. This means that the user should always run descriptive statistics to make sure that missing values are not mistakenly included in the analyses. Many data sets code missing values using negative numbers such as – 9 that would not make any substantive sense for the variable being measured. If such missing values are not recognized by the statistical software as missing, though, they may be included in the analysis – producing very strange results.

Many analyses involve the merging of numerous data sets, which may come from different sources; the articles reviewed for this chapter averaged more than six different data sets each. After merging together the needed information from all of these data sets, it is important to run descriptive analyses to note the number of cases with missing data for each variable. When there are large numbers of cases with missing data on one or more variables, the user should investigate further, to determine whether there is a systematic pattern that explains which cases are missing the needed variable. For example, the Polity IV data set only includes states that have a population of 500,000 by the end of the period of data collection, which included 162 countries in the version of the data set running through the end of 2007. The Correlates of War interstate system in 2007 included an additional 31 entities that were recognized as states but did not reach this population threshold, all of which would automatically be dropped from analysis due to missing Polity data. Unsuspecting users who use Polity IV to measure regime type in their studies will thus have no information about approximately one-sixth of the states in the COW interstate system, which may be systematically different from the larger, often wealthier, and often more conflict-prone states that are included in both lists. This may not be a matter of concern for every user, as many users may find the inclusion of such small states to be troublesome in its own right, but it is something that users should be aware of. Other data sets, particularly on economic matters, may exclude all communist economies or non-recognized states or entities such as Taiwan. The exclusion of such states may have important consequences for one's results, so the user should be sure to explore missing data points to determine whether there are any systematic patterns in the states or dyads that have missing data.

Similarly, many social science data sets use the standard COW numbering scheme to identify countries, but this scheme is not universally used, particularly for data sets collected by economists or by international organizations. Alternative numbering schemes include the ISO-3166-1 international standard, which includes lists of countries and territories by three-digit numeric codes as well as by two- or three-digit alphanumeric codes (all of which differ from the COW codes for the same countries), as well as collections that list countries by name rather than by a code number. While such differences can usually be handled relatively easily, some data sets may have different ways of handling countries that unified or divided over time. Care must be taken to make sure that Vietnam, Yemen, Germany, the Soviet Union/Russia, and similar cases are merged correctly; for example, if the name “Vietnam” is used before 1975, does this refer to North Vietnam, South Vietnam, or the aggregate of the two? Economic and environmental data sets may also combine countries into a single entry (most notably for Belgium and Luxembourg, although some sources may attempt to provide a single entry for entities like pre-1990 Germany or the post-1991 Soviet Union for the purposes of data continuity over time). Data sets may also split a single country into multiple entries (such as providing separate data for post-1997 Hong Kong and post-1999 Macao rather than including their figures with the rest of China, providing separate data for Zanzibar since its merger into Tanzania, or estimating separate pre-1991 totals for each individual Soviet republic). When using data sets from different sources, the user should be sure to check the correspondence between cases, to determine if any of these issues might be causing certain countries to be left out entirely or to include either too much data (as with Belgium-Luxembourg) or too little (as with Tanzania without Zanzibar or China without Hong Kong).

While these issues so far are related to intentional choices during the creation of a data set, data may also be more likely to be missing for certain types of countries, which could be related to other phenomena of interest in systematic ways. Lemke (2003), for example, notes that African states are much more likely than other states to have missing data on political regime type or on national power capabilities, and that some of the non-missing observations for African states are of lower data quality than is typical elsewhere. He speculates that this problem of missing and low-quality data may be at least partly responsible for an apparent “African peace” effect, in which African states appear to engage in less interstate conflict than would otherwise be expected. Ross (2006) also notes that states are more likely to have missing data on socioeconomic factors when they are less democratic and unconstrained by IMF agreements, with the consequence that many studies of socioeconomic relationships that do not attempt to account for this may be producing misleading results. While there may not always be an easy solution to any of these problems of missing data, as there may not be reasonably accurate estimates in any alternative data sets, the user should at least be aware of any systematic patterns in missing data and any possible implications for research findings.

Assess Robustness across Multiple Data Sets

Finally, while choosing the most appropriate data set for one's purposes is vital, it is also desirable to assess the robustness of one's results across alternative data sources where possible. The easy availability of multiple data sets for many important concepts makes it much easier to run the same analyses with several different data sets. If the central results do not change, then the scholar can have greater confidence that the results are robust and do not depend on quirks in a single data set. If the results do change, then the scholar should investigate more closely to determine what is different about the data sets. They may have different conceptualizations of the central concept, as with the different rivalry data sets discussed above. They may have used different sources to measure the concept, as with the COPDAB and WEIS data sets discussed above. They may also have different spatial-temporal domains, as might be the case if one covers a longer time span or one is limited to a certain type of actor.

Numerous articles have used multiple data sets to evaluate the robustness of results. While there may be substantial differences in measurement of rivalry between dispute density and perceptual approaches, for example, past studies have not identified many systematic differences in results between these approaches. Colaresi (2004) and Thies (2005) both report very similar results when using the Goertz/Diehl and Thompson rivalry data sets. There is also reason to believe that the specific data set that is used to measure trade should make a difference in results, given the number of divergent coding decisions (e.g., Gleditsch 2002; Keshk et al. 2004). Yet Keshk et al. find little difference in their ultimate result when using the Gleditsch data instead of their own, and Benson (2005) notes that an even more important source of different results is the sample of cases that is used; while several major trade data sets may have produced different results in prominent studies, very similar results are produced with each data set when using the full set of all dyads (rather than “politically relevant dyads”) and controlling for the size of their economies.

In other cases, though, running the same analysis with different data sets has identified important differences. Leeds et al. (2002) and Gibler and Sarkees (2004) note that there can be substantial differences when the same analyses are run using the COW and ATOP alliance data sets. These differences seem to stem from the respective data sets' coding rules, particularly with respect to the handling of several large multilateral alliances (Gibler and Sarkees 2004: 219). Similarly, Hegre and Sambanis (2006) note that several empirical results in the civil war onset literature are not very robust, depending on the specific data sets and measures being used. In cases such as these where different data sets produce different substantive conclusions, the user must be especially careful to decide which data set is more appropriate for the phenomenon being studied.

References

Arbetman, M., and Kugler, J. 1997. Political Capacity and Economic Behavior. Boulder, CO: Westview Press.Find this resource:

Barbieri, K., Keshk, O.M.G., and Pollins, B. 2009. Trading Data: Evaluating our Assumptions and Coding Rules. Conflict Management and Peace Science (26), 5471–91.Find this resource:

Bennett, D.S., and Stam, A. 2000. EUGene: A Conceptual Manual. International Interactions (26), 179–204.Find this resource:

Benson, M.A. 2005. The Relevance of Politically Relevant Dyads in the Study of Interdependence and Dyadic Disputes. Conflict Management and Peace Science (22), 113–33.Find this resource:

Bercovitch, J., and Fretter, J. 2004. Regional Guide to International Conflict and Management from 1945 to 2003. Washington, DC: CQ Press.Find this resource:

Bond, D., Bond, J., Oh, C., Jenkins, J.C., and Taylor, C.L. 2003. Integrated Data for Events Analysis (IDEA): An Event Typology for Automated Events Data Development. Journal of Peace Research (40), 733–45.Find this resource:

Bond, D., Jenkins, J.C., Taylor, C.L., and Schock, K. 1997. Mapping Mass Political Conflict and Civil Society: Issues and Prospects for the Automated Development of Event Data. Journal of Conflict Resolution (41), 553–79.Find this resource:

Brecher, M., and Wilkenfeld, J. 2000. A Study of Crisis. Ann Arbor: University of Michigan Press.Find this resource:

Bremer, S.A. 1992. Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War, 1816–1965. Journal of Conflict Resolution (36), 309–41.Find this resource:

Bueno de Mesquita, B., Smith, A., Siverson, R.M., and Morrow, J.D. 2003. The Logic of Political Survival. Cambridge, MA: MIT Press.Find this resource:

Buhaug, H., and Lujala, P. 2005. Accounting for Scale: Measuring Geography in Quantitative Studies of Civil War. Political Geography (24), 399–418.Find this resource:

Cingranelli, D.L., and Richards, D.L. 1999. Measuring the Level, Pattern, and Sequence of Government Respect for Physical Integrity Rights. International Studies Quarterly (43), 407–17.Find this resource:

Colaresi, M. 2004. When Doves Cry: International Rivalry, Unreciprocated Cooperation, and Leadership Turnover. American Journal of Political Science (48), 555–70.Find this resource:

Cukierman, A., and Webb, S. 1995. Political Influence on the Central Bank-International Evidence. World Bank Economic Review (9), 397–423.Find this resource:

Doyle, M., and Sambanis, N. 2000. International Peacebuilding: A Theoretical and Quantitative Analysis. American Political Science Review (94), 779–801.Find this resource:

Fearon, J.D., and Laitin, D. 2003. Ethnicity, Insurgency, and Civil War. American Political Science Review (97), 75–90.Find this resource:

Furlong, K., and Gleditsch, N.P. 2003. The Boundary Dataset. Conflict Management and Peace Science (29), 93–117.Find this resource:

Ghosn, F., Palmer, G., and Bremer, S.A. 2004. The MID3 Data Set, 1993–2001: Procedures, Coding Rules, and Description. Conflict Management and Peace Science (21), 133–54.Find this resource:

Gibler, D.M., and Sarkees, M.R. 2004. Measuring Alliances: The Correlates of War Formal Interstate Alliance Dataset, 1816–2000. Journal of Peace Research (41), 211–22.Find this resource:

Gibney, M., and Dalton, M. 1996. The Political Terror Scale. Policy Studies and Developing Nations (4), 73–84.Find this resource:

Gilmore, E., Gleditsch, N.P., Lujala, P., and Rød, J.K. 2005. Conflict Diamonds: A New Dataset. Conflict Management and Peace Science (22), 257–92.Find this resource:

Gleditsch, K.S. 2002. Expanded trade and GDP data. Journal of Conflict Resolution (46), 712–24.Find this resource:

Gleditsch, K.S., and Ward, M.D. 2001. Measuring Space: A Minimum Distance Database. Journal of Peace Research (38), 749–68.Find this resource:

Gleditsch, N.P., Furlong, K., Hegre, H., Lacina, B.A. and Owen, T. 2006. Conflicts over Shared Rivers: Resource Wars or Fuzzy Boundaries? Political Geography (25), 361–82.Find this resource:

Gleditsch, N.P., Wallensteen, P., Eriksson, M., Sollenberg, M., and Strand, H. 2002. Armed Conflict 1946–2001: A New Dataset. Journal of Peace Research (39), 615–37.Find this resource:

Goertz, G., and Diehl, P.F. 1993. Enduring Rivalries: Theoretical Constructs and Empirical Patterns. International Studies Quarterly (37), 147–71.Find this resource:

Golder, M. 2005. Democratic Electoral Systems around the World, 1946–2000. Electoral Studies (24), 103–21.Find this resource:

Hamner, J.H., and Wolf, A.T. 1997–8. Patterns in International Water Resource Treaties: The Transboundary Freshwater Dispute Database. Colorado Yearbook of International Environmental Law (9), 157–77.Find this resource:

Harbom, L., and Wallensteen, P. 2007. Armed conflict, 1989–2006. Journal of Peace Research (44), 623–34.Find this resource:

Hegre, H., and Sambanis, N. 2006. Sensitivity Analysis of Empirical Results on Civil War Onset. Journal of Conflict Resolution (50), 508–35.Find this resource:

Hensel, P.R. 2002. The More Things Change...: Recognizing and Responding to Trends in Armed Conflict. Conflict Management and Peace Science (19), 27–52.Find this resource:

Hensel, P.R., Mitchell, S.M., Sowers II, T.E., and Clayton L. Thyne, C.L. 2008. Bones of Contention: Comparing Territorial, Maritime, and River Issues. Journal of Conflict Resolution (52), 117–43.Find this resource:

Hewitt, J.J. 2005. A Crisis-Density Formulation for Identifying Rivalries. Journal of Peace Research (42), 183–200.Find this resource:

Howell, L.D. 1983. A comparative study of the WEIS and COPDAB data sets. International Studies Quarterly (27), 149–59.Find this resource:

Humphreys, M. 2005. Natural Resources, Conflict, and Conflict Resolution: Uncovering the Mechanisms. Journal of Conflict Resolution (49), 508–37.Find this resource:

Huth, P.K., and Allee, T. 2002. The Democratic Peace and Territorial Conflict in the Twentieth Century. Cambridge: Cambridge University Press.Find this resource:

Jaggers, K., and Gurr, T.R. 1995. Tracking Democracy's Third Wave with the Polity III Data. Journal of Peace Research (32), 469–82.Find this resource:

Keshk, O.M.G., Pollins, B.M., and Reuveny, R. 2004. Trade Still Follows the Flag: The Primacy of Politics in a Simultaneous Model of Interdependence and Armed Conflict. Journal of Politics (66), 1155–79.Find this resource:

King, G., and Lowe, W. 2003. An Automated Information Extraction Tool for International Conflict Data with Performance as Good as Human Coders: A Rare Events Evaluation Design. International Organization (57), 617–42.Find this resource:

Klein, J.P., Goertz, G., and Diehl, P.F. 2006. The New Rivalry Dataset: Procedures and Patterns. Journal of Peace Research (43), 331–48.Find this resource:

Lacina, B., and Gleditsch, N.P. 2005. Monitoring Trends in Global Combat: A New Dataset of Battle Deaths. European Journal of Population (21), 145–66.Find this resource:

Leeds, B.A., Long, A.G., and Mitchell, S.M. 2000. Reevaluating Alliance Reliability: Specific Threats, Specific Promises. Journal of Conflict Resolution (44), 686–99.Find this resource:

Leeds, B.A., Ritter, J.M., Mitchell, S.M., and Long, A.G. 2002. Alliance Treaty Obligations and Provisions, 1815–1944. International Interactions (28) (3), 237–60.Find this resource:

Lemke, D. 2003. African Lessons for International Relations Research. World Politics (56), 114–38.Find this resource:

Lujala, P. 2009. Deadly Combat over Natural Resources: Gems, Petroleum, Drugs, and the Severity of Armed Civil Conflict. Journal of Conflict Resolution (53), 50–71.Find this resource:

Lujala, P., Rød, J.K., and Thieme, N. 2007. Fighting over Oil: Introducing a New Dataset. Conflict Management and Peace Science (24), 239–56.Find this resource:

Marinov, N. 2005. Do Economic Sanctions Destablize Country Leaders? American Journal of Political Science (49), 564–76.Find this resource:

Marshall, M.G., Gurr, T.R., Davenport, C., and Jaggers, K. 2002. Polity IV, 1800–1999: Comments on Munck and Verkuilen. Comparative Political Studies (35), 40–5.Find this resource:

McClelland, C.A. 1983. Let The User Beware. International Studies Quarterly (27), 169–77.Find this resource:

Morrow, J.D., and Jo, H. 2006. Compliance with the Laws of War: Dataset and Coding Rules. Conflict Management and Peace Science (23), 91–113.Find this resource:

Pevehouse, J.C., Nordstrom, T., and Warnke, K. 2004. The COW–2 International Organizations Dataset Version 2.0. Conflict Management and Peace Science (21), 101–19.Find this resource:

Poe, S.C., Tate, C.N., and Keith, L.C. 1999. Repression of the Human Right to Personal Integrity Revisited: A Global Cross-National Study Covering the Years 1976–1993. International Studies Quarterly (43), 291–313.Find this resource:

Przeworski, A., Alvarez, M.E., Cheibub, J.A., and Limongi, F. 2000. Democracy and Development: Political Institutions and Well-Being in the World, 1950–1990. Cambridge: Cambridge University Press.Find this resource:

Ross, M. 2006. Is Democracy Good for the Poor? American Journal of Political Science (50), 860–74.Find this resource:

Russett, B.M., and Oneal, J. 2001. Triangulating Peace: Democracy, Interdependence, and International Organizations. New York: W.W. Norton.Find this resource:

Sarkees, M.R. 2000. The Correlates of War Data on War: An Update to 1997. Conflict Management and Peace Science (18), 123–44.Find this resource:

Singer, J.D. 1987. Reconstructing the Correlates of War Dataset on Material Capabilities of States, 1816–1985. International Interactions (14), 115–32.Find this resource:

Stinnett, D.M., Tir, J., Schafer, P., Diehl, P.F., and Gochman, C. 2002. The Correlates of War Project Direct Contiguity Data, Version 3. Conflict Management and Peace Science (19), 58–66.Find this resource:

Thies, C.G. 2005. War, Rivalry, and State Building in Latin America. American Journal of Political Science (49), 451–65.Find this resource:

Tir, J., Schafer, P., Diehl, P.F., and Goertz, G. 1998. Territorial Changes, 1816–1996: Procedures and Data. Conflict Management and Peace Science (16), 89–97.Find this resource:

Thompson, W.R. 2001. Identifying Rivals and Rivalries in World Politics. International Studies Quarterly (45), 557–86.Find this resource:

Vincent, J.E. 1983. WEIS vs. COPDAB: correspondence problems. International Studies Quarterly (27), 161–8.Find this resource:

Weidmann, N.B., Rød, J.K., and Cederman, L.E. 2010. Representing Ethnic Groups in Space: A New Dataset. Journal of Peace Research (47), 491–9.Find this resource: