International Cooperation and Regional Conflicts

Download 72 Kb.
Size72 Kb.
  1   2   3   4

International Cooperation and Regional Conflicts

in the post-Cold War World: Events Data for 12 Regional Conflicts, 1987-1999
Jon C. Pevehouse and Joshua S. Goldstein
I. Project Overview
This data set was generated for the purpose of investigating interactions in regional conflicts. Specifically, we were interested in the nature of reciprocity in regional conflicts. During the Cold War, the question of international cooperation and reciprocity revolved around U.S.-Soviet relations. In the 1990s and in the new millennium, the context for international security cooperation has shifted to regional conflicts. U.S. policymakers must revisit the question of regional cooperation and conflict on an almost daily basis. This can either occur in the context of direct U.S. relations with another state (such as Iraq, Iran, or Cuba) or in reference to a group of regional actors whom the U.S. wants to influence (Israelis & Palestinians, North & South Koreans, or Chinese & Taiwanese). The question of encouraging cooperation with the U.S. or with other regional actors is important to both policy makers and international relations theorists.
The traditional two-player game of the Cold War, which underlies most theoretical work on international cooperation, is usually inadequate to understand the dynamics of regional interaction. Not only does the larger number of players in a region make the strategies and reactions more complex, but large power differences may also change the nature of interaction. International relations literature has little to say about these non-major power interactions, especially in terms of empirical research on influence strategies and responses.
Over the last two years, our project has traced events in several regional conflicts including the former Yugoslavia, the Middle East, Somalia, Haiti, Cuba, the Korean peninsula, India-Pakistan, and China-Taiwan. We have generated this data set to test several hypotheses related to cooperation and regional conflict (for a review of these hypotheses and our findings to date, see Goldstein and Pevehouse 1997; Goldstein, Pevehouse, Gerner, and Telhami 1998; Pevehouse and Goldstein 1999a).
The files in this data set are organized as follows: all files ending in .events contain the actual events data as coded by KEDS. Files ending in .actors are the actor lists for each region. Each region contains some specialized actor codes which are described below as are the basic actor codes common to each data set. Files ending in .verbs are the verb patterns which code to a WEIS category. The WEIS codes utilized are listed below. Finally, files ending in .options and .class are KEDS preference files which are fully described in the KEDS manual. If your only interest is in using the events files to create time series or event count data, the .events files will be the only relevant files. To convert these to time series or event count data, use the KEDS COUNT program, available at the KEDS web site.
This data was generated using the Kansas Events Data System (KEDS). This program is free and available to the public at the KEDS web site: Details of the program (including the users manual) are available from that site. This codebook will present a very brief overview of how KEDS generates the data contained in these files.
KEDS uses a sparse parsing technique to code events data from machine-readable text. For this project, we have coded the leads (the first sentence) of Reuters newswire stories exclusively (on the merits and problems with this source see Huxtable and Pevehouse 1996). KEDS parses the subject (source), verb (event/action), and object (target) of a lead to discover who did what to whom. KEDS uses two dictionaries to code these actors and events: a verb and an actor dictionary. Each verb and verb pattern is tied to a WEIS code (see next section) which classifies the event/interaction into a nominal category. The actor dictionary contains persons, organizations, and other international actors which are coded into a three to five letter coding scheme (see next section). Each file labeled “.events” in this data set are the output from coding the Reuters stories on that particular region during the time period indicated. The dictionary files for each region are included as well. Actor dictionaries end in “.actor”, verb files end in “.verbs”. The files ending in “.options” and “.class” are KEDS preference files which are more fully described in the KEDS manual.
III. WEIS Codes and Actor Codes

Share with your friends:
  1   2   3   4

The database is protected by copyright © 2020
send message

    Main page