COW COW Bibliographic Essay


The Correlates of War (COW) Project traces its intellectual origins to the pioneering work of Sorokin, Wright, and Richardson. Yet the actual beginnings of the COW Project date to a grant proposal submitted to the Carnegie Corporation in 1963. Under the guidance of J. David Singer and Melvin Small, this grant funded an exploratory study of the conditions associated with the outbreak of war. In the years since those relatively humble origins and modest aims, the COW Project has produced over one hundred and fifty journal articles and book chapters, and a dozen or more books.

The scholarly contribution of the COW Project is evident in the extent to which it has become part of the fabric of international relations research. In reserach reported in fifteen bey journals that focus on quantitative international politics, COW was the most frequently cited data project in the period 1974 to 1986; COW accounted for thirty-one percent of the citations of the eight leading data projects (McGowan, et al., 1988). Furthermore, in a survey conducted in 1984 among one hundred and sixty-one specialists in international and comparative political research, thirty-nine percent found COW at the top of the list of currently archived data sets that "should be designated national data resources for maintenance, improvement, and expansion" (McGowan, et al., 1988: 107). The Data Development for International Research (DDIR) project, the umbrella organization for organizing data collection and integration efforts in international relations, is supporting eleven different data collection programs; perhaps not surprisingly, five of these are directly linked to the COW Project.

The COW bibliography contained within this website is not simply a list of all articles, chapters, or books that have used COW-generated data. Such a list might number enough pages to fill a journal. Although we applaud this work and are gratified that much of it has been done by leading scholars in the field of international relations (Russett, Starr, and Midlarsky to name a few), we confine our current list to published work of Correlates of War personnel and the loose collection of people called Correlates of War "Off-Campus Associates." It is hoped that a bibliography of related works may in the future by available for individuals interested in further exploring this field.

The former group is headed prominently by J. David Singer and Melvin Small, but graduate students and post-doctoral fellows at the University of Michigan who worked with the COW Project are also included. Off-Campus Associates are those scholars who have continued their research on international conflict after their formal affiliation with the University of Michigan ended and/or who work in the subject area, use COW data, and are involved (in a loose sense) in the COW enterprise.

The bibliography is not, however, an amalgamation of the vitae of the people who meet these criteria. We have only included those articles that use COW-generated data and/or reflect a topic or approach that is relevant to the COW mission of understanding the sources of conflict between states. As with works by individuals outside the COW Project, complete bibliographies from those within the COW Project may be included in the near future in order to provide additional avenues of exploration.

The COW bibliography is divided into four relatively distinct sections. The first, "Theoretical Rationale and Research Strategy," includes works concerned with the epistemological and theoretical bases of COW research. Not surprisingly, most of the early assumptions and strategies were outlined by the project director, J. David Singer; later work, reflecting such new and innovative approaches as expected utility models were developed by others including Bruce Bueno de Mesquita. Much of this work has been reprinted in the following two anthologies. Individual articles from these collections (and other COW anthologies) are cited in the bibliography for the convenience of the reader.

The second section, entitled "Constructing the Indicators and Generating the Data," includes selections that describe the theoretical and empirical bases of developing data on war and conflict. The COW Project and its team have also devoted considerable attention to deriving operational measures of various concepts such as systemic polarity and national capability. The major sources on war data are the two books by Singer and Small noted here, with many of the articles listed in the COW bibliography reprinted in the Singer and Diehl collection.

Although most of our time and financial resources at Michigan have gone into data generation, the bulk of COW published work falls into a category we call "Testing Some Models and Analyzing the Data." Data generation is not an end itself, but is a prerequisite for several other stages in research. The most important contribution that a scholarly enterprise can make is the construction and empirical testing of well-defined hypotheses and models based on such data. Appropriately, COW research has used the data to investigate conditions associated with the initiation, escalation, and outcome of conflict. A summary of central findings and theoretical directions of much of this work is found in Vasquez (1987). The relevant anthologies for this section are listed below.

Research into the origins and escalation of international conflict should not be a purely academic exercise, even though most decision makers still regard COW and other related work as such. Although the research in the previous section has relevance for policy as well as theory building, some COW work has been specifically focused on practical implications. The idea has been to discuss and apply social scientific findings to current problems. Thus, the final section of the bibliography is entitled "Practical Implications: Policy and Teaching." The two collections listed below are focused on the relevance of empirical research to policy problems.

Unlike a conventional list of references, the citations are presented in chronological order within each of the four sections. This allows the reader to note several changes and developments in COW research, reflecting the evolution of quantitative international political research.

The early years of the Correlates of War Project were conceptually oriented with a small band of scholars struggling to define a research strategy and initiate the enormous job of generating a number of new and crucial data sets. Empirical analyses were usually bivariate analyses concerned with the outbreak of war (and major power war in particular), with less attention focused on other forms of conflict. In many ways, these early studies were "brush clearing" exercises and constituted some of the first data-based work on their topics (e.g. the impact of alliance configurations on the incidence of war). Furthermore, the initial research was concentrated on the systemic level and its attributes; the belief was, and still is, that the characteristics of the international system are not only themselves critical in affecting the chances for war between states, but must serve as control variables when examining the effects of state behavior, interstates linkages, and so forth.

Over the history of the COW Project, the research output has evolved in several ways, although the basic thrust of the project has not been substantially altered. The data generation effort has concentrated more on updating existing files than developing new ones. Several exceptions, however, reveal a new direction for the project. The newly developed data sets on militarized disputes and the Behavioral Correlates of War (BCOW) indicate an extension of the original emphasis on war exclusively to a broader focus on all interstate conflicts involving the use or threat of military force.

Small and Singer also collected data on civil wars, although that data has not been utilized as extensively as was originally envisioned. The theoretical underpinnings also changed somewhat as new models, such as the expected utility model, were developed and became important off-shoots of the original COW framework.

Most COW published research now involves empirical testing instead of data generation. The empirical analyses are different both in their level of sophistication and the identity of the authors. Singer and Small conducted most of the preliminary empirical studies, but over time the list of scholars doing COW research has grown and has now reached its second generation. The kind of analyses also changed; the models tested now are more sophisticated, involving multilateral relationships. In addition, there has been a greater focus on the national level of analysis owing partly to an evolution in the kind of research questions asked but also in response to some of the empirical and theoretical limitations imposed by a focus on the systemic level.

In the over quarter century since the Correlates of War Project began, there has been appreciable progress in the social science community. Yet, the international conflict field remains only in the infancy stage, both pedagogically and intellectually, in understanding the correlates of conflict between states. It is our hope that the research listed on this site has nutured, and will continue to nurture, furture scholars interested in this vital concern.

Go to the Correlates of War Bibliography


Article courtesy of Paul F. Diehl, University of Illinois at Urbana-Champaign

References in Essay