Research Design and Method in International Relations: Editors' Introduction.
Green, Kim, and Yoon submitted their article to International Organization in Fall 1999. The reviewers immediately recognized the importance and controversial nature of its argument. We had already begun to see references to the manuscript in other papers and considered it important to get the article into print. But because of the challenges it poses and concerns raised about the fixed-effects method, we decided to ask several scholars to write formal comments on the paper to be published alongside the article. Everyone accepted the task with alacrity. The comments were peer-reviewed and all the articles revised. We are all indebted to the authors in the symposium for their willingness to participate.
Although they do not accept the necessity of the fixed-effects model for international relations, John R. Oneal and Bruce Russett nonetheless demonstrate the sensitivity of Green, Kim, and Yoon's findings to time period and other methodological choices. Oneal and Russett show that with reasonable alterations in Green, Kim, and Yoon's approach, the primary findings of the so-called liberal peace, including the benefits of economic interdependence, can be recovered. Nathaniel Beck and Jonathan N. Katz develop a methodological critique, arguing that fixed effects are always inappropriate for binary dependent variables and are often inappropriate for continuous dependent variables. Although they accept Green, Kim, and Yoon's main point about the importance of unmeasured heterogeneity, they conclude that the fixed effects cure is typically worse than the disease. Gary King synthesizes the contending positions in this debate and adds his own perspective. He highlights a point made by all the authors that the real solution involves getting better and more complete data on the omitted variables. Moreover, as an interim solution, he shows that the bias introduced by unmeasured heterogeneity can often be anticipated and used to interpret better the substantive effects of key variables. King concludes with suggestions on how methodologists might devise better methods to correct the problem of unmeasured heterogeneity.
Although framed in more methodological terms than is common for articles in 10, the issues raised are of sufficient importance that we decided the symposium warranted publication. And even though the methodological issues go far beyond international relations, as King's reference to the statistical study of salamander mating habits indicates, they apply more forcefully to questions asked in world politics than perhaps in any other area of the social sciences. As King concludes, Green, Kim, and Yoon identify a fundamental problem that is underappreciated by practicing scholars. We may not yet have a solution, but it is only through active debate that progress can occur.
|Printer friendly Cite/link Email Feedback|
|Author:||Gourevitch, Peter; Lake, David A.|
|Article Type:||Brief Article|
|Date:||Mar 22, 2001|
|Previous Article:||Investing in the Peace: Economic Interdependence and International Conflict.|
|Next Article:||Dirty Pool.|