the Similarity of States: Using S to Compute Dyadic Interest Similarity

Abstract

Several leading international relations theories argue that the degree of interest similarity is an important determinant of dyadic conflict and cooperation. Empirical scholarshave long wrestled with operationalizing and measuring this central, yet elusive, concept. Signorino and Ritter's (1999) S algorithm, combined with multiple data sources, provides an attractive solution to this problem. To date, however, many scholars have failed to take full advantage of this solution. In this research note we examine the properties of S via simulation and with real data sources, highlighting its virtues and potential limitations. In particular, we stress the need to include multiple data sources in the computation and provide scholars with an easy-to-use tool to greatly simplify this task.Keywords interstate dyads, interest similarity, preference similarity, Scompute Signorino and Ritter (1999) propose a methodology for calculating the similarity of two column vectors, S, which has the potential to fundamentally change the way scholars of international relations (IR) measure the key concept of dyadic interest similarity. 1 Not only is S intuitively appealing, but it allows for various weighting schemes and the easy combination of multiple variables. These features should make S attractive to IR scholars on a theoretical level. To date, however, examination and use of the new method has been limited. Many studies continue to employ other algorithms, and others, while using S, fail to take full advantage of the method. In this research note we examine the properties of S,

Strategic Demands, Credible Threats, and Economic Coercion Outcomes

“…Usually issue salience measures are based on the five-level ordinal scale used by Hufbauer, Schott, and Elliott (1990) and more or less distinguish between the traditional categories of ''high'' and ''low'' politics. Simultaneously, operational indicators of states' common foreign policy interests have become more sophisticated (Sweeney and Keshk 2005), allowing one to control for exogenous dissimilarities that might push a sender to demand more substantial policy changes from a target. For instance, sanctions between the United States and Canada typically involve disputes over trade policy whereas US sanctions against Iran are related to security issues for reasons separate from the US bargaining power vis-à-vis these two states.…”

Choice or Circumstance? Adjusting Measures of Foreign Policy Similarity for Chance Agreement

“…The 'rmac' package (Kirk 2010) in R (R Development Core Team 2011 implements the computation of the interrater-agreement indices as described by Fay (2005). 3 In principle, similarity measures can be calculated with any relational data set (Sweeney and Keshk 2005). Besides alliance data, the use of data on voting in the UN General Assembly has been popular (Gartzke 1998 The subsequent conversion of the dissimilarity into a similarity score is also the same.…”

“…1 Despite its growing popularity, few studies have subsequently examined the properties of S. While Bennett and Rupert (2003) and Sweeney and Keshk (2005) have pointed to some empirical and conceptual problems of S, they have not suggested feasible alternatives.…”

Addressing Endogeneity in Actor-Specific Network Measures

“…Self-dyads of the type i − i are commonly assigned the value of 3 (Sweeney and Keshk 2005), following the logic that a state will defend itself if attacked (Bueno de Mesquita 1975).…”