International Organization

Research Notes

Spatial Effects in Dyadic Data

Eric Neumayera1 and Thomas Plümpera2

a1 Department of Geography and Environment, London School of Economics, and the Centre for the Study of Civil War, International Peace Research Institute Oslo (PRIO). E-mail: [email protected]

a2 Government Department, University of Essex, UK, and the Centre for the Study of Civil War, International Peace Research Institute Oslo (PRIO). E-mail: [email protected]


Political units often spatially depend in their policy choices on other units. This also holds in dyadic settings where, as in much of international relations research, analysis focuses on the interaction or relation between a pair or dyad of two political units. Yet, with few exceptions, social scientists have analyzed contagion in monadic datasets only, consisting of individual political units. This article categorizes all possible forms of spatial effect modeling in both undirected and directed dyadic data, where it is possible to distinguish the source and the target of interaction (for example, exporter/importer, aggressor/victim, and so on). This approach enables scholars to formulate and test novel mechanisms of contagion, thus ideally paving the way for studies analyzing spatial dependence between dyads of political units. To illustrate the modeling flexibility gained from an understanding of the full set of specification options for spatial effects in dyadic data, we examine the diffusion of bilateral investment treaties between developed and developing countries, building and extending on Elkins, Guzman, and Simmons's 2006 study. However, we come to different conclusions about the channels through which bilateral investment treaties diffuse. Rather than a capital-importing country being influenced by the total number of BITs signed by other capital importers, as modeled in their original article, we find that a capital-importing country is more likely to sign a BIT with a capital exporter only if other competing capital importers have signed BITs with this very same capital exporter. Similarly, other capital exporters' BITs with a specific capital importer influence an exporter's incentive to agree on a BIT with the very same capital importer.


This article is based on equal authorship. We thank Zach Elkins for providing the source data for the weighting matrices used in this article and thank Steve Gibbons, Ian Gordon, Rob Franzese, and two referees for many helpful comments. Stata ado-files generating spatial effect variables for any of the forms of contagion listed in using basic link functions can be downloaded from xs27E8 or xs27E8 For ado-files that generate spatial lags with more complex link functions, contact the authors.