This paper discusses the analytical problems associated with the increasing use of numerical databases in quantitative cross-national research. The authors highlight the need for a better understanding of the way these databases are constructed and their inherent ambiguities and they examine the existing critical literature. The paper makes a case study of data on military interventions and analyses the problems of classification that the authors identified in several well-known databases. The paper concludes that more attention should be given to modelling uncertainty rather than attempting to achieve clear-cut precision.
Working Paper No. 50 (series 2), London, UK; Crisis States Research Centre, 19 pp.