Increasingly in political science, models are developed and tested against datasets which have quite complicated structures. Sometimes it happens that the standard statistical tests are not directly applicable. Furthermore one's intuition as to what counts as strong evidence in favour of a particular model can be misleading. In such circumstances it may be necessary to think more formally about the situation in order to establish valid interpretations of the evidence. This Note seeks to illustrate the value of such formal thinking by reanalysing the evidence in Budge's recent article on parties' policy positions.I. Budge, ‘A New Spatial Theory of Party Competition: Uncertainty, Ideology and Policy Equilibria Viewed Comparatively and Temporally’, British Journal of Political Science, 24 (1994), 443–67. This draws on H-D. Klingemann, R. I. Hofferbert and I. Budge, Parties, Policies and Democracy (Boulder, Colo.: Westview Press, 1994).