Paul V. Warwick, "Toward a Common Dimensionality in West European Policy Spaces," Party Politics, January 2002, vol. 8, no. 1, pp. 101-122.

First Paragraph:
One of the most active areas of research on liberal democracies concerns the compositions, policies and survival prospects of coalition governments in West European parliamentary systems. This literature has been profuse both in empirical studies exploring possible causal influences on coalition behavior as well as in theoretical formulations, particularly formal theoretical formulations, designed to account for coalition outcomes. Despite this activity, how much we know with reasonable certainty remains unclear: the empirical work often focuses on hypotheses and relationships rather than on fully developed theories, while the theoretical literature tends to operate at a level of complexity and abstraction that makes empirical testing diffi- cult. Ideally, what the topic needs is a stronger linkage between theory and testing.

Figures and Tables:
Table 1. Principal components analysis of Laver and Hunts eight cross-national issues
Table 2. Party system dispersion along LH components
Table 3. Principal components analysis of party manifestos (CMP) data
Table 4. Second-stage principal components analysis of LH and CMP components
Table 5. Party system dispersion along CMP components

Last Paragraph:
In addressing these reservations, the key step will be to determine how well party positions on the three dimensions posited here, and derived from the two sources utilized here, contribute to explaining parliamentary and coalition behavior. This step is not as easily broached as it might appear, for its success is conditional on our knowing which theories ought to perform well in this regard. In other words, we cant judge the data confidently until we know that we have good theories, just as we cant judge the theories convincingly until we know we have good data. What the present analysis has shown is that it may be reasonable to conceive of policy spaces in West European democratic systems as composed of various combinations of dimensions drawn from a common set of three and, secondarily, that currently available expert and manifesto data might yield good measures of party positions on these dimensions. It would certainly be a boon to the study of democratic government if the thorny issue of measuring policy spaces turns out to be so readily resolvable, but a good deal of further work will be required before we know whether this is indeed the case.