Coalition Predictions with Ordinal Policy Position Data

Abstract

The policy-oriented coalition models that were developed in the 1960s and 1970s generally required that the positions of political parties were measured at ordinal level. More recent models of coalition formation, however, have tended to assume that interval-level data on parties' policy positions are available. This need for interval-level data has raised serious practical problems for empirical research. Analysts frequently wish to examine patterns of coalition formation after several, or many, elections. Unfortunately, detailed interval-level data on parties' positions are generally available at only one or two points in time. While it can often be plausibly argued that parties will occupy fairly stable ordinal positions over time, it is extremely doubtful whether this is also the case for their interval positions.