Testing Spatial Voting Models Using Simulated Data

Markus Wagner, Thomas Meyer, David Johann

The debate on how voters use party positions to determine their vote choice has remained inconclusive. This is partly because it is difficult to distinguish empirically between rival spatial models of voting. We argue that simulations can provide important new insights. Specifically, we use simulations to see what results we estimate for each spatial model with the advantage of knowing which specific model our simulated 'voters' based their vote choice on. We therefore proceed by first simulating sets of voters who choose parties based on one specific spatial model. Then, we run rival models on these data and examine the resulting model parameters. We find that we usually underestimate the 'true' effect of spatial voting effects on vote choices. Moreover, it is impossible to distinguish empirically between directional and compensational voters. Our findings encourage researchers to test these models in experiments or to consider the theories’ observable implications.

Department of Government
Publication date
Austrian Fields of Science 2012
506014 Comparative politics
Portal url