A Flexible, Unified, Normalized Euclidean (FUN) Voter Utility Function

This paper proposes a new unified voter utility function with both proximity and directional components. This improved unified spatial voter utility function enables researchers to predict voter choices from survey data and test the relative explanatory power of theories of voting behavior measured by opinion or exit poll surveys by combining survey question results. Decision-making processes may be flexibly varied by weighting individual questions, even, if desired, by individual voter. The utility function is normalized to facilitate comparisons between voter groups and to estimate the likelihood of voting, able to incorporate, we believe, virtually any voting behavior theory. Tested against a subset of 2012 ANES time series data, the model performs well