Analysis of Dimension Expansion in Spatial Modeling of American Voter Behavior
Understanding voter behavior has the potential to give us key insights and reflections on election outcomes and even serves useful in managing campaigns and predicting election outcomes. Yet political scientists have yet to create an accurate model of voting behavior. In fact, the most popular theories assert that there isn’t much political sophistication to understand or interpret. Spatial Theory however, asserts that there is political sophistication but in the form of an underlying ideological framework, or an Ideology. Spatial Theory seeks to understand and predict voter behavior by relying on ideological similarities between candidates and voters. But spatial theory isn’t perfect either. Spatial models may be one of the most predictive models of voter behavior, but it is undeniably missing out on some predictive feature that affects voting behavior. The goal of this research is to attempt to find these missing features and to see if they can be adjusted for. If these features exists and can be accurately adjusted for, they may complete the model. The most obvious of these potential features are common demographical features that have been known and empirically shown to have different voting behaviors as populations. I hypothesize that these demographic features affect a voter’s world view and value systems. I believe that if I adjust for the effects of these features in my measurement of ideology, my model will get closer to more accurately reflecting voter behavior.