Dimensionality in Congress
Update: A revised version of this paper, given as a poster at the 2011 Summer Meeting of the Society for Political Methodology, is available here (PDF).
In collaboration with Jacob Montgomery and John Aldrich, I am interested in understanding the relationship between observed (measured) and unobserved (true) dimensionality in Congress. In an ongoing project, we employ Monte Carlo simulations of legislative voting behavior, followed by dimensionality-reducing scaling techniques, to identify the parameters under which we might observe roll-call scalings similar to those we find in empirical data. Our findings suggest that the typical account of the dimensionality of ideology in Congress, “one-and-a-half dimensions,” may arise under a large variety of “true” dimensionality settings.
One of our papers, given at the 2010 APSA, is available for viewing or download here [PDF].
The slides from our paper presentation may be seen below.