Data set on “Party Positions in Multiple Dimensions across Time and Space” available
05 November 2014
Estimates (or more correctly: predictions) of parties’ positions in multiple dimensions in different countries, spanning the decades from 1945 (1920 for the USA) to 2010 is now available here on my website.
The data are available in csv-format as summaries of the posterior distribution (means, standard deviations and 95 percent prediction intervals) and as draws from the posterior distribution (with 5, 100, and 2000 replications). I plan to extend the data by adding more dimensions and will also provide software for getting ‘quick-and-dirty’ estimates using a Laplace approximation, which is considerably faster than the computationally intensive Monte Carlo integration approach used for the Political Analysis article on which the published estimates are based.