Designing Quantitative Historical Social Inquiry: An Introduction to Stan

Abstract

This paper provides an overview of newly-developed software (called Stan) to sample from an arbitrary posterior distribution using a variant of Hamiltonian Monte Carlo. We use Stan to replicate and extend part of the analysis in Cox and Katz (2007) on the relationship between votes and seats in the U.S. House of Representatives. Together, Cox and Katz (2007) and Stan illustrate a new way forward to combine historical analysis with quantitative modeling as recommended in Wawro and Katznelson (2012).