Study parliamentary activity in the absence of individual voting data.

Summary When attempting to account for voting behavior in parliamentary bodies, researchers are often forced to collect incomplete data, as the practice of recording and publishing individual voting results is far from widespread across and even within institutions. While the actors present and the final decision are known, the distribution of votes and, consequently, the motivations, behaviors, and attitudes of individual voters are generally unknown. The aim of this article is to propose a range of solutions to this type of difficulty, adapting them specifically to the case of parliamentary activity. To do so, a multivariate probit model developed by Moritz Marbach (2015, 2017) is used. Placing himself in a Bayesian framework, he proposes to derive the distributions of voting choices from the exact stochastic distribution of decision data. In this first version of a paper in progress, we revisit solutions in the literature to the problem of the absence of individual voting data in parliaments. The properties of the model are then introduced. A concluding section reviews its limitations and application prospects.