As part of my PhD project on rules of debate and polarisation in the UK House of Commons, I gathered historical data covering the period 1811-2015. Some of of these data are available on this page (and more will be released soon).

Session dates for the UK House of Commons, 1811-2015

Download in JSON format

Download in XML format

Polarisation in the UK House of Commons, 1811-2015

Building on a classification accuracy approach introduced by Peterson and Spirling (2018), and my PhD work, the graph below shows levels of polarisation in the UK Parliament for the period 1811-2015. “Polarisation” is defined as how predictive language use is of party membership in a given session (measured on a 0-1 scale), and is estimated on the basis of records of parliamentary speeches from Hansard. In simple terms, the measure is generated by fitting a training algorithm to a set of parliamentary speeches in a session with party labels, and taking the accuracy from predicting a held-out subset of speeches (which are also labelled). The data for the graph can be viewed in the table below, and are also available in .csv format via this link, or in .json via this link. Please cite this paper if you intend to use these data for your research.