Over the past decade, congressional approval ratings have steadily dropped to the lowest in history. A Gallup Poll as recent as October 11th, 2015 states that only 13% of Americans view Congress with approval. As the rift between constituents and their representatives widens, the need for tools to closely track representatives grows. WordShoal is an algorithm developed by Benjamin E. Lauderdale (London School of Economics and Political Science) and Alexander Herzog (Clemson University) that measures political polarization by analyzing the speech patterns of senators through debates on the senate floor.
The Library of Congress records all speech on the senate floor in debates. Debates can be about anything the senate is speaking about at that particular moment and can include any number of senators. Speech positions are estimated for each of these senators by finding relationships between keywords along party lines. For instance, words like pro-life, defense, and immigration are more often spoken by Republican senators, while words like equality, pro-choice, and rights are more common in Democratic speeches. While the meaning of words change with current events and context, accumulating scores over hundreds of debates minimizes contextual variance, allowing for more confident senator speech scores.