Polarization is increasing. How does this play out in the Senate?

_Code for this project can be found on _GithubIn our last article, we showed empirically that polarization in the American Congress in increasing. In this article, we dive deeper and explore these trends in the Senate specifically. We specify another metric, party loyalty, as a proxy value to polarization. This metric helps us further confirm our empirical findings on the trend of polarization and to answer the question: are individual Senators becoming more loyal to their party? Do Senators who are more loyal to their party stay in power longer than those who reach across the aisle?Party loyalty is defined as the percentage of Senators of a given party which votes in line with the majority of the party. To illustrate the concept we take the vote for the Affordable Care Act as an example. In this vote all Democrats voted for it, while all Republicans voted against it. The party loyalty is therefore 100% for both parties.As can be seen from the screenshot below, which shows this result of the vote for the Affordable Care Act, we can also see the name and state of each Senator. That allows us to investigate the party loyalty for each Senator and state. The website provides us with all bills voted on in the Senate between 1989 and 2020. Having data for over 30 years gives us the possibility to investigate the dynamics over time.

Image for post

In order to calculate the average party loyalty for a given year, we proceed as follows: for every bill a Senator faces, we assign them the value1 if the voted with party majority, and a zero if they did not. Afterwards we average all these values a Senator accumulated over a years time. With that we obtain the average party loyalty of an individual Senator. If we now average the values of all Senators within a party, we get the average party loyalty of the entire party within a year. Below, a mathematical notation for this concept is given.

Image for post

Furthermore, it is important to talk about what the numerical boundaries of our concept of party loyalty are. From our Affordable Care Act example, it should be clear that the maximum average party loyalty across all Senators of one party is equal to 100% or 1. That is because everybody voted in line with the majority of the party.The lower boundary of the average party loyalty across all Senators of one party, though, cannot be any lower than 0.5. That is because how we define a party’s majority. A majority is defined as having more people voting for it than against it.Assuming a positive even amount of Senators, N, within one party, and given the discrete nature of votes, the minimum amount of votes needed for a majority is equal to (N/2 + 1). All these Senators will, following our definition, be assigned a value of one, whereas the rest (N/2 - 1) gets assigned the zero. If we now calculate the average across all Senators we get the following equation which shows that the average value is always going to be strictly above 0.5. The mathematical notation below also shows that this holds for an uneven number of Senators within a party.

Image for post

Of course we also have to address the case of an even amount of Senators where exactly half of the Senate votes one way, and the rest votes the otherway. In that case we manually set the party loyalty to 0.5.

Average Party Loyalty across all Senators and Parties over TimeBefore looking into how the average party loyalty developed within each party, we start by looking into the development average party loyalty across Senators and parties. This figure is obtained by averaging the party loyalty figure of each Senator over both parties for a given year.

#data-visualization #politics #statistics #america

Polarization in the Senate
1.35 GEEK