(Technical) Partisanship, while not explicitly present in the first few United States Congresses, has been a central tenet of American democracy since its birth. With few exceptions, the United States has operated a largely two party system: Loyalists vs. Patriots, Pro-Administrations vs. Anti-Administrations, Federalists vs. Democratic-Republicans, Democrats vs. Republicans, Democrats vs. Whigs—the list continues.

To be sure, these parties did not always exist at perfectly opposite ends of the political spectrum. In the Era of Good Feelings following the War of 1812, for example, the Federalist party collapsed, leaving the United States with single-party rule under the Democratic-Republicans. And, as this WorthHiding article points out, significant variation in ideology also exists within parties.

There have been many attempts at conveying partisan behavior in quantitative terms. The most well known of these attempts is NOMINATE, a metric created by political scientists Keith Poole and Howard Rosenthal. NOMINATE uses roll-call voting behavior to calculate 'ideal points' for legislators in two dimensions.

Poole and Rosenthal argue that throughout the entire course of American political history, partisan behavior can be adequately described using only two axes. The first axis—typically assigned to the x axis when graphed—corresponds generally to fiscal behavior. Values toward 1 are more conservative, while values toward -1 are more liberal. The 2nd dimension—assigned usually to the y axis—illuminates social voting behavior. In the early Congresses, for example, the 2nd axis separated northern abolitionists from their southern counterparts.


Using a variant of NOMINATE called DW-NOMINATE, I graphed the average ideal points for each political party from the conception of the United States to the 18th Congress. Below, I have included charts and commentary which help to reveal the partisan dynamics of this early era in quantitative terms.

Purple bubbles indicate parties in the House of Representatives. Orange bubbles indicate parties in the Senate. The size of the bubbles (by radius) designate relative size to the other parties in Congress at that time.

spectrum_1
In the 1st Congresss of the United States, the Senate parties fall below their House counterparts on the 2nd DW-NOMINATE dimension, perhaps indicating a difference in ideology between the state legislatures that elected the Senators and the white landowning males who elected the members of the House. While there is a clear separation between the parties in the 1st dimension, none is present in the 2nd.

spectrum_2
In the 2nd Congress, a more obvious divide emerges between the parties in the 2nd dimension. (Pardon the text overlap—STATA is not wonderful!)

As the following figures show, however, the balance of membership was tilted only slightly towards the Pro-Administrations.
house_2
senate_2

spectrum_4
Skipping to the 4th Congress, we see that the Anti-Administrations became the Democratic-Republicans and the Pro-Administrations became the Federalists. Their partisan divides remain do not change significantly.

spectrum_5
In the 5th Congress, the Federalists of the House and Senate nearly perfectly converge—so much so that their respective labels are hardly readable.

spectrum_9
In the 9th Congress, the growth of the Democratic-Republicans becomes visually evident.

spectrum_16
In the 16th Congress, the Democratic-Republicans continue to grow—and the House Federalists find themselves close in voting behavior to the Democratic-Republicans.

The imbalance in party membership becomes more pronounced:
house_16
senate_16

In the 18th Congress, however, a great fracture occurs. The Election of 1824 had no clear nominee from the Democratic-Republicans, leading to political turmoil:
house_18
senate_18

Mapped ideologically, the 18th Congress looks like the following:
spectrum_18

A mess!


Cover image: an artists rendition of the United States Capitol in 1800, prior to being bruned by the British.