Supreme court justice agreement

NOTE: Portfolio piece pulled and republished from my old blog
Purpose: Show overlap in voting records and decision splits

Data source: Supreme Court Database (Justice centered rows)

Date retrieved: 07/05/2015 (data for 11/8/2010 to 07/01/2014)

Tools Used: R (ggplot2), Excel (colored table)

R script and data: 715_supreme_court overlap R

Full dataset and Excel table: 715_supreme_court overlap table data

Summary: Most people have the impression that the supreme court is split into two voting blocks (liberal and conservative) with Justice Kennedy acting as a swing vote making most of the decisions. This is false. 

There are 9 justices on the Supreme court. In each case, the justices can be part of the majority, the minority, or recuse themselves (not be a part of a case). This means the number of votes in the majority can range from 1 to 9 (1 in the incredibly unlikely situation where 8 justices recuse themselves). Below is the frequency of majority vote count outcomes for the current court since Kagan joined in 2010 (11/8/2010 to 07/01/2014 data).

supreme court majority vote count

Contrary to what most people think, almost 40% of the time the justices are in complete agreement. 

NOTE: Reddit /user/LittleToke noted that this number does not include absentee numbers. For more accurate numbers on complete agreement and other cool supreme court viz check out his post here!

So how often do the justices agree with each other? In the table below, blue names are those appointed by Democratic presidents and red by Republican presidents. The bottom three rows show overall average agreement for each justice, and agreement with the other Democrat and Republican appointed justices.

overlap between justices

The lowest agreement can be found between Alito and Ginsburg, but amazingly this “low” agreement is still more than half the time (58%). It is likely that the most divisive cases get the media attention and thus lead to the impression that the court is more divided than it is in actuality. But that is a question for another day.


1. Agreement data calculated from the “majority” column of the larger dataset. This means these values are calculated from whether the justice agreed with the majority or minority. There are other values (such as opinion) and vote (concurrences and dissents) that could be used to estimate not just voting agreement but ideological agreement.

2. In the data table recused’ justices data was not counted against them, they were simply not included in the averages for agreement between justices.

3. Image Credit: