Notes: State data for Louisiana are not included because Louisiana has a unique, two-round primary system in which candidates who receive more than 50 percent of the vote in the open primary are elected without having to participate in a general election. State data for Nebraska are not included because Nebraska has a unicameral legislature and candidate party aliation is not listed on the ballot. States in gray have a bias below 5%. However, in Figure 1A, some states that have a higher bias are also displayed in gray because there is no arrangement of districts that could reduce the bias due to the scarcity of seats (i.e. districts are already as close to accurate representation as possible). The 10 states displayed in gray for this reason are: Alaska, Colorado, Delaware, Maine, Montana, North Dakota, New Mexico, South Dakota, Vermont, and Wyoming.

Sources: U.S. House bias: U.S. House data for the 2012 and 2014 elections were compiled from Ken Kollman and others, “CLEA Upper Chamber Elections Archive,” Regents of the University of Michigan, available at http://www.electiondataarchive.org/clea-upper-chamber-elections- archive.php (last accessed April 2019). U.S. House data for the 2016 election were compiled from MIT Election Data and Science Lab, "U.S. House 1976–2016," available at https://doi.org/10.7910/DVN/IG0UN2 (last accessed April 2019). State Senate bias: State legislative data were compiled from Carl Klarner, "State Legislative Election Returns, 1967-2016: Restructured For Use," Harvard Dataverse, available at https://doi.org/10.7910/DVN/DRSACA (last accessed April 2019).