On April 5, 2013, three political scientists published their own study of the effect of California’s top-two system, used for the first time in a regular election in 2012. The three authors are Thad Kousser of the University of California, Justin Phillips of Columbia, and Boris Shor of the University of Chicago. The paper can be read at this link, courtesy of the Social Science Resource Network.

The authors used survey research to determine the average political attitudes of the voters of each congressional district. Instead of just asking the voters if they are “liberal” or “conservative” or various other crude categories, it asked them their opinions about 46 political issues. Then the authors used the same data (which had already been gathered by Project Vote Smart) for all candidates for Congress in California, for both 2010 and 2012. The purpose of the study was to see if there was a greater fit between average political attitudes within each congressional district and the identity of the winner, or a lesser fit, or no change at all, between 2010 and 2012.

In 2010, California used a semi-closed primary, whereas in 2012 it used a top-two primary. The study determined that the top-two system did not result in a closer fit between the political attitudes of the average voter within a particular district, and the type of candidate elected in that district.

The study paid particular attention to congressional races in which two candidates from the same party were the only two candidates who appeared on the November ballot. The study concluded that the top-two primary has not halted the continuing partisan polarization of California’s elected members of Congress, and says if anything, polarization has increased and the quality of representation has declined. The paper explores ideas about why this has happened. The paper itself is 31 pages; the appendix lists the 46 policy questions that were answered both by congressional candidates and voters.