The Statistics of Redistricting

Now that the U.S. Census Bureau has tallied the 2010 population figures, the redistricting process has begun in earnest. The census data are used by state officials to realign congressional and state legislative districts taking into account population shifts since the 2000 census and assuring equal representation for their constituents in compliance with the "one-person, one-vote" principle of the 1965 Voting Rights Act. The new voting districts are drawn once a decade, after each census, to make sure that all congressional districts have roughly the same number of people. In addition, jurisdictions with a history of voting discrimination are required to pre-clear all proposed changes with the U.S. Department of Justice. Research Design Associates statistical experts have worked with the U.S. DOJ and voting rights organizations since the late 1980s to analyze redistricting plans and to assess voting practices for possible unlawful violations. RDA statistical experts have testified about voting issues in federal courts in 12 states.

The redistricting stakes are huge. Redistricting battles are occurring all across the U.S. Eighteen states are either adding or losing Congressional seats. The big winners are Texas which gained four seats and Florida which gained two. New York and Ohio are the big losers, each losing two seats. Georgia gained one seat. For the four U.S. regions, losses tended to occur in the Northeast and Midwest while the gains occurred in the South and West. The one exception to this pattern is Louisiana which lost one seat. There is much pressure on state officials who recarve the electoral maps: to try to keep their grip on power or pry it away from opponents while taking into account pressure from those who want to lessen the impact of partisan politics. The creation of oddly shaped districts has been going on for a long time. They take their nickname, gerrymander, from a 19th century Massachusetts governor, Elbridge Gerry, who drew a district in the shape of a salamander.

RDA statisicians and voting analysts use an array of mapping and statistical tools to assist their clients in assessing the impact of changes in district boundaries and voting practices. Sophisticated mapping software and inferential statistical methods such as correlation, regression and maximum likelihood are used to analyze proposed voting changes. We also conduct exit polls to determine whether or not voters enjoy the equal opportunity to participate in the politcal process and elect representatives of their choice. In my next blog, I'll write about some of RDA's recent voting cases.