My collaborator Micheal McDonald and I are now analyzing the data that resulted from the crowd-sourcing participative electoral mapping projects we were involved in and others. Our earlier article, analyzing redistricting in Virginia established that members of the public are capable of creating legal redistricting plans, and in many ways perform better than the legislature in doing so.
Our latest analysis, baed on data from Congressional redistricting in Florida, reinforces this finding. Furthermore, it highlights some of the limits of reform efforts, and the structural tradeoffs among redistricting criteria.
The reforms to process in Florida, catalyzed by advances in information technology, enabled a dramatic increase in public participation in the redistricting process. This reform process in Florida can be considered a partial success: The adopted plan implements one the the most efficient observable trade-offs among the reformer’s criteria, primarily along the lines of racial representation by creating an additional Black-majority district in the form of the current 5th Congressional District. This does not mean, however, that reform was entirely successful. The adopted plan is efficient, but is atypical of the plans submitted by the legislature and public. Based on the pattern of public submissions, and on contextual information, we suspect the adopted plan was drawn for partisan motivations. The public preference and good-government criteria might be better served by the selection of the other efficient plans – that were much more competitive, and less biased, at the cost of a reduction of the majority-minority seat.
Most of these trends can be made clear through information visualization methods. The figure below draws a line representing the Pareto (efficient) frontier in two dimensions to illustrate the major trade-offs between number of partisan seats and other criteria.
One of the visuals is below — it shows the tradeoffs between partisan advantage and different representational criteria, based on the pareto-frontier of plans publicly available.
(Click on the image below to enlarge)
These frontiers suggest that some criteria are more constraining on Democratic redistricters than on Republican redistricters. On the one hand equipopulation is equally constraining on Democratic and Republic partisan seat creatio. However, the data suggests a structural trade-off between and Black-majority seats and Democratic seats.
More details on the data, evaluations, etc. appear in the article (uncorrected manuscript).