The North Carolina Supreme Court ordered the General Assembly to draw new congressional and state legislative maps in a hastily-written order released on February 4. The legislature has to submit the maps to the trial court by February 18.

As far as the criteria the court majority wants the General Assembly to use, what they offer is thin and, on its face, contradictory. They offer some possible measures on pages 6 and 7 of their order:

There are multiple reliable ways of demonstrating the existence of an unconstitutional partisan gerrymander. In particular, mean-median difference analysis, efficiency gap analysis, close-votes, close seats analysis, and partisan symmetry analysis may be useful in assessing whether the mapmaker adhered to traditional neutral districting criteria and whether a meaningful partisan skew necessarily results from North Carolina’s unique political geography.

The contradictory part is that those suggested measures are notoriously weak in dealing with political geography since they cannot indicate if a difference between vote share and seat share between parties is due to how districts are drawn or where voters live. In other words, they do not address if a “meaningful partisan skew necessarily results from North Carolina’s unique political geography.”

Michael D. McDonald and Robin E. Best, who advocate for a gerrymandering measure that incorporates mean-median difference analysis, acknowledged that weakness and proposed to overcome it through creating sets of plans using algorithms (page 320):

Prong 4 answers by evaluating a plan against a set of alternative plans that have been produced using neutral criteria. The form we propose is through algorithms used to generate compact, contiguous, and equipopulous districts drawn without conscious partisan considerations

Assuming that the court majority was serious when they wrote that they wanted evaluations to consider “North Carolina’s unique political geography,” they are asking the General Assembly to submit congressional, North Carolina House, and North Carolina Senate maps that fall within the “impose median maps” section of an analysis I wrote on February 3:

This would incorporate data from some of the plaintiffs’ expert witnesses to create guidelines for map drawers (legislative or court-appointed) to follow. For example, one expert witness for the plaintiffs found that the median congressional map was 9-5 Republican while another found 8-6 was slightly more likely than 9-5.

There are a couple of questions that justices would have to answer as part of their ruling to further define what median maps are. First, should map drawers seek to keep municipalities together when drawing maps? Second, should map drawers be required to draw median maps at the county-cluster level as well as the statewide level? Answering “no” to both of those questions would be more favorable to Democrats and is basically what the court imposed in Common Cause v. Lewis in 2019. Gov. Roy Cooper and Att. Gen. Josh Stein asked the court to only consider median outcomes at the statewide level.

The court could also order some county clusters to be changed.

Here is a range of possible district partisan leanings based on median outcomes, using Dave’s Redistricting Atlas (DRA) program:

-Congress: 9-5 or 8-6 Republican

-NC Senate: 27-23 or 26-24 Republican

-NC House: 65-55 to 63-57 Republican

If the General Assembly submits maps within those ranges, they should be within a safe harbor, unless the real intent of the court majority’s order is to maximize Democratic Party wins.