Racially polarized voting analysis

Racially polarized voting (RPV) analysis estimates how voters of different racial groups voted in an election, given that individual ballots are secret. It is a cornerstone of redistricting litigation and remains critically important even after Callais as circumstantial evidence of systemic patterns.

Why RPV analysis is necessary

Because ballots are secret, we cannot directly observe how a Black voter or a white voter voted. But we can observe, for each precinct, the racial composition and the vote totals. RPV analysis uses statistical methods to infer group-level voting patterns from these aggregate observations.

Ecological regression (ER)

The simplest method is ecological regression, which regresses the precinct-level vote share for a candidate on the precinct-level racial composition. If precincts with higher concentrations of Black voters consistently show higher vote shares for Candidate A, that suggests Black voters prefer Candidate A. Ecological regression is easy to implement but prone to the "ecological fallacy"—inferring individual behavior from group data.

Ecological inference (EI)

Ecological inference, developed by Gary King (Harvard), uses a statistical model that accounts for the bounds of the data (if a precinct is 70% Black and Candidate A got 80% of the vote, at least some Black voters must have voted for Candidate A). EI produces better-calibrated estimates than simple regression and is the dominant method in expert testimony. Common implementations include the eiCompare R package, which is widely used in redistricting expert reports.

Georgia Senate district example

In a hypothetical Georgia State Senate district encompassing parts of Muscogee County, an analyst applied ecological inference to the 2022 U.S. Senate race. Results showed Black voters supported the Democratic candidate at approximately 94%, while white voters supported the Republican candidate at approximately 79%. This level of polarization—where the two groups' candidate preferences are nearly opposite—is what courts have historically required for a finding of racial bloc voting.

What RPV cannot tell you

RPV analysis cannot tell you why voters vote the way they do—it does not distinguish partisan loyalty from racial solidarity. Courts recognize this ambiguity. The post-Callais significance of RPV analysis is as evidence that a minority group has a distinct political identity, which is relevant to both the Gingles framework and to historical patterns of polarization supporting an intent claim.