Daily Kos Elections is pleased to announce the publication of a comprehensive statement of methodology that explains how we calculated the 2020 presidential election results by congressional district, down to the last detail. Armed with this document, anyone wishing to will be able to replicate our calculations (which can be found here) in full or double-check all of our work, which we encourage.
The bulk of our guide is devoted to the two hardest problems posed by calculating presidential results by congressional district: votes that are unassigned to any geographic precinct (often those cast before Election Day), and precincts that are split between multiple districts. We've developed a number of different methods for allocating these votes to the proper district, and our confidence in their accuracy is high. The necessity of these approaches underscores, however, that some calculations are estimates (albeit very carefully considered ones), and highlights the need for state election authorities to calculate this data themselves.
We've also included a section that explains which methods we've used to allocate unassigned votes and split-precinct votes on a state-by-state basis, and that includes links to all of the raw election results we've relied on for our calculations. Our methodology is applicable to calculating district-level results of other statewide races (such as for Senate or governor), and for state legislative districts as well as congressional districts. As we begin the process of calculating the most recent presidential results by legislative district, we will update this document.
You can find all of our district-level data at this bookmarkable permalink. We invite any and all questions about our methodology statement and will continue to refine it based on your feedback.