As discussed on the PRESIDENTIAL RESULTS BY CONGRESSIONAL DISTRICTS Project page,
the results denominated as the PRCD information are compiled independently by Polidata in conjunction with various sponsors.
As such, they are compiled right after the election from source material, most of which is still on paper or, at any rate, not data files.
Regardless, most of this work is accomplished before the availability of precinct-level election datasets prepared by State Election Authorities.
This means that there may be some differences in the raw numbers, the percentages, or even the 'winner' of the district between the Polidata information
and numbers that could be compiled from information finally available from the States.
As part of the ongoing Polidata data library development processing, precinct-level datasets are acquired and/or compiled when available.
As time allows, comparisons are made between the State version and the Polidata version for the PRCD data.
It is also important to understand that these numbers are, generally, the result of compilation of precinct-level information, many of which are 'split',
i.e., they contain portions of more than one district and/or are in jurisdictions that require the deallocation of
centrally-counted absentee and/or early votes. The sheer size of the compilation effort may result in clerical errors.
The deallocation process, for both split units and centrally-counted votes, requires judgment as to the factor to be used to allocate the votes to districts: in such districts there are no 'hard' numbers.
Moreover, even when the final datasets are available, inconsistency exists as to how votes in these situations are addressed by the dataset developer.
Differences, when noted, may be available here. Generally speaking, the datasets are not updated in a piecemeal fashion unless it is a clear and significant error.
