As an example, the PGM Project has always been clear that Robert Charles Anderson's Great Migration works are used "unless there is more recent published research that corrects or adds" to them. Which there has been, and will continue to be, including revisions and updates from the GM team itself.
Anderson's GM and Richardson's RA are significant works in their genre because they are broad, comprehensive, with a consistent data structure/format, relatively accessible even to the beginner, and they exhaustively document their sources and explain their methods, allowing their research and conclusions to be reproduced, extended, or refuted by others. It would be ridiculous to hold such publications to any standard of perfection, since that is not how historical scholarship works, but their rigor and overall reliability are indisputable.
Large complex Projects like PGM and MC need a shared baseline around which their many participants can find common ground. Without it, the work would be unmanageable; thousands of significant profiles could resemble this G2G thread, an outcome the WikiTree leadership strives mightily to avoid. It makes absolute and complete sense that the Projects would choose GM and RA respectively, because the qualities mentioned above that these works share (broad, accessible, consistent, reproducible, falsifiable) also make them wonderfully suited to collaborative wikiwork.
Of course history doesn't freeze the moment someone writes it down. There is always new research and the possibility of new data coming to light which could revise or upend prior conclusions. That is how scholarship rightly works. Both Projects already acknowledge this.
I wonder if we are getting wrapped around the axle unnecessarily by the word "master" in the original question? I found it a curious choice of word, especially because some seem to be choosing to interpret it in the sense of superlative (making it a matter of opinion and debate) rather than one of the neutral senses, like principal or even from which duplicates [or derivatives] are made. Perhaps it is worth pointing out that neither PGM nor MC refers to GM or RA as a "master" source on their Project page, a fact which I know because I wrote both of them.
Could we simply agree that these are starting sources, and not "master" anythings, and might that allow us to de-escalate the rhetoric a bit?