It assumes that the ratio of true-to-false statements repeated is the same regardless of political correctness. (If a true PC statement is four times more likely to be repeated than a false PC statement, then a true PI statement is four times more likely to be repeated than a false PI statement.) I’m not sure that’s true.
But this does give us the conditions required to assume a PI statement is more likely to be true than a PC statement, which is valuable.
LINK: Someone on math.stackexchange ask if politically incorrect conclusions are more likely to be true by Bayesian Logic. The answer given is pretty solid (and says no).
It assumes that the ratio of true-to-false statements repeated is the same regardless of political correctness. (If a true PC statement is four times more likely to be repeated than a false PC statement, then a true PI statement is four times more likely to be repeated than a false PI statement.) I’m not sure that’s true.
But this does give us the conditions required to assume a PI statement is more likely to be true than a PC statement, which is valuable.