How does this not entail being mistaken about our experiences?

Suppose I believe I had chicken for lunch yesterday, when in fact I had pork. That does not mean that, when I had lunch, I actually thought pork tasted like chicken (or, I subjectively experienced the taste of chicken when I actually subjectively experienced the taste of pork - this convolution is my issue with the concept). If the point is, "Sometimes our memory is incorrect," it seems to be wholly uninteresting and hardly worthy of a top-level post - as it adds little insight to this established fact.

On your earlier point: have your dreams changed, qualitatively, since you were a child? How would you even know? Unless you kept rather detailed diaries, then compared them to a large sample of other people to control for age-related changes, it'd be impossible to tell. Moreover, however complex your cultural upbringing, you had one cultural upbringing (perhaps a multicultural one, but one upbringing). Populations do not experience muti-cultural upbringings, they experience different cultural upbringings.

My point is that attaching percentages to probabilities here is largely an exercise in futility. You are using a piece of evidence to fit into a large claim - we can be mistaken about our subjective experiences. Fact X isn't either evidence for Y or not-Y; the vast majority of the time Fact X doesn't really have much bearing one way or another. In short, it's very weak evidence for a highly specific (and, as I've argued, incoherent) proposition. I'll happily concede that some people may dream in color and not remember this (or vice versa), but