I think there’s more to it than that, although I’m not at all arguing our memories are reliable.
I have a suspicion that it has something to do with how we store and retrieve memories beyond simple conflation of things we know. Like we store the “abstract”, and when our brains are tasked with retrieving the whole thing, it fills in the gaps with what it thinks ought to be there, based on our knowledge and experience.
So for example, with antiquated quotes, we remember the information content of the quote but not the exact wording, and when we go to remember it, our brains fill it in using modern day language. Like for example the Berenstain Bears. We may remember how it’s pronounced, but when we retrieve it from memory, we fill in the actual spelling with how we would normally spell a word pronounced like that - Berenstein, even though it’s incorrect.
Or Curious George having a tail. We recall that he’s a monkey, he’s naughty, and that he’s drawn with a certain art style. If we don’t explicitly notice that he does NOT have a tail, when we remember what he looks like, we’ll assume he’s got a tail, because duh, he’s a monkey.
The flip side is when you DO explicitly notice those differences- I did notice that it’s Oscar Mayer at some point, so that stuck in my head. Otherwise, most people would remember “Meyer”, since that’s the most common way you spell the word pronounced like that.
Most of the examples of the Mandela Effect are things that are outliers- spelled weirdly, phrased weirdly, or otherwise out of the ordinary, and unless someone specifically recalls that they’re unusual, their brains will fill in the commonplace spelling/phrasing.
I don’t have an explanation for the events- I mean, I recall Mandela being released from prison and going on to a political career in South Africa, so that one never happened to me.