Hi all, my question is whether a mediator must be measured in the hypothesized order. If we manipulated X, then measured Y and M (not M and Y), it is fine?

I've seen both in previous work. But does it influence the findings?