The distinction you highlight between the expected absolute deviation and the absolute expected deviation seems extremely important; Jordan Ellenberg says that the study is measuring one and calling it the other, and goes on to say (emphasis added)
Let me be clear: I don’t think the authors are trying to put one over. This is a mistake — a somewhat subtle mistake, but a bad mistake, and one which kills a big chunk of the paper. Science should not have accepted the article in its current form, and the authors should withdraw it, revise it, and resubmit it.
Thanks for the link. Ellenberg gives a great analogy to stock prices and points out something important: the study measured the absolute expected deviation
|) for the predictors, but measured the expected absolute deviation ) for the reporters.
This difference can explain the observed discrepancy between the two groups without invoking the “end of history illusion” because
%20%3E%20|c%20-%20\operatorname%20E(p)|).
This difference came about because obtaining an X+10-year-old’s report of current personality is equivalent to drawing a random sample from an X-year-old’s distribution of actual future personality, but the X-year-olds’ are providing point-estimates (of something like the mean) from their distributions of predicted future personality.
The distinction you highlight between the expected absolute deviation and the absolute expected deviation seems extremely important; Jordan Ellenberg says that the study is measuring one and calling it the other, and goes on to say (emphasis added)
Thanks for the link. Ellenberg gives a great analogy to stock prices and points out something important: the study measured the absolute expected deviation
|) for the predictors, but measured the expected absolute deviation ) for the reporters.This difference can explain the observed discrepancy between the two groups without invoking the “end of history illusion” because
%20%3E%20|c%20-%20\operatorname%20E(p)|).This difference came about because obtaining an X+10-year-old’s report of current personality is equivalent to drawing a random sample from an X-year-old’s distribution of actual future personality, but the X-year-olds’ are providing point-estimates (of something like the mean) from their distributions of predicted future personality.