FiftyTwo comments on Calibrate your self-assessments - LessWrong

68 Post author: Yvain 09 October 2011 11:26PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (117)

You are viewing a single comment's thread.

Comment author: FiftyTwo 10 October 2011 01:11:29AM 5 points [-]

I've always had this problem when asked to self assess on personality traits. E.g. 'Are you extroverted?' Compared to what baseline? My friends? Some hypothetical average person?

Comment author: scav 10 October 2011 08:08:08AM 4 points [-]

Not a problem: you have correctly identified a meaningless question. The correct answer is "mu".

I have read, and maybe don't take my word for it, that so called "personality traits" don't reliably correspond to objective properties of a person. Or rather, that we apply them as labels to a person on seeing one set of behaviour we associate with it, but they aren't predictors of any of the other kinds of behaviour that typically get assigned the same label.

I don't have a reference to it here at work, but at home I have a psychology book which contains a paper about it. Maybe someone else here will have a clearer memory of what I'm talking about?

Anyway, that's my main objection to so-called personality tests that give you a numeric personality score on one or more axes. I just want to ask "what are the units on that number?", "how was this calibrated against objective measurements?". The Myers-Briggs test goes one step further and combines 4 uncalibrated numeric scores into a "personality type". Might as well be a horoscope.

Comment author: Morendil 10 October 2011 08:44:36AM 3 points [-]

clearer memory of what I'm talking about

This paper by Borsboom (pdf) maybe? (Specifically attacking the Big Five.)