@nashdelirium @kareem_carr @Notgiamatti There's a model for how to calculate this distortion in from the tweet below. If x is the true proportion, then the distorted proportion is ArcLo(c Lo(x) +(1-c) Lo(.5)) where c is an empirical constant, Lo is the log
441 followers
2,594 followers
@CopyrightLibn @bmwiernik @grimalkina Super interesting discussion, reminds me of this paper. https://t.co/PgIvYDpjLd
6,843 followers
@jakehofman @eredmil1 @JessicaHullman @DrDavidLandy That'd be my explanation too https://t.co/WuIbLPtc1k
6,843 followers
@JSEllenberg This is basically what you'd expect given human perception of proportions. See https://t.co/BCUfkRShyw https://t.co/Pkdz1ZX8qg