Correction for 'Causal inference, probability theory, and graphical insights'

Abstract

One should clarify that the numbers in Table 1 of Stuart Baker's article 1 contain unfortunate numerical errors. The numbers with Y = 0 and Y = 1 for the not treated (X = 0) and the treated (X = 1) under the heading ‘ U = ?’ should read as given in the succeeding text. Indeed, these numbers are just the sum of the respective cell frequencies n the U = 1 and U = 0 columns, and pr(Y = 1) equals 260 ∕ (260 + 140) = 0.65 for the not treated and pr(Y = 1) equals 180 ∕ (180 + 220) = 0.45 for he treated. The numbers given in the published paper result in pr(Y = 1) = 300 ∕ (300 + 100) = 0.75 for he not treated and pr(Y = 1) = 100 ∕ (100 + 300) = 0.25 for treated. Corrected part of Table 1: