Biometrika One Hundred Years
Stephen John Senn · International Journal of Epidemiology · 2002
Biometrika is Pearson's child and has something of its father's eccentricity. Regarded by many as the most important of all statistical journals, like Pearson it has excelled in many things but also has curious blind spots. Fisher is only represented by his brilliant paper of 1915 on the correlation coefficient. His next contribution on the subject was summarily dismissed by Pearson and was published in Gini's journal Metron (1921) instead. As early as 1922 Fisher had his revenge, albeit not in the pages of Biometrika. He showed that Pearson had got the degrees of freedom for his own chi-square test wrong. Naturally, Pearson's thunderbolt in reply was in Biometrika, where he wrote, ‘The above description of what seem to me very elementary considerations would be unnecessary had not a recent writer in the Journal of the Royal Statistical Society appeared to have wholly ignored them. He considers I have made serious blunders in not limiting my degrees of freedom by the number of moments I have taken’. Nowadays, of course, attentive lecturers dock marks from any students who make Pearson's mistake.