The Degree of Accuracy of Statistical Data 9 



V 



Another matter that should be spoken of in this connection is 

 the method of computing the v's , i. e., the moments about an 

 arbitrary vertical. Professor Pearson, in one of his articles, pub- 

 lishes a table of the first six powers of the integ'ers from i up to 

 30, and refers to this table several times. Professor Davenport, 

 in his Statistical Methods, copies this table, and says : "This 

 table is useful in calculating moments." This is not true, except 

 in very rare instances, and then only to a very insignificant ex- 

 tent; but the statement coming from such authorities is likely to 

 imduly influence inexperienced computers and cause them to 

 waste much valuable time. Instead of multiplying the ordinates 

 by the successive powers of the corresponding abscissas to obtain 

 the moments, it will be found much better to proceed as in the 

 following example : 



No doubt this is the method generally followed by expert 

 biometricians but as the only reference books on the subject in 

 existence recommend the long and roundabout way, I feel justified 

 in calling attention to the matter. 



VI 



In vol. I, part IV, of Bioinetrika, Professor Pearson generalizes, 

 what he has called "Francis Galton's Difference Problem;" that 

 is, he has obtained a general law giving the difference between 

 individuals competing for a set of prizes. The work is very 

 brilliant, and shows well Professor Pearson's mathematical 



95 



