82 



University Studies 



Vol. Ill APRIL, 1903 No. 2 



I. — T]\e Degree of Accuracy of Statistical Data 



BY CARL C. ENGBERG 

 I 



This paper is written as a protest against the unnecessary re- 

 finement of statistical computations as carried out by the biome- 

 tricians of to-day. These practices are well illustrated by the 

 case of the college freshman, who in his zeal and desire for abso- 

 lute accuracy used TT to fifteen decimal figures in the determina- 

 tion of the size of the micrometer divisions in his microscope, 

 although the change in focus necessary even for the other eye 

 of the same observer will, in many cases, alter the size of the 

 object observed by as much as half a division. While this is an 

 extreme example, it is not much worse than the performances of 

 all inexperienced computers, or even than those of many distin- 

 guished mathematicians who are experienced computers, but who 

 have had their practical sense killed by impractical theories. 



In order to give greater weight to my remarks, I shall discuss 

 cases taken from the works of prominent biometricians, espe- 

 cially those of Professor Karl Pearson, the originator and 

 developer of the science. 



87 



