NOVEMBEK 14, 1919] 



SCIENCE 



463 



Percentage havins 



weight should be 



" 1 



" 2 



" 3 



" 4 



" 5 



" 6 



" 7 



" 8 



" 9 

 10 



0.00 



0.15 



1.54 



8.46 



23.42 



32.86 



23.42 



8.46 



1.54 



0.15 



0.00 



For the present purpose it will be more use- 

 ful to express these same data in the following 

 form: 

 Percentage having weight — 



or above should be 100.00 



ISTow a similar tabulation may be made in 

 the ease of grades. An examination of the 

 grades assigned by teachers shows a distribu- 

 tion which, while it is very far from having 

 the bilateral symmetry of the theoretical de- 

 parture law, nevertheless gives evidence of a 

 somewhat similar continuous relation through- 

 out the range of the assignment. The follow- 

 ing data, based on the tabulation of over 

 one hundred thousand grades from a large 

 number of school and college teachers,^ are 

 probably fairly typical : 

 Percentage of grades which were — 

 50 or above was 100 



2 See article by author, "A Standard of Inter- 

 pretation of Numerical Grades," School Beviev), 

 Vol. XXV., p. 412, June, 1917. 



Let us plot these two sets of data on the 

 same diagram, so that they may be easily com- 

 pared (Fig. 1). The method of translation 

 from grades to weights may then be made 



W, ASSIGNED WEIGHT 



w\ \ . 



G. ASSIGNED GRADE 



Fig. 1. 



clear by a single illustration. The grade 

 curve shows that the grade 85 or ahove is 

 attained in YO per cent, of all cases, while the 



Fig. 2. 



weight that should be attained in the same 

 percentage of cases is shown by the weight 

 curve to be 4.9 or above. It logically follows 

 that the grade 85 corresponds to the weight 

 4.9; etc. 



