28 



EMPIRICAL STUDIES OF MEASUREMENT 



N-I30S 



J4N-C51 



)= 516 

 i= (of 

 v= IS 

 /n~ S 



FIG. 8. 



var. B ,. D var. C 



-r- and similarly the median ^ 



var. A * C var. D 



with the median 



A 

 B 



n'~~~7i Tn ^ s combination would be made by taking the median 



U V9.1". 



of both the-- - 



A 



var. B and B' var' A 



ratios, an equivalent of the 



double entering involved in the Pearson Coefficient, or more easily 



B_ var. A , A var. B 

 still by taking A ' var. B B var. A 



