THE STOP-WATCH AND THE ASSOCIATION TEST 



173 



as is obvious, the mean variations might in a given pair of series 

 be identical, and the averages be identical, yet the discrepancies 

 between corresponding times be large. As a matter of fact, 

 the data from which Table 1 was compiled shows a measure of 

 discrepancy far above that indicated in the table. Many long 

 actual (chronoscope) reactions have short stop-watch records 

 and vice versa, in every series. The tendency to reduce varia- 



TABLE 1 



tions which is shown in Table 1 by almost all experimenters, is 

 shown in the detailed data in even greater degree by everyone. 

 This tendency to react automatically with the stop-watch after 

 a relatively constant interval is somewhat masked in the aver- 

 ages by a number of erratic very long and very short reactions 

 having little correspondence with the actual variations. While 

 the presentation of the entire mass of data is not feasible, Table 

 2 gives a clear indication of the general nature of these data. 



