24 PSYCHOLOGICAL EFFECTS OF ALCOHOL. 



the case of caffein." While it was "very difficult to distinguish the 

 two from one another when the dose was small," with doses of 24 

 to 40 c.c., such as were used in his work with the ergograph, Rivers 

 regarded it as probable that the alcohol would be recognized; conse- 

 quently he adopted the additional precaution of comparing two dif- 

 ferent doses. In other words, in experiments on alcohol, Rivers felt 

 obliged to supplement the doubtful efficacy of control mixtures by the 

 systematic arrangement of his experiments. We carried this process 

 to its only logical conclusion in experiments on alcohol, i. e., to develop 

 as far as practicable the controls that are dependent on the nature of 

 the experiments as well as those that are dependent on their systematic 

 arrangement. These internal preventatives of the effect of bias we 

 believe to be particularly effective in our experiments, since one of the 

 main principles of selection of measurable phenomena was the greatest 

 possible freedom of the process from the interplay of arbitrary and 

 capricious voluntary modification. The danger of such bias must 

 have been much greater in ergographic experiments, in which the 

 complex interrelation between capacity and effort is subject to large 

 and uncontrollable variations. It must have been peculiarly great 

 while the experimenter served as subject. After mature consideration, 

 in view of the impracticability of completely masking the "taste" of 

 the alcohol, in view of our systematic precautions against voluntary 

 modification of our experimental results, and in view of the character 

 and variety of our subjects, we decided that the regular use of highly 

 flavored control mixtures be abandoned, except in the case of the 

 psychopathic subjects, in whom the knowledge that alcohol was being 

 taken might conceivably have produced some agitation. We believe 

 that the nature and sj^stematic arrangement of our experiments, on 

 the latter of which Rivers himself came finally to rely in alcohol experi- 

 ments, contain more efficient controls than could be produced by the 

 use of doubtful control mixtures. 



SUBJECTS. 



The selection of subjects presented a number of practical difficulties. 

 In accordance with our program (p. 267), it seemed desirable to inves- 

 tigate the effects of alcohol on total abstainers, occasional users, 

 moderate users, habitual drinkers exceeding 30 c.c. of absolute alcohol 

 a day, and on excessive drinkers. Of these groups, the first and last 

 proved most difficult to secure. 



With respect to the first group the practical difficulties were social 

 and moral, on the one hand, and theoretical on the other. In the first 

 place, we were loth to assume responsibility for administering alcohol 

 to total abstainers for a series of experimental days. There was a 

 small but serious risk of initiating a practice that might become 

 habitual and excessive. In the second place, we were confronted by 



