120 PSYCHOLOGICAL EFFECTS OF ALCOHOL. 



the lower level, motor reactions. However, Subject IX is the only 

 one who shows it in the curves. The speech-habit category does not 

 appear to be consistently affected by alcohol in any other subject. 

 But the average of the alcohol days shows for all subjects some increase 

 in the speech-habit category; the miscellaneous category decreases in 

 5 cases out of 6, the total change being the same, as shown in table 15. 

 The effects reported by Riidin and Fiirer seem established, but 

 require much heavier doses than are here given to produce them. 

 The present effects, though small, are in the same direction. As 

 regards the egocentric category, ordinarily the one of the greatest 

 psychological meaning, Subjects II and III show the nearest approach 

 to a consistent tendency in the direction of an increase of egocentric 

 responses under alcohol. Nothing can possibly be read into the figures 

 for the remaining categories as an alcohol effect. There is no doubt 

 that, as Partridge 1 remarks, sufficient alcohol will produce great changes 

 in the character of associative responses; but beyond the results of such 

 experiments as those of Riidin and Fiirer, or the unanalyzed data of 

 Partridge, we are unable to say in what direction they are, or whether 

 they would be in a uniform direction for different subjects. 



"FREQUENCY" OF THE RESPONSE WORDS. 



The Kent-Rosanoff Frequency Tables 2 make possible this sort of 

 measurement. For comparing the usualness of response on alcohol 

 and normal days, it was thought necessary to divide the Kent-Rosanoff 

 series of 100 into 2 series of 50 words each, so that comparison should 

 not be had with material that had been used before. The two series 

 should, of course, be so selected as to show in central tendency the 

 same frequency of response in each. It did not seem that this should 

 be left to chance, for a series of words like dark, mutton, or short, would 

 be much more likely to have " usual" or "frequent" responses than 

 one composed of words like anger, religion, or memory. Various sys- 

 tematic means of selecting two equal series were tried, that finally 

 employed being as follows: 



So far as was known, the normal median frequency value of the Kent- 

 Rosanoff series of associations is represented by the figures 9.0. It 

 was then determined for each stimulus word how many subjects out of 

 a thousand had given a reaction word which was less frequent than this. 

 Thus, in the case of table, 733 persons did so, in the case of dark, 

 352 persons, etc. All the words were arranged in the order of the 



'Partridge, Studies in the Psychology of Intemperance, New York, 1912. 



2 These frequency tables were prepared on the following basis: "From the records obtained 

 from these subjects, including in all 100,000 reactions, we have compiled a series of tables, one 

 for each stimulus word, showing all the different reactions given by 1,000 subjects in response 

 to that stimulus word, and the frequency with which each reaction has occurred." (Kent and 

 Rosanoff, Am. Journ. Insanity, 1910, 67, pp. 37 and 317.) 



