68 JUNE E. DOWNEY 



sensitiveness to graphic form. It was, however, necessary in the present 

 test to rely in choosing assistants merely upon the author's general im- 

 pression as to an individual's interest in and sensitiveness to personality 

 differences and, particularly, his "feeling" for the characteristics covered 

 by the terms extravert-introvert. The instructions were very simple. 

 Each judge or assistant was given the collection of paired writings of the 

 same sentence, each marked with its identifying number, and asked to 

 choose which sample had been written by an extravert, which by an in- 

 trovert and to record his choice by number under the properly headed 

 column on the record sheet. No names were given. The writing of the 

 faculty group and that of the student group was kept separate and the 

 records listed in two distinct series. 



It was discovered later that the experiment might be conducted as a 

 group test by use of a Bausch-Lomb balopticon. The pairs mounted to- 

 gether on a card could be projected upon a screen. The resulting magnifica- 

 tion was found to be helpful to many judges. In the present paper, how- 

 ever, the results with the balopticon will not be discussed. 



The results may be summarized by giving the percentage range of right 

 judgments for individual judges; the mean for the group of twelve judges 

 with the average variation from this mean; and the percentage of right judg- 

 ments on each pair. In the case of samples obtained from the faculty 

 group, an objective study was made of the writing in order to determine 

 whether certain measurable graphic characters could be associated with 

 temperamental tendencies. 



In terms of percentage, the range of right judgments for the individual 

 judges was from 52 per cent made by one judge, to 80 per cent made by 

 three judges. The mean percentage of correct judgments was 72 per cent 

 with a mean variation of 6.6. Mere chance would give 50 per cent correct 

 judgments. The actual result is 72 per cent correct or 22 per cent better 

 than chance. Seven of the twelve judges scored above 75 per cent right 

 judgments, or 25 per cent above chance. The results, as a whole, are some- 

 what better than those reported when sex was judged from handwriting. 

 In the latter investigations, however, no selection of subjects was attempted 

 while in the present instance very definite selection was made of the penmen 

 whose writing should be used in the experiment. It should be emphasized 

 that this selection was in terms of penmen and not in terms of graphic 

 samples — a procedure which would have invalidated the whole experiment. 

 In judging sex in handwriting a similar procedure would involve choice of 

 handwriting specimens on the basis of scores received by penmen on a 

 questionnaire relative to sex characteristics. 



