360 



MEMOIRS NATIONAL ACADEMY OF SCIENCES. 



[Vol. XV, 



"Number 24. Make a cross on the caudal appendage of the squirrel. 



"Number 25. Make a cross at the orifice of the jug. 



"Number 26. Make a cross on the superior aspect of the pulpit. 



"Number 27. Make a cross on the major protuberance of the bludgeon. 



"Number 28. Make a cross on the sinister extension of the swastika. 



"Number 29. Make a cross on the cephalic extremity of the homunculus. 



Item 1 is a double item and is scored 



right minus wrong," since sometimes 



The score is number of items right, 

 everything is crossed . 



Difficulty of items. — The per cents of successes for the items are shown in the second table (Table 426). 



Figures of table 42b show per cent successes for separate items of test. Group I, 460 men of non-English 

 development battalion, including a few high-grade English-speaking men from training cadre. Group II, 173 

 English-speaking men from depot brigade and from company of heart defects. 



While there are some inversions in columns 2 and 3, the order of difficulty at the two levels of language ability ia 

 not greatly different, and the columns may be said to be approximately consistent. 



Correlations of group with individual language test. — The coefficients of correlation between the group test and the 

 individual test, taken in toto and separately in its ' ' verbal " and ' ' performance " halves, in the cases of the non-English 

 and the English groups mentioned above, are as follows: 



Correlations of group with individual linguality test scores. 



Group II 

 (English). 



Group with performance individual 



Group with verbal individual 



Group with total individual 



Thus the group score appears to be related to either and to both halves of the individual test about as well as the 

 two halves of the individual test are related to each other. Furthermore the relation is not altered by a change in the 

 level of language ability. 



Thus it seemed legitimate to combine the English and non-English groups and to compare (for the combined group) 

 total individual score with group score. The following values were found: 



Comparison of group and individual linguality test values (group of 860). 



Group test. 



Average 



Standard deviation 



Coefficient of correlation (630 cases) 



11.71 

 7.55 



The equation of correlation is therefore: Individual score =1.13 (Group score) +13.95. Inspection of the table 

 of correlation shows that the correlation is linear. 



Ratings. — The correlation of .79 was considered high enough to justify the selection of critical points on the group 

 test as those corresponding to the critical points on the individual test, determined by way of the equation of correlation. 



The group test is much harder than the individual test. It is not possible to give a rating of E by the group test. 

 This is not surprising when it is remembered that the points in the E group involve gestures and other aids, and not 

 direct understanding of words 



