June 1, 1906.] 



SCIENCE. 



843 



enees; so far as the facts go they show a 

 slight advantage in favor of the science 

 examinations and a slight inferiority of 

 the mathematics examination. The facts 

 are given in Table V. 



The reliability of any one single exam- 

 ination mark from the dozen or more given 

 as the test for entrance is practically im- 

 portant because the disrepute into which 

 entrance examinations have fallen in the 



tional knowledge of the cleverness of each 

 candidate's coaching. But the general fact 

 that the entrance examinations do not 

 measure at all accurately the candidate's 

 capacity can be demonstrated from two 

 lines of evidence— the great variability of 

 the marks of the same individuals in dif- 

 ferent special branches of the same subject 

 (such as — Vergil and Cicero in Latin) and 

 in precisely the same subject, in the case of 



Average, 



.43 



.40 



.30 



.55 



Table VI. 



Class of 1901. Correlation between Entrance Mark in Vergil and College Work in Latin. 



Entrance Mark in Vergil. 



62 64 66 68 70 72 74 76 78 



84 86 88 90 



2 3 12 1 3 2 13 



opinion of high school teachers is largely 

 due to the failure in some examination of 

 a pupil known to be of first-rate ability in 

 the subject. I give in Table VI. the facts 

 concerning the relation of the entrance 

 mark in Vergil of one class to achievement 

 in college Latin courses as a sample. The 

 correlation as calculated is almost zero. 



It is the opinion of many sagacious ob- 

 servers that one defect of the traditional 

 entrance examinations is that they meas- 

 ure the cleverness of the student's coaching 

 rather than his own fundamental capaci- 

 ties. This hypothesis as it stands can not 

 be tested by my records without the addi- 



examinations repeated because of initial 

 failure. 



The average range of difference of the 



Table VII. 



Range of Difference of the Same Indimdual's 



Separate Entrance Marks in Latin. Only 



last trials used. Class of 1905. 



