664 



SCIENCE 



[N. S. Vol. XXXIII. No. 852 



In the same diagram the black lines repre- 

 sent in the same manner the grading by forty 

 teachers of rather large numbers of B-students 

 during five sessions of the university. The 

 former passing grades were A, B and 0, and 

 these were assigned, on the average, to 24 per 

 cent., 35 per cent, and 32 per cent, of the 

 students. Since the B-group is the largest of 

 these, it is best to use it for comparison with 

 our present M-group. The diagram shows 

 what fractions of the B-students were given 

 the grade of A (at the top) or a grade lower 

 than B (at the bottom of the diagram). The 

 scale of drawing, for comparison's sake, has 

 been adjusted so that the central column of 

 this case is identical with the central column 

 for the M-students of the present system. 

 The diagram clearly shows that a large frac- 

 tion of the B-students were formerly unjustly 

 marked other than B, in figures, 24 per cent, 

 of them. And worse than this, many stu- 

 dents were graded two steps lower by one 

 teacher than by another. For example, take 

 a student in the center of a file of B-students. 

 Under one teacher, at a', b', in the diagram, 

 such a student standing just in the middle 

 between a* and b' would invariably receive 

 the grade of A. Under another teacher, a 

 student of the same rank of scholarship, 

 standing in the middle between a^ and b", 

 could not entertain the slightest hope of get- 

 ting a higher grade than C. Moreover, the 

 best of the B-students, at the point of a° in 

 the diagram, would get no B under this 

 teacher. Such enormities of divergence have 

 forever disappeared since the introduction of 

 the new system. Even the extreme cases of 

 the M-students, AS B^ and A", B°, overlap 

 to some extent. The ratio expressing the lack 

 of uniformity in grading was twenty-four to a 

 hundred as short a time as two years ago. It 

 is now ten to a hundred. I am confident that 

 it will still further decrease in the near fu- 

 ture. The improvement already attained is 

 reaUy greater than it appears in the diagram, 

 although it is there conspicuous enough. The 

 teachers who varied so enormously under the 

 old system were all teachers of very large 

 classes, so that there was no possible excuse 



for this variation. I must point this out also 

 for another reason. Professor Dearborn, in 

 discussing our former variations, thought that 

 they were probably to be explained by the fact 

 that teachers of small and advanced classes 

 had been included in the statistics. They 

 were not. 



I am sure that any institution can in the 

 same way make its grading more .and more 

 uniform, but only under the condition that 

 the faculty does not expect mere legislation 

 to accomplish the desired result. Our experi- 

 ences in Missouri prove conclusively that it is 

 absolutely necessary for the faculty to have a 

 special committee which collects each semester 

 the cumulative results of the grading of each 

 teacher and calls attention to those deviations 

 from the rules which seem to be unwarranted. 

 We publish each semester among the teachers 

 of the university a statistical table showing 

 how each teacher has thus far graded his 

 students. I regard this publicity as essential 

 to the success of our system. Besides this, 

 our committee collects data as to the relative 

 standing of special classes of students, for 

 example, athletes, fraternities, women, fresh- 

 men, etc. 



Although the method used by the individual 

 teacher for ranking the members of his class 

 is irrelevant to the system of grading under 

 discussion, I wish to mention briefly the 

 method used by myself because there is no 

 end of complaint on the part of college teach- 

 ers concerning the large amount of time 

 which the teacher has to spend on examina- 

 tions. I use almost exclusively the conjec- 

 tural or completion method which was intro- 

 duced into psychological science by the late 

 Professor Ebbinghaus for the measurement of 

 general intelligence, but can be used with 

 equal success for the measurement of the 

 student's rank in a class. The following is 

 an example of an examination blank in which 

 certain words have been omitted. The stu- 

 dent must fiU in the blank spaces in such a 

 manner that the whole becomes meaningful. 

 The fact of simultaneous color induction ia well 

 illustrated by the Bering window. This consists 

 of two narrow windows at some distance from 



