404 



SCIENCE 



[N. S. Vol. XXXVI. No. 926 



Distribution of Grades at Eeed College," 

 which appeared in the June Y, 1912, issue of 

 Science. 



Without wishing to criticize the Eeed Col- 

 lege system which the writer considers amply 

 adequate to bring about uniformity in grading 

 at Reed College itself, there yet remains the 

 question, is the Eeed College system the one 

 most likely to be adopted by other educational 

 institutions? The extent to which the ques- 

 tion of grading is at present agitated justifies 

 the conclusion that there is a tendency toward 

 the standardization of grades — not only within 

 the walls of a given school, but among the 

 schools themselves. As long as no absolute 

 units for mental measurement exist, the one 

 essential factor for standardized grades is 

 that the form of the distribution of these 

 grades be identical for the various institu- 

 tions. If this is true, the questioa arises 

 what form of distribution shall be adopted? 



The form adopted by Reed College is a 

 modification of the normal curve, skewed to 

 take account of the " selected nature of col- 

 lege students." It is this question — the ad- 

 visability of skewing the normal curve — which 

 is to be considered in this paper. 



From present indications, if the various 

 educational institutions throughout the coun- 

 try were suddenly obliged to adopt a uniform 

 type of distribution for their grades, it is prob- 

 able that the normal type of distribution is 

 the one which would be most readily accepted. 

 As soon as deviations, either in one direction 

 or the other, are suggested, each faculty would 

 insist on skewing the curve in the direction 

 which would best express, in the opinion of the 

 faculty, the local conditions. If the different 

 school faculties are encouraged to express in 

 their grades the degree of selection which 

 they believe to exist in their student body, the 

 personal element will again become so strong 

 as to eliminate the uniformity which a scien- 

 tific system is supposed to introduce. In a 

 faculty composed mainly of " home talent " 

 there is a tendency to suppose that its par- 

 ticular students are a specially selected group 

 with respect to intelligence, and under such 



conditions it may happen that a faculty will 

 adopt a form of distribution in which all of 

 the students are supposed to be above the gen- 

 eral average instead of only three fourths, as 

 is done in the Reed College curve. 



Until a system of absolute mental units 

 has been invented, it is futile to try to make 

 grades represent absolute accomplishment. 

 The best that can be done under the condi- 

 tions is to let the grades express relative ac- 

 complishment. It should be clear that skew- 

 ing the normal curve is aa attempt to make 

 grades represent absolute accomplishment in 

 that the degree of deviation from the normal 

 type of distribution is supposed to measure 

 the degree of superiority or inferiority, above 

 or below some hypothetical absolute accom- 

 plishment of the population at large. Since 

 we do not know enough about the intellectual 

 capacity of the population at large to gain 

 general acceptance for such numerical values 

 as we might choose to assign, perhaps it would 

 be better, for the present, to be satisfied with 

 measuring relative accomplishment. To jus- 

 tify the deviation from the normal curve 

 shovm by the Eeed College system, it seems 

 that the following two factors should permit 

 of quantitative statement. 



1. The numerical degree of the selection 

 with respect to scholarship. 



2. The difl'erence in the form of the distri- 

 bution of the selected group, from that of the 

 unselected group. 



With respect to (1) it is doubtful whether 

 the data at our disposal are of such convinc- 

 ing validity as to justify our giving it nu- 

 merical values, especially when we consider 

 not only those students who fail to attend the 

 university after finishing their high-school 

 work, but also those who drop out during the 

 first, second and third high-school years. 

 Many students leave high school because their 

 parents have not the means to enable them to 

 continue; some students get dissatisfied with 

 the school work and prevail upon their pa- 

 rents to let them go out and do " real work." 

 It is not necessary here to indicate the many 

 reasons for which children leave the high 



