666 



SCIENCE 



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



possess. I must say that, judging from my 

 own experience with students, this argument 

 is a fallacious one. I have never had an 

 excellent student in psychology who did not 

 have a good knowledge in languages, in his- 

 tory, in biology, etc. Indeed, I am inclined 

 to assert that his knowledge of these other 

 sciences, his comprehension of the relation 

 existing between psychology and these other 

 sciences, made him an excellent student in 

 psychology. And this, I think, holds good 

 also for other studies, more or less. I there- 

 fore believe in varying credit. We ought not 

 to make conditions in college unnecessarily 

 different from conditions in life. In life too 

 our credits vary. 



The amount of credit which we give for 

 different scholarly accomplishments depends 

 on our view as to what is the distribution of 

 such accomplishments in a large number of 

 students. The division lines which we draw 

 between the different grades, as explained 

 above, also depend on our assumption of a 

 definite curve of distribution. I am inclined 

 to think that, owing to the insufficient data 

 which we possess at present, we have to base 

 our conclusions on the normal curve. Three 

 years ago I had the honor of speaking on this 

 very question before this section of the Amer- 

 ican Association. I then criticized certain 

 conclusions drawn by Professor Hall. Since 

 my criticism has been slightly misunderstood, 

 I wish to repeat what I wanted to point out. 

 I did not want to belittle Professor Hall's 

 work on the distribution of scholarly abilities. 

 But I held, and still hold, that Professor Hall 

 found his students distributed in accordance 

 with the normal curve simply because he be- 

 lieved, while he was teaching and examining 

 these students, that they ought to be dis- 

 tributed thus. I shall try to make this 

 clearer. 



Fig. 2 shows two entirely different curves 

 of distribution. You may be surprised that 

 these are the same students in the same sub- 

 ject. The only difference is that the broken 

 lines represent the outcome of a very difficult 

 examination, requiring a large amount of 

 reasoning; the continuous lines represent the 



outcome of an examination of the kind most 

 commonly given by instructors, which enables 

 most of the students to respond to a majority 

 of the points correctly, especially when there 



is no time limit, or when the time of com- 

 pleting the task — of such enormous signifi- 

 cance in life — is not taken into account at all. 

 College teachers usually assert that the curve 

 of distribution is not the normal curve, but a 

 skewed curve like that of the continuous lines. 

 They are usually ready to explain this by 

 referring to the elimination of poor scholars 

 in the high schools and lower schools. I Jiave 

 considerable doubts as to this elimination. 

 Is the work done in a high school really so 

 much like that done in college that there is a 

 large previous elimination of poor college stu- 

 dents? Our curves clearly show that the 

 skewed distribution which an instructor finds 

 is likely to be simply the result of the kind of 

 examination which he gives. If he believes 

 that the distribution ought to be like the 

 curve of the continuous lines, he will give his 

 examinations accordingly. If he believes 

 otherwise, he will give his examinations 

 otherwise. 



I do not admit, then, that anybody has 

 proved thus far that the distribution of schol- 

 arly accomplishments in college is like the 

 normal curve or like a curve skewed either 

 way. Under these circumstances, if for any 

 purpose we have to assume, at least provi- 

 sionally, a particular distribution, I see no 

 other possibility than that of regarding schol- 

 arly accomplishments as based on biological 

 properties and assuming the normal curve as 

 the most probable one. 



