532 



SCIENCE 



[N. S. Vol. XL. No. 1032 



than like those of beginners in graduate work. 

 The teacher thus develops in himself the illu- 

 sion that his average senior, however illogical 

 this is, stands above the average of his own 

 group, and that all the seniors deserve un- 

 usually high marks, that is, in comparison 

 with freshmen. But let these seniors enter 

 the graduate school, and some of them will 

 be found, by the different kind of test there 

 employed, to be almost incapable of doing any 

 graduate work at all, because they are deficient 

 in originality, initiative, resourcefulness, what- 

 ever you call it, in their chosen line. 



This tendency to compare freshmen and 

 seniors is so deep-seated that there is no hope 

 of eradicating it by simply calling attention 

 to it. As in college, so you find it in the 

 high school. My former colleague in Missouri, 

 Professor C. Alexander, found in an (unpub- 

 lished) investigation of the grading of high 

 schools, that the freshmen are reported partly 

 as average scholars, partly as superior, and 

 partly as inferior; but the seniors, there, too, 

 are reported mostly as high-grade scholars. 

 The low-grade scholars are said to have been 

 eliminated. Now some of these high-grade 

 scholars, obviously not the worst, enter the 

 state university. One should think, then, that 

 our college teachers in the freshmen classes 

 would find it a difficult task to separate from 

 this whipping cream any more plain milk. 

 But the contrary is true. Our teachers com- 

 plain constantly of the poor scholarship of 

 these " selected " college freshmen. 



All this shows, by the way, how unfounded 

 the statement is which we hear again and 

 again that the normal, i. e., symmetrical, curve 

 of distribution is inapplicable to college stu- 

 dents because they are supposed to be a 

 selected group. Only then would the s,ym- 

 metrical curve of distribution be inapplicable, 

 if the college freshmen under consideration 

 had been selected by freshmen tests from col- 

 lege freshmen, or if the college seniors had 

 been selected, by tests appropriate to seniors, 

 from the entire group of seniors. There is no 

 reason why the symmetrical curve should be 

 inapplicable to the entire group of freshmen, 

 or to the entire group of seniors, or to the en- 



tire group of graduate students or to any group, 

 provided only that the group is complete as a 

 group. That the group came into existence by 

 selection from a different group does not seem 

 to matter when each new group is confronted 

 with new kinds of tasks. There are those who 

 say that it is easy to prove, by examination 

 tests of the ordinary, traditional type, that 

 college students must be regarded as a selected 

 class^ in the sense that their distribution is 

 not represented by a symmetrical, but by a 

 skewed curve. I have already, a few years ago, 

 called attention to the f acf that such examina- 

 tions are unreliable. Simply make the exam- 

 ination difficult and set a time limit; the curve 

 appears skewed one way, most of those tested 

 crowding in the direction of low ability. 

 Make the examination easy and abolish or 

 greatly extend the time limit ; the curve ap- 

 pears skewed the other way. I offer to prove 

 at will by an examination left to my choice 

 that any group of students is distributed either 

 way. Just tell me in advance which way you 

 want the curve skewed. 



Por the practical problems of college admin- 

 istration this question as to the exact nature 

 of the curve of distribution is really of minor 

 importance. If, however, we just have to 

 make an assumption, it is safest to assume the 

 symmetrical normal distribution. We have 

 assumed in Missouri that the distribution is 

 either normal or very nearly so and experi- 

 enced no inconvenience. We have reduced 

 the lack of uniformity between teachers to 

 one third of its former amount simply by 

 the adoption of scientifically justifiable defi- 

 nitions, and a reduction to that amount is 

 worth while. But to eradicate the last third 

 is a complex problem of the future, so com- 

 plex that it may never be completely solved. 

 As has been indicated, it seems to involve 

 problems of our whole educational system and 

 even of the broader social organization of the 

 nation. 



Max Meyer 



3 Compare the two tendenoies, conflicting with 

 each other, according to Cattell, Fopular Science 

 Monthly, 1905, p. 372. 



* Science, 33, p. 667, 1911. 



