January 24, 1913] 



SCIENCE 



135 



tion of men in mental traits up and down 

 from a common medioci-e condition and the 

 wide variability of the same person's per- 

 formance in the same test. 



In the case where two groups are com- 

 pared in respect to some condition (call it 

 T) and in respect to some possible symp- 

 toms thereof (call them 8,^, So, ■••, 8n) it is 

 almost imperative that the amounts of T, 

 S-^, So, etc., possessed by each individual be 

 at least roughly determined. Unless this 

 is done we can not tell how symptomatic of 

 T each S is, nor compare the significance of 

 these S's with others tested in some other 

 investigations; we can only say of each of 

 them that it is or is not, in some undefined 

 degree, symptomatic of T; and perhaps 

 that it is more or less so than others of the 

 ;Sf's tested in the particular investigation. 

 For example, in testing tests for general 

 mental ability, Mr. Terman, who merely 

 chose five very bright boys and five very 

 dull boys without determining how bright 

 or how dull they were, has left all his re- 

 sults undefined in quantity and incompar- 

 able with any one else's. An otherwise ad- 

 mirable study has thus had far less influ- 

 ence on educational diagnosis than it should 

 have had. It is also desirable that the dif- 

 ferences within either group in respect to 

 T be small relatively to the difilerence be- 

 tween the central tendencies of the two 

 groups, and that the extent to which each 

 S separates the first from the second group 

 be stated in an exact and commensurate 

 quantity. 



Unfortunately, the great majority of 

 studies of sane versus insane, ordinary 

 versus feeble-minded, bright versus dull, 

 color-blind versus color-keen, musical versus 

 non-musical, and the like, have left one or 

 more of these three requirements unful- 

 filled. 



I can not quote a study that is beyond 

 reproach in all these respects, but the fol- 



lowing facts from Dr. Simpson's test of 

 tests of general intellectual ability will il- 

 lustrate them roughly. Dr. Simpson de- 

 fines the first of his two groups by the fact 

 that they were all teachers or graduate stu- 

 dents in a university. The group com- 

 pared with them were all "men in New 

 York City who had never held any position 

 demanding a high grade of intelligence," 

 who spoke "English as their mother 

 tongue. ' ' 



Two of them were persons earning comfortable 

 livings for their families, but men recognized by 

 their associates as being dull. Eleven others were 

 staying at the Salvation Army Industrial Home at 

 the nominal salary of $1.00 per week in addition 

 to board and room, until work could be secured. 

 One of these held a somewhat responsible position 

 at the time, acting as assistant superintendent of 

 the home. He stood high in the most significant 

 tests. The remaining seven were found in a mis- 

 sion on the Bowery where they were being helped 

 somewhat until they could find employment. 



As a measure of the extent to which each 

 symptom or test separates the "Poor" from 

 the "Good" groups, the degree of over- 

 lapping is taken. For example, the facts 

 for four tests — giving the opposites of 

 words, memorizing lists of words, marking 

 the A's on a sheet of printed capitals, and 

 discriminating lengths — were: The point 

 reached by 50 per cent, of the "Good" 

 group was reached by none of the "Poor" 

 group in the opposite test or in memory of 

 words; by 15 per cent, of them in the A 

 test ; and by 33 per cent, in the case of dis- 

 crimination of length. The point reached 

 by 76^ per cent, of the "Good" group was 

 reached by none of the "Poor" group in 

 the opposite test ; by 5 per cent, in the case 

 of memory of words ; by 20 per cent, in the 

 case of the A test, and by 62 per cent, in 

 discrimination of length. The percentages 

 of the "Poor" group reaching the points 

 reached by 88 per cent, and by 94 per cent. 



