ESTIMATION AND TESTING OF FINER GRADATIONS 117 



has, then, a much broader range of information on 

 which to base his judgment of the pupil's intelli- 

 gence than has the experimenter, who has merely 

 taken a half -hour's time to record the response of 

 the child to a small number of tests. But, on the 

 other hand, there are some great disadvantages in 

 this method. The teacher does not, as a rule, stop 

 to consider on what concrete facts of observation 

 his judgment is based: the symptoms which de- 

 termine his judgment are not controllable as regards 

 their real importance; when the intelligence of sev- 

 eral pupils is compared, the comparison is based on 

 different facts that are not strictly comparable with 

 one another. Finally, the teacher is often by 110 

 means clear as to what he is to understand by l intel- 

 ligence' when he makes his decisions. 



These considerations are enough to show that the 

 estimation of intelligence is far from an easy matter, 

 that it is not something to be demanded of every 

 teacher as a matter of course. What is desired here 

 is to seek the middle road between two opposite 

 sources of danger: on the one hand we are threat- 

 ened with the danger that the teacher can not free 

 himself, when he estimates intelligence, from the 

 pedagogical habit of judging his pupils by their 

 schoolroom performance, in which event the rank- 

 order for intelligence will be no more than a copy of 

 the rank-order for class work, corrected in a few 

 points. If the teacher tries to avoid this tendency, 

 then there arises the other danger that he goes at 

 the selection blindly and that the resulting rank- 

 order becomes a mere product of chance. 



Evidently, the estimating of intelligence by the 

 teacher makes contribution not only to the psychol- 



