PAPERS OX PSYCHOLOGY AND EDUCATION 4S5 



learned to read or not before the age of 8 years will 

 tend to modify 10 of the 26 tests between the 8 and 12 

 year groups, inclusive, with respect either to the actual 

 subject matter of the tests or the vocabularies involved. 

 Academic training varying from grammar grades to 

 college will modify 8 of the remaining 18 tests of the 

 scale. 



In summing up. let me say that the most pressing need 

 now is that analysis of tests which will enable us to know 

 what it is that we are testing in a given case. Our 

 knowledge of the result can be valid only in so far as we 

 know what our tests are. I have not mentioned here the 

 influence upon test scores or mental development of sen- 

 sory defects, certain generalized motor disabilities such 

 Elf -pastic paralysis, poor nutritional condition, poor 

 health or attitude and temperament. Each one of these 

 points could be elaborated into a long discussion. In 

 taking my examples from the Stanford Binet scale, I 

 do not mean to say that that is the only one concerned. 

 Performance scales are involved in their own way. I do 

 not at all think, either, a performance scale can take the 

 place in all respects of a scale testing ability with lan- 

 guage forms. The higher grades of intelligence can 

 not be tested with any performance scale yet devised, 

 though such scales may be used to test Avide ranges of 

 innate intelligence. The very general movement toward 

 the development of group scales is a false economy in the 

 development of mental testing. TVhat we need is more 

 intensive work on the side of analysis of the character 

 and quality of tests themselves and the conditions under- 

 lying their application. 



