§ 10. Method of Comparing Defective Children with Ordinary Children We now have two series of grades in the same measurements, one set from mentally defective children and the other from ordinary school children. The usual method of comparing such results is to compare the records of one set of individuals with the central tendency of those of the others of the same age and sex. But in this case there were not enough defectives of any age to make the results gained from such treatment of any value, consequently a different method had to be adopted. The method used in dealing with the majority of the measurements was one which enabled me to compare the records of all of the defectives with those of all the ordinary children without restriction as to age or sex. Another very decided advantage is the fact that the units of grading are identical throughout all the measurements, as will be evident from the following description. - The difference between the record of each defective in any test and the median for an ordinary child of the same age and sex was found. This difference was then transmuted into positive or nega tive multiples of the probable error as the case required. For in stance, referring to Table III., No. 23, whose age is 12 years 7 months has a mark of 12 in the second opposite test. Table XXII. gives th median for that test at that age to be 15.4 + (the actual figures by interpolation being 15.6); the difference therefore is 3.6. Th probable error in the standard for that age is <3.0 (2.9), the - 3.6 is 1.249 times the probable error. Hence individual 2 is 1.25 times the probable error below the grade reached by 50 pe cent. of ordinary children of the same age and sex in this test. B thus transmuting the difference between the grading received b defectives and ordinary children respectively in every test into mu tiples of the probable error of the appropriate age and sex I ca compare the records of the 150 defectives tested with the 500 or 6 ordinary children, just as if I had 150 idiots and 600 school childr all of the same age and sex. Not only by this method can I consid all my cases together, but each test is, so far as is possible, co parable with every other, irrespective of whether the trait examin is physical or mental. This, so far as I know, has not yet be done. For instance, a mark of 0 in height, -2.5 in a memory t -4 in the part-whole test and +1.19 in the measure of perc tion, means that the individual was just at the median of children general as to height; nearly twice as far below the median in the of intelligence as in the memory test; while in the test of percept |