64 HAZEL M. STANTON 



is evidently due to the natural development and growth of children schooled 

 through the twelve grades of our day schools. 



The norms made from unselected groups for the transfer of raw scores 

 to centile ranks are school grade norms rather than age norms. Since raw 

 scores are naturally higher for children as they advance in school grade, the 

 norms for each school grade offer an interpretation of that developmental 

 growth at successive periods commensurate with the natural increase. 

 The process might be compared to the known fact that a child who is tall 

 when young tends to be tall during successive periods of his growth. His 

 height increases with growth, but when he is compared at successive ages 

 to a group of similar age his height has increased at sufficient rate to classify 

 him as tall at those successive periods. Have we a similar tendency of 

 imcrease in test scores such that they increase at sufficient rate to classify 

 a pupil high in capacities at successive periods, low in capacities, or average, 

 et cetera? 



Our study of 146 re-retests, those children who have been measured three 

 times with two three-year intervals occurring between tests over a period 

 of seven years' development for each child, and about one dozen pupils who 

 have been measured four times with three three-year intervals over a 

 period of ten years' development, will give us further light on this subject, 

 but no attempt will be made in this paper to present data for pupils with 

 three and four tests. 



With the brief discussion of raw scores for the children's groups, we will 

 now consider the talent profiles of the retest groups. The three groups of 

 children's retests show the following variation in talent profiles, each group 

 profile of Ti and T2 representing the average centile ranks. Ti is the solid 

 profile and T 2 , the dotted profile. Notice the interlacing of the Ti and 

 T 2 profiles in each group. The centile rank variations in each test are 

 negative as well as positive. On the whole, the average talent profile for 

 each group tends to remain in the same area for T x and T 2 . These data 

 suggest the tendency for each pupil's talent profiles for Ti and T 2 to be 

 similar; for example, if a child has high capacities, average capacities, or 

 low capacities when in the fifth grade of day school, that child tends to have 

 the same capacities when in the eighth grade of day school. Such a pre- 

 diction holds true from the viewpoint of talent profile classifications for 

 54 per cent of the children's groups. In other words, 54 per cent of the 

 pre-adolescent, adolescent, and post-adolescent groups have the same talent 

 profile classification in the restests as they did in their first tests, 43 per cent 

 of the groups had a classification either one step higher or one step lower, 

 so that 97 per cent had the same or similar talent profile classifications in 



