Smith's data showed almost precisely the same figures as those of Dearborn. From the Liberal Arts class of 1910 (one hundred and sixty students) those were chosen whose records were complete in both high school and university. This gave a total of one hundred and twenty students. On the basis of their standing, as based on the grades assigned in all subjects studied, they were ranked in order for each year of high school and university. They were then separated into quintiles on the basis of these rankings, and their standing in these various quintiles observed from year to year.
When the students, on the basis of their general high school average (for the four years), are distributed through their respective quintiles in the university (general average again) the results are as shown in the table on page 183.
SHOWING THE RELATIONS BETWEEN HIGH SCHOOL RECORDS AND UNIVERSITY RECORDS (SMITH). _See Text for Explanation_
------------------------------------------------ | University Average H. S. Average|---------------------------------- | 1st | 2nd | 3rd | 4th | 5th |Quint.|Quint.|Quint.|Quint.|Quint. -------------+------+------+------+------+------ 1st Quintile | 54% | 17% | 17% | 4% | 8% 2nd Quintile | 25% | 29% | 17% | 13% | 16% 3rd Quintile | 17% | 25% | 20% | 21% | 17% 4th Quintile | 0% | 25% | 25% | 33% | 17% 5th Quintile | 4% | 4% | 21% | 29% | 42% ------------------------------------------------
In considering this table it is apparent that if the high school students were distributed through the various university quintiles on a purely chance basis, and without any reference to their high school records, there would tend to be twenty per cent of each high school quintile in each of the university quintiles. Any percentage higher than this twenty per cent thus indicates some significant relation between the two sets of grades. On the whole there is a close relation indicated. The tendency is clear for those in a given high school quintile to be found in or near the same quintile in their university work. The relation is particularly close in the highest and lowest quintiles. In the intermediate quintiles there is more or less shifting about.
In the same way it is possible to classify all students in quintiles during their first high school year, and then to trace their careers through the following three years of high school and four years of college. The following tabulation shows the results when this was done. The figures show the percentage of each quintile in first year high school who were found in the same quintile in the various later years.