25' failures per pupil there is an increase of only 1
per cent over that for failures. The number of failures does not seem
directly to condition the pupil's ability to graduate or to continue to
in school.

5. TIME EXTENSION FOR THE FAILING GRADUATES

We shall now inquire further what extension of time for graduating
characterizes the failing graduates in comparison with the non-failing
ones.

The distribution according to the period for graduation for the 1,936
pupils who graduate was shown by the summary lines of Table VIII. In
the same table the non-failing graduates are included (but distinct).
No pupil graduates in less than three years and none takes longer than
six years; 9.8 per cent of the number finish in less than 4 years; 19.7
per cent take more than 4 years. The small number that finish earlier
than four years may be due in part to the single annual graduation in
several of the schools. Some of the schools admitting two classes each
year graduated only one, and the records made it plain that some pupils
had a half year more credit than was needed for graduating.
Considering, however, that about 42 per cent of the graduates had no
failures, they should have been able to speed up more on the time
period of getting through. They were doubtless not unable to do that.
But some principals hold the conviction that four years will result in
a rounding out of the pupil more than commensurate with the extended
time. More than 35 per cent of those who did finish in less than four
years are graduates who had failed from 1 to 11 times. In the
conventional period of four years 77 per cent of the non-failing and 64
per cent of the failing graduates complete their work and graduate (see
p. 59, for the means employed). The percentages of non-failing
graduates for each time period are given below.

THE PERCENTAGES OF NON-FAILING GRADUATES FOR EACH PERIOD

Time Period in Years 3 1/2 4 1/2 5 1/2 6
Per Cent of Non-Failing 80.4 50.0 46.5 19.3 13.3 .. ..

This continuous decline of percentages representing the non-failing
graduates shows that they have an evident advantage in regard to the
time period for graduating. Their percentages are high for the shorter
time periods and low for the longer periods. But by reference to Table
VIII we quickly find that the slight extension of the time period for
the failing graduates is not at all commensurate with the number o