and only four improve their ranks and these to the extent of only two
places each. This would indicate a high degree of constancy. Two
different performances can be compared as above described. The abilities
on successive days can be determined by taking the average rank of the
first day and comparing it with the average rank of the second day.

If the effects of practice are to be studied, the experiments must be
kept up for many days, and each student's work on the first day compared
with his work on succeeding days. Then a graph can be plotted to show
the improvement from day to day. The average daily speed of the class
can be taken and a graph made to show the improvement of the class as a
whole. This might be plotted in black ink, then each individual student
could put on his improvement in red ink, for comparison. A group of
thirty may be considered as furnishing a fair average or norm in this
kind of performance.

In connection with this simple performance, making marks as fast as
possible, it is evident that many problems arise. It would take several
months to solve anything like all of them. It might be interesting, for
example, to determine whether one's speed in writing is related to this
simple speed in marking. Each member of the class might submit a plan
for making such a study.

The foregoing simple study illustrates the procedure of psychology in
all experimentation. A psychological experiment is an attempt to find
out the truth in regard to some aspect of human nature. In finding out
this truth, we must throw about the experiment all possible safeguards.
Every source of error must be discovered and eliminated. In the above
experiment, for example, the work must be done at the same time of day,
or else we must prove that doing it at different times of day makes no