s made in the fields of sensation and memory, where good
experimental procedure has been developed.

Another general line of attack worthy to be mentioned alongside of the
experimental is the _comparative method_. You compare the actions of
individuals, classes or species, noting likenesses and differences.
You see what behavior is typical and what exceptional. You establish
norms and averages, and notice how closely people cluster about the
norm and how far individuals differ from it. You introduce tests of
various sorts, by which to get a more precise measure of the
individual's performance. Further, by the use of what may be called
double comparison, or "correlation", you work out the relationships of
various mental (and physical) traits. For example, when many different
species of animals are compared in intelligence and also in brain
weight, the two are found to correspond fairly well, the more
intelligent species having on the whole the heavier brains; from which
we fairly conclude that the size of the brain has something to do with
intelligence. But when we correlate brain weight and intelligence in
human individuals. {15} we find so many exceptions to the rule (stupid
men with large brains and gifted men with brains of only moderate
size) that we are forced to recognize the importance of other factors,
such as the perfection of the microscopic structure of the brain.

Tests and correlations have become so prominent in recent
psychological investigation that this form of the comparative method
ranks on a par with the strict experimental method. A test is an
experiment, in a way, and at least is often based upon an experiment;
but the difference between the two lines of attack is that an
experiment typically takes a few subjects into the laboratory and
observes how their mental performances change with planfully changed
conditions; whereas a test goes out and examines a large number of
persons under one fixed set of conditions. An experiment belongs under
what we called "general psychology", and a test under "differential
psychology", since the first outcome of a test is to show how the
individual differs from others in a certain respect. The results may,
however, be utilized in various ways, either for such practical
purposes as guiding the individual's choice of an occupation, or for
primarily scientific purposes, such as examining whether intelligence
goes with brain size, whether twins resemble each other as