In a completely randomised experiment, there are two potential reasons for variability in
the response.
The different factor levels
Changing the value of a factor can result in changes to
the response. The purpose of the experiment is to assess the magnitude of
these changes. This is called explained variation since
it can be explained in terms of the explanatory factor.
Other differences between the experimental units
There is usually natural variability between the experimental units and
in the measurement process. Two response measurements at the same factor level will usually differ. This natural variability
cannot be explained in terms of the factor and is called unexplained
variation.
The distinction between explained and unexplained variation is critically important.