The second problem with interpreting results such as those shown in Figure 14.1 is that the apparent difference between the before-treatment observations and the after-treatment observations may simply be the result of chance.
Notice that there is variability in the day-to-day observations; this variability is a natural part of behavior and measurement.
Although the results appear to suggest a pattern of higher scores before treatment and lower scores after treatment, the “pattern” may be nothing more than normal variability. You may recognize this problem as the traditional question of statistical significance.