From Example 6.4 (page 148), for the cognitive behavioral therapy group the sample mean change in weight of 3.0 pounds was significantly different from 0. However, Example 7.8 (page 194) showed it is not significantly different from the mean change for the control group, even though that group had a negative sample mean change.

How do you explain this paradox? (Hint: From Sections 7.1 and 7.3, how does the se value for estimating a difference between two means compare to the se value for estimating a single mean?)