Example 8 in Section 9.3 described a study that used a cognitive behavioral therapy to treat a sample of teenage girls who suffered from anorexia. The study observed the mean weight change after a period of treatment. Studies of that type also usually have a control group that receives no treatment or a standard treatment. Then researchers can analyze how the change in weight compares for the treatment group to the control group. In fact, the anorexia study had a control group that received a standard treatment. Teenage girls in the study were randomly assigned to the cognitive behavioral treatment (Group 1) or to the control group (Group 2).The figure shows box plots of the weight changes for the two groups (displayed vertically). The output shows how MINITAB reports inferential comparisons of those two means.

T-Test of difference = 0 (vs not =):
T-Value = 1.67 P-Value = 0.102 DF = 50

a. Report and interpret the P-value for testing H0: μ1 = μ2 against Ha: μ1 ‰  μ2.
b. Summarize the assumptions needed for the analysis in part a. Based on the box plots, would you be nervous if you had to perform a one-sided test instead? Why?
c. The reported 95% confidence interval tells us that if the population mean weight change is less for the cognitive behavioral group than for the control group, it is just barely less (less than 1 pound), but if the population mean change is greater, it could be nearly 8 pounds greater. Explain how to get this interpretation from the interval reported.
d. Explain the correspondence between the confidence interval and the decision in the significance test for a 0.05 significance level.

MINITAB Output Comparing Mean Welght Changes group cog behav 29 control Difference =hu (cogbehav)-mu (control) Estimate for difference: 3.45690 958 CI for difference: -0.70643, 7.62022) N Mean StDev SE Mean 1.4 1.6 3.01 7.31 26 0.457.99 - cog.behav control group