A study16 of treatments for obesity for rural women examined the impact of a six-month behavioral weight loss program delivered by phone either one-on-one with a counselor or to a group via conference call. A sample of rural women classified as obese by their BMI were randomly assigned to the two conditions. Table 7.24 shows software results of twosample comparisons of mean weight loss (in kg) over the six months.

(a) State the assumptions on which each significance test is based, and explain how to interpret the results.

(b) Assuming equal population standard deviations, a 95% confidence interval for the difference in mean weight changes for group versus individual treatments is (1.45, 9.35). Interpret.
TABLE 7.24 t-test for Equality of Means Sig. Mean Std. Error t df (2-tailed) Difference Difference wtloss Equal variances 2.82 25 0.008 5.4 0.522 assumed Equal variances 2.91 23 0.009 5.4 0.538 not assumed