The same text is in the attached file.

QUESTIONS TO ANSWER
? Description of data analysis procedures (one paragraph)
o Selection/justification of your basic statistical approach
o Test(s) of assumptions underlying the statistical approach
o Any modification made based on your test(s) of assumptions

I would do a 2-sample t-test without equal variances assumed (you could do it with equal variances assumed, it probably wouldn't make a difference because the standard deviations of the groups are pretty similar). The test statistic is calculated as:

So, in this situation, t = (-1.1111 - (-5.0111)) = 3.9/√2.567 = 7.697
√(1.89888/18 + 2.72178/18)

This value for t is significant at the 0.001 level (2-tailed probability; compare to a t-distribution with 34 degrees of freedom). Therefore you can reject the null hypothesis that the means of group 1 and group 2 are equal, and assume the alternative hypothesis that they are different.

This test assumes that the two groups are normally distributed. This was shown with the Q-Q plots (the plots are linear) and with the Kolmogorov-Smirnov and Shapiro-Wilk tests (the p-values are large, so the null hypothesis of normality can not be rejected). Therefore, there are no modifications that need to be made.

? Description of findings (paragraph or two)
o Summary table comparing groups
o Accompanying narrative statement on CI of differences and t-test
o Description of effect size and power of the analysis
The summary tables are below (the simplest way to summarize the data is with the box-and-whisker plot). Group 1 has a larger mean (less weight loss?) and a smaller variance than group 2.

T-test: As was shown in the t-test above, the two groups have statistically significantly different means.

Confidence interval: To calculate the 95% confidence interval for the difference of the means, we use the following formula:

-1.1111 - (-5.0111)) ± (2.032)(2.3468)(.3333)
3.9 ± 1.588
(2.312, 5.488)
(the t-value of 2.032 comes from the 95% critical value for a t-distribution with 34 degrees of freedom)

Effect size: The desired effect was 1.92, and, as the entire 95% confidence interval is greater than 1.92, you can be confident that the actual difference between the group means is larger than 1.92 kg.

Power: At the top of the page, the power was calculated as 0.9979. The power equals 1 - β, where β is the Type II error rate ("false negative" error, i.e. the alternative hypothesis is actually true, but you do not reject the null hypothesis). Here, 0.9979 = 1 - β, so β = 0.0021. That means that there is a 0.21% type II error rate.