Hi
Which test would you recommend to look at the effect of gender on weight loss in the active group adjusted for the placebo group?
Thank you
If you have longitudinal weight versus time data,and it looks fairly linear in the active group, try a multiple regression of weight change on time, placebo/active (0/1), gender (0/1) and initial weight.
alternatively do separate studies in males and females and compare differences of regression slopes of weight with time between active and placebo in each study,
In this case gender is a nominal variable and the weight loss is a continuous variable. Since we can get the weight loss values based on gender, we can use independent t-test. So we run the mean weight loss of males against that of females and that will show whether the differences in weight loss can be affected by gender. Other wise we can rank the weight loss and make it an ordinal variable, eg weight loss difference of 5 kg is mild weight loss, 10 kg is moderate and above 15 kg is severe weight loss. Knowing the number of females and males who fell in these categories, we can use the Chi-square test.