[学术动态] 求助，一道统计实验设计的问题！ [推广有奖]

A veterinary researcher would like to study theeffectiveness of a new treatment for a type of bladder inflammation in cats. Ifuntreated the inflammation lasts about 7 to 10 days. Based on the results ofstudies with humans, the researcher hopes that the new treatment will shortenthe duration of the inflammation to about 3 to 5days. The outcome variable ofinterest is the duration of the inflammation measured in days. The inflammationwill be artificially induced by placing a mixture of alcohol and aspirin intothe bladder. Treated cats will receive the new medication; control cats willreceive a placebo medication. The researchers will be blinded; they will notknow which cats received which treatment until the experiment is complete. Theresearcher is considering the following two designs.

A: twenty cats will be used. Randomly assign 10 cats totreatment and 10 cats to control.
B: ten cats will be used. Randomly assign 5 cats to thegroup 1 and 5 cats to group 2. The cats in group 1 will receive the treatment;after a wash – out period of five weeks, they will receive the placebo. Thecats in group 2 will initially receive the placebo; after a wash – out periodof five weeks, they will receive the treatment.

The researcher strongly believes that there are no lastingeffects of this inflammation and that the five-week period is long enough to insurethat there will be no carry – over effects. Give an outline of how you wouldanalyze the data for each of these designs. Discuss the relative merits of thedesigns and make a recommendation to the researcher.

完全没有思路啊，请教各位大侠。