We wanted to test the effect of community level sanitation intervention(CLS) integrated with nutritional supplementation(NS) on child growth in rural poor settings. We proposed factorial design where children are randomly grouped in to four groups(control, CLS, CLS+NS and NS) and will be followed for a total of 9-months. First the community will be randomized to receive CLS. i.e. A1( for those who receive and A0 for those who do not receive) . Then, within each village,households will be randomly assigned to control or to nutritional supplements [CLS will be provided at community level and NS will be at individual level]. Is it possible to estimate the sample size using simulation techniques given the above complex design and assumed multiple testing/intervention? 2). How could we consider attrition rate/missing values? If not, how one can handle such complex design using conventional method?

Thank you in advance,

Sincerely,