We're designing a controlled before-after study for which we'll use difference-in-difference in child growth stunting (in kids under five years) as the main outcome. We have 8 intervention areas and plan to select four control areas matched on relevant (area level) indicators. The intervention consists of an integrated package of community and clinic based care and community education. We have baseline info on the current stunting rates in the area, i.e. 40%, and assume stunting in control areas to be slightly lower at 30% (but will conduct a baseline to assess).
We assume stunting in control areas to reduce with 5%, to 25%, over the (4 year) time line of the study. We are interested in detecting a minimum diff-in-diff effect of 10%, so a reduction in the intervention group from 40 to 25%. The baseline and end-line measurements will be in different populations, as the measurement is done in age bound category of 0-4 year old, i.e. these are two cross sectional measurements.
Can anyone advise on calculating sample size for this? Or refer me to a calculator?