Dear Naresh

Please visit the BA Superhelp Desk http://ssc.bibalex.org/helpdesk/introduction.jsf.

It is a global initiative by The Library of Alexandria, and the WHO Collaborating Centre at the University of Pittsburgh are launching a project to improve productivity in developing countries by establishing a Research Methods Help Desk.

Regards

thank you very much Sir!

If measurement is done by personal Interviews I would recommend random walk techniques stratified by daytime

I will assume that the slum is relatively small, and that going from one end of it to the other is not a big issue.

If you have a census of the slum from where to select your random sample, then that will be the best strategy, as it usually requires a smaller sample size.

If you don't already have a census of the slum, it may then be cheaper to do cluster sampling, and then conduct a census only in the clusters selected (and assume that the cluster not selected are similar to the ones selected). If as you say clusters are too homogeneous, then you will need a larger sample size, but still the study can be cheaper than the cost of a full census plus simple random sampling.

I am also trying to do study on the same. I prefer to go for Stratified sampling based on distance from government health institutions. The notified slum list are framed into two strata then by Simple random sampling, choose the no of slums you want.

For HH selection, still i am thinking how much sample size i must take.