Hi, I do not understand what you mean "managed to correct for the oversampling and weights". If you mean that you modified the sampling weight in the DHS data for correcting oversampling, this may not be an legitimate correction. It is true that in almost all DHS surveys, oversampling and under-sampling may happen in certain areas of the country because of the limited total sample size and the domain sample size requirement. But the oversampling and under-sampling are already taken into account in the sampling weight calculation. If you use the sampling weight presented in the DHS data file, there is no worry at all. By contrary, if you do further corrections of any kind by intending to correct the oversampling or under-sampling, you may introduce a bias in your analysis and which may result invalid statistical conclusions. No DHS document or report suggest any treatment for oversampling or under-sampling, so I am curious where you get this point.
Regarding the Indian DHS 2005-2006, the only difference compared to most of the other DHS surveys is that sampling in the urban areas was a three stages sampling, while most of the DHS surveys use a two stages sampling. Again, there is no need to correct oversampling in any way as explained in above.

Thanks.

Ruilin