For the year 2015, a lot of decision variables like DECFAMVISIT, DECLEAVETOWN have a lot of entries in the not in universe. I wanted to clarify it is because this data is only asked from specific houses which are included for state modules. My study focuses on these decision variables as the variables of interest. Should I drop variables which are NIU or is there another way to resolve the issue. Would be very thankful for your help. Also wanted to confirm this happens only for 2015 not for previous surveys ?
You’re correct that these variables (at least the two you listed) are asked only in the state module. You can see this on the variable’s Universe page. You will only be able to use these variables in analyses of individuals who participated in the state module, and you’ll want to drop the NIU cases. You’re also correct that this is only the case for the 2015 survey.
IPUMS DHS is planning to make the variable SSMOD available in the near future, which indicates whether households were in the state module. In the meantime, you can pull this variable from the original DHS data (available through dhsprogram.com) and merge it with your IPUMS extract using the instructions found here.