The DHS Program User Forum - RDF feed
https://userforum.dhsprogram.com/index.php
clustering standard error while using IR file
https://userforum.dhsprogram.com/index.phpindex.php?t=rview&goto=23518&th=11279#msg_23518
I am analysing the effect of a policy (that was passed at a regional level) on women characteristics. I am using a combined DHS women file (IR) collected across rounds.
My question is: Initially I adjust Standard error by regions due to the fact that women's in the same region are exposed to the same shocks. I now realise that more than one woman was interviewed in a household. The household number is available in the IR file and I can see there are 2 OR MORE women's in the same household. So, I have to take into account outcome correlation within a household. Can I just use the household number as my cluster variable?
Kind regards
]]>id709nvz2021-09-29T21:28:55-00:00Re: clustering standard error while using IR file
https://userforum.dhsprogram.com/index.phpindex.php?t=rview&goto=23522&th=11279#msg_23522
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:
We stand by our usual recommendation for calculating standard errors. There are relatively few households with more than one eligible woman. The effect of an adjustment for household clustering would be negligible.]]>Bridgette-DHS2021-09-30T13:29:17-00:00