When weights are not supported [message #3188] |
Mon, 03 November 2014 20:06 |
Yohannes
Messages: 6 Registered: July 2014
|
Member |
|
|
I have DHS data for over 30 countries which I plan to use for a multi-country analysis using a STATA platform. I have already de-normalised the weights (following DHS suggestion) but the model I want to fit, multi level probit/logit model, does not support weights. The question I have is what do you do when the platforms you use for analysis do not support weight, but you know that ignoring weights is not an option because sample sizes vary enormously between countries, and in some cases countries that have half the size population than the larger countries have two or three times larger samples.
|
|
|