| error after svyset for higher spatial resolution [message #7026] |
Thu, 13 August 2015 10:35  |
fabienne
Messages: 6
Registered: March 2015
|
Member |
|
|
Hello everyone
I am working with DHS VI from Guinea (2012), particularly with household-level data and I use Stata for statistical analyses.
The smallest geographical frame in the data is the region (ADM-1), but I need data at higher spatial resolution (ADM-2). I have therefore reclassified the cluster location by préfécture using the GIS data. I thus now have an additional variable which identifies the préfécture for each household. The minimum number of clusters I per préfécture is 2.
My aim is to produce summary estimates of variables such as hv204 or hv207 at préfécture level.
My syntax is:
svyset hv021 [pw=weight], strata(hv022)
svy:mean hv204, over(prefecture)
However I get an error message after this command:
no observations
r(2000);
If I ignore the svy design, I get estimates for all préféctures, I suspect it must be related to the svyset command? Is it because in one préfécture I have only 2 clusters?
Or I have to adjust the weights, or the strata(hv022) or the PSU (hv021) and if yes, how would I need to change it?
Of note, the error does not occur if I use a categorical variable for tabulation (using svy: tab prefecture hv025)
Thank you very much.
Kind regards, Fabienne
|
|
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
