The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Sampling and Weighting Webinar June 2015 » Modifying sampling weights
Modifying sampling weights [message #4355] Tue, 12 May 2015 11:11
baremma2002 is currently offline  baremma2002
Messages: 20
Registered: April 2013
Dear panelists,

I would be grateful if you could address the following questions:

1) Sampling weights are designed for the data to be representative at national or regional level and spatial modeling taking into acount sampling weights using the geoRglm package of R requires that those weights are integers, which is
not the case for DHS data. Then, how can we modify them in order to use them in geoRglm package?

2) We can link DHS data with external data. Are sampling weights to be dropped in this case?
3) Is it a good way to round them down, up or by default (which will bias the estimates) or else to multiply them by 100 for example and round them after in order to use them in geoRglm even if
multiplying the sampling weights by a factor will affect the uncertainty of the estimates (standard errors)?
Previous Topic: Clarification about webinar participation
Next Topic: Sample design, linear models and multilevel analysis
Goto Forum:

Current Time: Tue May 11 03:18:27 Coordinated Universal Time 2021