Re: How would I use weights in this scenario? [message #11407 is a reply to message #11342] |
Tue, 13 December 2016 15:29 ![Go to previous message Go to previous message](/theme/default/images/up.png) |
Bridgette-DHS
Messages: 3139 Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Senior Research Associate, Wenjuan Wang:
Quote:From your previous description, I understood that you want to examine how access/quality of healthcare is associated with infant mortality and you plan to link DHS and SPA data at the regional level, right? If this is the case, I see the following major steps of the analysis and the weight variable you would use in each step.
You first need to generate health service variables of interest at the regional level based on SPA datasets. For example, you want to look at the availability of C-section services in the region, you may generate this indicator: proportion of health facilities in the region that offer C-section. When generating this indicator, you will use SPA facility weight. Sometimes you also want to use provider data, for example, % of facilities with at least half of the providers who provide delivery services trained in C-section. You will need to use both provider weight and facility weight to create this indicator.
After you prepared these SPA indicators, you then merge them into the DHS dataset, for example, BR file using the region variable. You do need to make sure the region categorization is the same between DHS and SPA for the same country. They are not always the same.
After merging, let's say you use survival analysis to test the association between healthcare access variables and risk of death before one. You will apply the DHS weight.
Hope this helps!
|
|
|