| Regions and MultiLevel Models [message #19267] |
Tue, 19 May 2020 08:27 |
Yawo
Messages: 45
Registered: February 2019
|
Member |
|
|
Hello,
I am involved in a research project examining HIV risk behaviors, pooling data from recent surveys in 14 sub-Saharan African countries.
I plan to fit a 3-level HLM model , with individuals nested in regions , which are in turn nested within countries (individuals --> regions --> countries). My concern is which variables to use for the level 2 (region) differences.
DHS provides two possible options:
v023: stratification
v024: regions.
The v023 stratification variable is already used to weight (svyset) the data, so I am not if I can use it again as a proxy for regional/community differences.
The v024 would have been appropriate, except when I pool the data, the regions codes get messed up, since each country numbers the regions sequentially 1, 2, 3, 4 with unique labels. Once pooled, these variable needs to be regenerated into a long a list of regions, each with unique codes. This is tricky, and can be accomplished with some coding. But before doing so, I wonder if:
(1) there is/are other ways to recreate the region variable to ensure uniqueness across cases
(2) there is any other variable to use to capture community differences in my HLM model (at level 2)
(3) it is ok to use the v023 stratification variable as a proxy.
Thanks in advance for your attention.
cY
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
