The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Weighting data » v022, mv022, hv022 (the sample strata variables) in 2003 Burkina Faso stata dataset
v022, mv022, hv022 (the sample strata variables) in 2003 Burkina Faso stata dataset [message #3276] Sun, 16 November 2014 20:36 Go to next message
ykang12 is currently offline  ykang12
Messages: 8
Registered: July 2014
Location: USA
Member
To whom it may concern,
In the 2003 Burkina Faso stata dataset, v022, mv022, hv022 (the sample strata variables) are not available. How could I get these variables?
Thanks,
Yunhee
Re: v022, mv022, hv022 (the sample strata variables) in 2003 Burkina Faso stata dataset [message #3279 is a reply to message #3276] Mon, 17 November 2014 14:51 Go to previous messageGo to next message
Reduced-For(u)m
Messages: 292
Registered: March 2013
Senior Member


The standard way to stratify is to use the combination of "region" and "rural/urban". You can generate these as:

egen strata = group(v024 v025), label

This comes from the FAQ under "using data files": http://dhsprogram.com/faq.cfm
Re: v022, mv022, hv022 (the sample strata variables) in 2003 Burkina Faso stata dataset [message #3285 is a reply to message #3279] Mon, 17 November 2014 18:30 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 772
Registered: January 2013
Senior Member
Just a clarification that the standard way is to use the strata used at the time of selection of the sample, and you can usually work this out by referring to Appendix A in the survey report. In many DHS surveys the strata are urban and rural areas within regions, and the example given is correct in those cases, but not always. I confirmed that this survey did actually use urban and rural areas within each region as strata, so the code is good for this survey.
Re: v022, mv022, hv022 (the sample strata variables) in 2003 Burkina Faso stata dataset [message #3315 is a reply to message #3285] Thu, 20 November 2014 05:33 Go to previous messageGo to next message
ykang12 is currently offline  ykang12
Messages: 8
Registered: July 2014
Location: USA
Member
It is clear now and thank you so much!
Yunhee
Re: v022, mv022, hv022 (the sample strata variables) in 2003 Burkina Faso stata dataset [message #4323 is a reply to message #3315] Fri, 08 May 2015 08:25 Go to previous messageGo to next message
albena is currently offline  albena
Messages: 6
Registered: February 2015
Member
Dear DHS Forum,

I've been following the discussion on how to combine DHS data sets over time and for different countries. I however couldn't find an answer to my problem but this thread seems to be quite related to it.

I'm working with around 64 children's recode data files for different countries and different points in time and facing the problem that in some cases the strata haven't been reported (e.g., Burundi 1987) and instead there's only the v101 "region" variable, or reported as "country specific" (e.g., Burkina Faso 2010). That's why I decided to use as a strata variable the grouped "region" and "place of residence" variables (v024 and v025). It's however a very tedious work to adjust the number of regional divisions for the same country at different points in time. The regional division is important for a comparison over time. I don't know whether it would otherwise be correct to just use the "region" variable for the same country as it is at point 't' and use it as it is at point 't+1', without being interested into how many regional parts was the country split during both periods. After all, no matter the regional division, the surveys have a representative sample for the country as a whole.

One rather general question regarding the DHS recodes. I saw that for the DHS from the 80s there is only a household recode (HH) available but not a household members recode (PR). I was wondering if I still could use these to look at the years of schooling of the household members?

Thanks a lot for your help!
Albena
Re: v022, mv022, hv022 (the sample strata variables) in 2003 Burkina Faso stata dataset [message #4385 is a reply to message #3276] Mon, 18 May 2015 11:41 Go to previous message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 772
Registered: January 2013
Senior Member
The strata variables were only created starting in DHS phase II (roughly 1990 onwards), so you won't find them in any DHSI survey.

I'm not sure what you mean by "adjust the number of regional divisions for the same country at different points in time". Can you explain?If you just use region, then you won't be fully taking into account the stratification, which in most cases is urban/rural areas within region. Also note that regions change over time, and in some cases, the region may have the same name, but the borders of the region are not the same. See the spatial data repository for more information about boundaries for each survey.

On the question of using the HH datasets, note that these are not recode datasets, but are households that mirror the original household questionnaires for survey in DHS Phase I. For the most part these do not contain much information on household members, but where questions on education do exist, you could use these to look at years of schooling.
Previous Topic: Generating new variables in dataset with weighting
Next Topic: Weighting Data without SPSS/Stata
Goto Forum:
  


Current Time: Thu Jun 30 21:43:20 Coordinated Universal Time 2022