anemia among lactating mothers [message #17468] |
Wed, 27 March 2019 02:31 |
meseret
Messages: 3 Registered: March 2019 Location: Ethiopia
|
Member |
|
|
Hello,
I am using the 2016 Ethiopian DHS to see the spatial distribution and factors associated with anemia among lactating mothers. However I am having difficulties matching the total number of lactating mothers with the final report table which is 4554,but on individual women data set the number/sample size is 4252 ,so how can I equalize number/sample size?
my second question is steps of DHS data manipulation using stata,I mean which comes first merge,sample weight,recode,drop case/observation
third,which types of sample weight can I use simple or complex?
forth how can I prepare excel event data from GPS data which contain anemia prevalence,latitude and longitude?
fivith how can I know and exclude cluster which had no GPS data?
Thank you,
|
|
|
|
Re: anemia among lactating mothers [message #17634 is a reply to message #17486] |
Wed, 01 May 2019 12:07 |
Bridgette-DHS
Messages: 3203 Registered: February 2013
|
Senior Member |
|
|
Following is a response from: DHS Lead Nutrition Research Associate Rukundo Benedict and DHS Senior Geospatial Analyst, Trinadh Dontamsetti:
To match the total number of lactating women with anemia use the code below. Remember, you have to ensure that the women are not pregnant and you will also need to apply weights.
use " ETIR70FL.DTA"
*Is woman pregnant?
ta v213
*Anemia variable
ta v457
*Breastfeeding variable
ta m4_1
*lactating women with anemia
gen wt=v005/1000000
ta v457 if m4_1==95 & v213==0 [iw=wt]
Any data manipulation will depend on your research question. If all the variables you require are in the IR file, you do not need to merge with another data set. Similarly, because of the two-stage sample cluster design of the surveys, it is important to use the sample weights. In your analysis you can decide to use complete cases and thus may drop those without the required variables. I suggest you review the DHS guide to statistics as it covers many of your questions.
You will have to download GPS data by making a request on The DHS Program website. This requires a short description of the project and a justification for requesting cluster-level GPS data. Once approved, the data will be sent in the form of a shapefile, which includes a DBF table file. This can then be imported to a CSV. The cluster numbers in this DBF table will match exactly to the cluster numbers from the survey dataset, meaning users can attribute anemia prevalence to the cluster (which already comes with latitude and longitude information).
Any clusters whose GPS could not be verified are marked MIS for missing and have altitude values set to 9999 and lat/long values set to 0,0.
|
|
|