[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: DHS Ghana merging question

From   Evans Jadotte <>
Subject   Re: st: DHS Ghana merging question
Date   Thu, 05 Nov 2009 16:14:33 +0100

Somebody else worked on that issue and probably solved it. Maybe he can help.


Amadou DIALLO wrote:

Normally, there is a doc file contained in the zipped child anthro
file that describes how to merge (if I recall, it is said to merge
with the child KR dataset). I am not sure you use the correct
variables for merging. I thinks also you should rename anthro
variables in the women IR and children KR files before attempting to
merge. Please notice that the women (what DHS team calls "individual")
files contain information about "partner/husband" educational level.

Since I am working extensively with DHS datasets, please contact me
off network so that I can try to help you.

Best regards.


2009/6/22, Friedrich Huebler <>:

In your message of June 11 you described how you are merging the
household member data with the anthropometric data. When I reproduce
your approach the data from all household members, adults and
children, appear in the merged file.

You write now that "only information for children is kept" but without
seeing the Stata commands you used it is difficult to understand what
you are doing differently.

The household member file lists the line number of the mother and
father of each child up to 14 years of age in the variables hv112 and
hv114. You can use this line number to create variables with the
parents' age, level of education, etc. As an example, assume you want
to identify the parents' ages. First, sort the data by household ID
and household member line number.

. sort hhid hvidx

Next, generate variables with the mother's and father's age.

. by hhid: gen mage = hv105[hv112]
. by hhid: gen fage = hv105[hv114]

Two FAQs on the Stata website provide information related to your problem.

"How do I create a variable recording whether any members of a group
(or all members of a group) possess some characteristic?"

"How do I create variables summarizing for each individual properties
of the other members of a group?"


On Sun, Jun 21, 2009 at 6:43 AM, Tharshini
Thangavelu<> wrote:
Hi all,

I have  a specific question concerning the Ghana DHS survery. I am trying
clean the dataset so to analysis using a 2SLS method and of IV method. My
dependent variable will be the anthropometric variables controlling for
economic factors, mothers education, fathers education ect.

Basically my question is : how does paternal (fathers and mothers)
improve the child health in Ghana?

I know how to merge but is still confusing! I am trying as the information
by DHS, that is. I merge first the anthropometric data set with the
member data. Then, I merge the resulting file with the individual data
to finally merge it with the child recode data file.

The first merging (anthropometric and household member data) only
for children is kept,which I think is correct.

Ex. HV105 "the age of household member" gives information about the age of
the household member. When merging the anthropometric data with the
member data, as intuitive says HV105 excludes all age above 5 years.

In order to answer my problem question, I need mothers' fathers' age,
level, their height and weight and other variables. So, What I basically
want is
to locate each child with their parents.

Following the information for merging given by DHS, I should now merge the
resulting file with the women file. But this doesn't give me completely
result that I am expecting. That is, I need womens age, height and weight.
know for sure that these variables exist but when merging, they are

If someone has worked with DHS data set perhaps can give me some advise.
I do one-to-many merging but perhaps I should use many-to-one merging but
doesn't seem reasonable.

Forskarbacken 8 / 101
114 16 Stockholm
Phone +46 (0)735 53 43 90
*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index