The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Reproductive Health » Binary Variables Expanded to Six Categories
Binary Variables Expanded to Six Categories [message #1598] Sun, 16 March 2014 16:55 Go to next message
amira.elshal.1@city.ac.uk is currently offline  amira.elshal.1@city.ac.uk
Messages: 52
Registered: April 2013
Senior Member
Dear Sirs,

I am working on the Individual Recode files of Egypt's EDHS 2000 and 2008 using Stata.

I understand that a variable such as who attended at the delivery (m3) is coded by DHS as a set of binary variables (m3a, m3b, etc.)
for each possible option (doctor, nurse/midwife, etc.). What I do NOT understand is why m3a is expanded into m3a_1, m3a_2,...m3a_6. Is that to deal with the case of a woman giving birth to 6 children for example?

Thank you.

Amira
Re: Binary Variables Expanded to Six Categories [message #1607 is a reply to message #1598] Mon, 17 March 2014 14:17 Go to previous messageGo to next message
Reduced-For(u)m
Messages: 292
Registered: March 2013
Senior Member

Amira,

The _1, _2.... are in fact for different children - all children born within the last 3/5 years. From the recode manual (https://www.dhsprogram.com/pubs/pdf/DHSG4/Recode4DHS.pdf)


The maternity history contains up to six entries, relating to births in the three/five years preceding interview. The use of three/five years is country specific. The entries are in reverse order, such that the first entry relates to the last birth in the last three/five years. There is an entry for all children born in the last three/five years including all twins. The period of three/five years includes months 0 to 35/59 prior to the interview, with month 0 being the month of interview. If there are more than six births in the last three/five years then only the last six are included in the maternity history. Each of the following variables, duration of breastfeeding (M5), duration of postpartum amenorrhea (M7) and duration of postpartum abstinence (M9) may have several cases coded 97 "Inconsistent" since the duration of breastfeeding, amenorrhea or abstinence was impossible in the interval between the birth and the following birth or date of interview if the most recent birth (only the date of interview in the case of breastfeeding).
Re: Binary Variables Expanded to Six Categories [message #1623 is a reply to message #1607] Wed, 19 March 2014 09:34 Go to previous messageGo to next message
amira.elshal.1@city.ac.uk is currently offline  amira.elshal.1@city.ac.uk
Messages: 52
Registered: April 2013
Senior Member
Thank you for your informative reply.

The question now is how I can calculate a maternal indicator if a woman has had more than 1 birth during the last five years.

For example, according to the Guide to DHS Statistics, the ANCC (at least four visits) is 100 times the quotient of the numbers of women who received antenatal care for their last birth, according to grouped number of visits (four visits), divided by the number of women with a birth in the last five years, expressed as a percentage. How can I update this formula to account for women having more than 1 birth during the last five years?

Thank you again.

Amira
Re: Binary Variables Expanded to Six Categories [message #1629 is a reply to message #1623] Wed, 19 March 2014 12:57 Go to previous messageGo to next message
user-rhs is currently offline  user-rhs
Messages: 132
Registered: December 2013
Senior Member
Amira,
I would strongly advise you to read the DHS Recode Manual (link: https://www.dhsprogram.com/pubs/pdf/DHSG4/Recode6_DHS_22Marc h2013_DHSG4.pdf) and the questionnaires at the end of the DHS Final Report for Egypt to familiarize yourself with the data. This is what I personally do whenever I'm stumped about what variable to use or the correct way to use a variable I'm interested in. There is a lot of good information in there you can use: what questions were asked, how were they stored in the dataset, what the _1, _2, _3 in a variable name mean, why some variables have missing values for all observations, etc.

The answer to your question is answered in the DHS Recode Manual.

RHS


NB. FYI, typically the antenatal and postnatal care variables are asked only of the most recent birth, instead of ALL of the births in the last 5 years presumably to improve the reliability of the response and minimize recall bias.


Re: Binary Variables Expanded to Six Categories [message #2443 is a reply to message #1629] Wed, 18 June 2014 12:08 Go to previous messageGo to next message
amira.elshal.1@city.ac.uk is currently offline  amira.elshal.1@city.ac.uk
Messages: 52
Registered: April 2013
Senior Member
Dear RHS,

I have taken your advice and checked both the DHS Recode Manual and Egypt's EMW questionnaire. I have also checked the individual data file I have. However, in Egypt, I have noticed that antenatal and postnatal care variables are asked for all births, not only of the most recent birth. So, if 1 woman gave birth to 2 children in the last five years, I know how many ANC visits she had for each of her 2 children. Does I have to account for the ANC visits of her last birth only when I calculate the variable "ANC - at least 4 visits"? Or her ANC visits for both children?

Thanks.

Amira
Re: Binary Variables Expanded to Six Categories [message #2446 is a reply to message #2443] Wed, 18 June 2014 13:22 Go to previous message
user-rhs is currently offline  user-rhs
Messages: 132
Registered: December 2013
Senior Member
It depends on your research question. If you are interested in # ANC visits per pregnancy for all pregnancies, then you would use all of the data from each woman. If you are interested in the most recent pregnancy, then you would use just the variables with subscript _1 in the name. Note that if you use all of a woman's pregnancies, then you would have to account for "nesting" within the individual woman (as her care-seeking behaviour during pregnancy may be the same for all pregnancies).
Previous Topic: DHS Malawi youth-specific
Next Topic: Computing timing of postnatal checkup using Stata
Goto Forum:
  


Current Time: Thu Mar 28 18:31:01 Coordinated Universal Time 2024