The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Merging data files » DHS Pakistan 2017-2018 (Merging File IR and BR)
DHS Pakistan 2017-2018 [message #26597] Thu, 06 April 2023 08:29 Go to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Expert,

I am working on the study about women education on fertility 2017-2018 Pakistan data. I need to analyze the variable, total number of children ever born (BR file) with women education attainment and other characteristics of women in the IR file. I want to merge IR file which contain ever married women with the BR file because I need total children ever born to analysis with women characteristics.

OR need to merge these variables v149 (IR file), v024 (IR file), v013 (IR file) with the v201 (BR file).

Kind regards.
Re: DHS Pakistan 2017-2018 [message #26599 is a reply to message #26597] Thu, 06 April 2023 11:05 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

In the IR file, the cases are women. The number of children ever born is a characteristic of the woman. It is coded as v201 in the IR file. All the variables you mention are in the IR file.

In the BR file, the cases are children, but the BR file includes virtually all the information about the mother on the child's record. You do not need to merge the IR and BR files. All the variables you mention are already in the BR file.
Re: DHS Pakistan 2017-2018 [message #26601 is a reply to message #26597] Thu, 06 April 2023 23:29 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member

Thanks for the response.

In the PDF file of the PDHS survey 2017-2018, A total of 15,671 households were selected for the survey, of which 15,051 were occupied.
But in the data file in IR the number of observations is 15,068. Kindly confirm the exact sample size.

Regards.
Re: DHS Pakistan 2017-2018 [message #26602 is a reply to message #26601] Fri, 07 April 2023 07:25 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member
We do not distribute data on households or Individuals with result of interview different of 1 (complete).

You should be looking at HV015 in the household recode dataset. The number of households interviewed (completed interviews) is: 14540. This matches the number in the Final Report: Table 1.1 Results of the household and individual interview.
Re: DHS Pakistan 2017-2018 [message #26612 is a reply to message #26599] Sun, 09 April 2023 06:31 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member

Thanks for the reply.

I am new user of stata. it is very difficult to analyze my data.

My dependent variable is children ever born and independent variable are women education attainment wealth index and other characteristics of women. I want to apply Poisson regression on my data. I am not familiar to survey data analysis in stata. Are these syntex are better for my analysis?

gen swt= v005/1000000
la var swt "Sampling weight"
svyset v021, strata(v023) weight(swt) vce(linearized) singleunit(missing)
svy linearized : poisson CEBR i.WEA i.AgeG i.Region i.WI i.PR

Or we can use this syntex

poisson CEBR i.WEA i.AgeG i.Region i.WI i.PR [pweight = swt]

your suggestion is required.

Kind regards.
Re: DHS Pakistan 2017-2018 [message #26617 is a reply to message #26602] Mon, 10 April 2023 07:00 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Experts,

I am waiting for your response.

If I am using the below mentioned syntax for IR file. They give me a note in the results:

Note: 4 strata omitted because they contain no population members.

Is this any problem? and also, in the results they don't show in the region results of Gilgit Baltistan.

or we can use simple this syntax for Poisson.

poisson CEBR i.WEA i.AgeG i.Region i.WI i.PR

Kindly help me in this matter.
Regards.


Re: DHS Pakistan 2017-2018 [message #26619 is a reply to message #26617] Mon, 10 April 2023 19:29 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

I recommend using svyset but with the singleunit option "centered", as follows:

svyset v021, strata(v023) weight(swt) vce(linearized) singleunit(centered)
svy linearized : poisson CEBR i.WEA i.AgeG i.Region i.WI i.PR

I don't know how all of your variables are defined, but in terms of adjustments for the survey design, it is better to use all the components of svyset, not just the weights.

You also asked about regions being omitted. If you run "tab v024, summarize(v005)" you will see that gb and ajk have weights of 0, meaning they will be dropped from your analysis. You can read about this in the final report on the survey and in several other postings on the forum. The government of Pakistan requires that these two areas not be included in national estimates. Please follow that requirement. These areas can be analyzed separately but not pooled with the rest of Pakistan.
Re: DHS Pakistan 2017-2018 [message #26648 is a reply to message #26619] Wed, 12 April 2023 23:32 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Tom Pullum,

Thanks for the detailed answer.

How we can analyze the GB and AJK separately. My all variables are available in IR file.sv005 in IR file, it is mentioned that GB and AJK data. But when I used it in analysis, they also show that there is no population member.

Regards.
Re: DHS Pakistan 2017-2018 [message #26654 is a reply to message #26648] Thu, 13 April 2023 08:04 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

Please search the forum for other references to the Pakistan 2017-18 survey and specifically the analysis of GB and AJK. DHS has nothing to add to what has already been posted. The Government of Pakistan, together with the National Institute of Population Studies (NIPS), are responsible for the special treatment of these two areas.

Re: DHS Pakistan 2017-2018 [message #26812 is a reply to message #26654] Wed, 03 May 2023 10:50 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Tom Pullum,

I am working in IR file. As you know in my previous messages my independent variable is v149 and my dependent variable is v201. I need a mother's and father's education attainment variable. Your support needed.

Kind Regards.
Re: DHS Pakistan 2017-2018 [message #26813 is a reply to message #26812] Wed, 03 May 2023 11:57 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

In the woman's interview, she is asked about the education of her husband/partner, and her responses go into the IR file. V715 and v729 for her partner correspond with v133 and v149 for the woman. The woman may be widowed, remarried, etc., so you don't know whether the current husband was the father of all her children (v201) but usually that will be the case. The woman's report of her husband's level of education may not agree exactly with what he would say himself. However, I think it will be ok to use v715 and v729 in your analysis.
Re: DHS Pakistan 2017-2018 [message #26815 is a reply to message #26812] Thu, 04 May 2023 05:22 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Tom Pullum,

In my analysis, v729 is also included. but there is endogeneity problem between dependent variable(v201) and independent variable (v149 and v729), that's why I need some instrument variable to solve this issue. Mother education may be a good instrument variable. it is not necessary in v149 all women are mother. Your opinion is highly appreciated.
Regards.
Re: DHS Pakistan 2017-2018 [message #26818 is a reply to message #26815] Thu, 04 May 2023 07:32 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

My only opinion would be that you should try to confirm the validity of the assumptions behind the use of any instrumental variable. Note also that there is usually a strong association between the education of the father and the mother because of homogamy.


Re: DHS Pakistan 2017-2018 [message #26819 is a reply to message #26815] Thu, 04 May 2023 08:32 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Tom Pullum,

I appreciate your opinion. I have confirmed that the mother and father education is good instrument variables for my analysis, that's why I am asking about these variables. v149 and v729 is women education and husband education data but it is not necessary that all women and men are mother and father. In v201, there are 1,950 women has no children. I need mother and father education data. Kindly help me in this matter.

Regards.
Re: DHS Pakistan 2017-2018 [message #26821 is a reply to message #26819] Thu, 04 May 2023 11:15 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

Sorry--I don't know what you are looking for. In the IR file you have the education variables for the woman and her reports of the education of her current husband. If v503 (number of unions) is 1, which it is for 98% of women in this survey, then the husband is the father of all the children. If you go to the CR file, you will find the woman's husband and get his reports of his own education, which could be more accurate but only marginally. That's all the advice I can give you.
Re: DHS Pakistan 2017-2018 [message #26829 is a reply to message #26821] Fri, 05 May 2023 12:13 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Tom Pullum,

Thanks for your reply, but I am asking about mother education data not women education data. My point is that in IR file there is v201, there are 1,950 women has 0 children, 1,983 women has 1 child, 2,305 women has 2 children, 2,376 women has 3 children and it's up to 15 children. If am I right. If women have 0 children, they are not mother. I need a data that women are mother. In IR file, V149 included of women education data and in this woman also included that they are not mother. I am agreed your point, v503 (number of unions) is 1, which it is for 98% of women in this survey. The 98% women are married but it's not necessary they have child because in IR file they also show the women has 0 child. I need mother education data.

second, I also need the variable of Marital status of women that they mentioned, not married and married. no need of the variable of woman may be widowed, remarried, etc.
Re: DHS Pakistan 2017-2018 [message #26836 is a reply to message #26829] Mon, 08 May 2023 05:51 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member

Dear Tom Pullum,

I am waiting your response.
Regards.
Re: DHS Pakistan 2017-2018 [message #26837 is a reply to message #26836] Mon, 08 May 2023 08:43 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

I'd say that your units of analysis are women. If a woman has had no children, then it is true that she is not a mother, but 0 is a valid number of children. If you wanted to calculate the mean number of children ever born, you would include the women for whom v201=0. If you just want to analyze women who have had at least one child, then DROP the women with v201=0.
Re: DHS Pakistan 2017-2018 [message #26840 is a reply to message #26837] Mon, 08 May 2023 12:46 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Tom Pullum,

Thanks for reply but I am not interested to calculate the mean number of children ever born that they are mother. I am interested to calculate the mother education data because in IR file, V149 included of women education data and in this woman also included that they are not mother. Is there any variable available that we can calculate the mother education data. any possibility? secondly, I am also asking about the variable of Marital status of women that they mentioned, not married and married. no need of the variable of woman may be widowed, remarried, etc.

Regards.
Re: DHS Pakistan 2017-2018 [message #26842 is a reply to message #26840] Tue, 09 May 2023 08:03 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Tom Pullum,

I am waiting for your response. In IR file there is variable v501, married (women: v501 = 1), Widowed (women: v501 = 3), Divorced (women: v501 = 4), Separated (women: v501 = 5) data are available but never married (women: v501 = 0) data are not available. I need also this data of v501=0. Kindly help me in this variable as well as my previous message that the variable of women education.

Regards.
Re: DHS Pakistan 2017-2018 [message #26843 is a reply to message #26842] Tue, 09 May 2023 08:05 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3199
Registered: February 2013
Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

DHS data are used all the time to analyze children and their mothers, sometimes with the children as the cases and sometimes with the mothers as the cases. In the IR file, the women are the cases. In the KR file, the children are the cases. The v variables always refer to the women, whether you are using the IR file or the KR file. If v201=0 then she has had no children. If v201>0 then she has had children. Other v2* variables tell you whether she has living children, has children living with her, etc. Whatever definition of "mother" you may be using, you can apply it. I can't give more help. Maybe some other users can step in.
Re: DHS Pakistan 2017-2018 [message #26844 is a reply to message #26843] Tue, 09 May 2023 08:15 Go to previous messageGo to next message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member
Dear Tom Pullum,

thanks for your response.

In IR file there is variable v501, married (women: v501 = 1), Widowed (women: v501 = 3), Divorced (women: v501 = 4), Separated (women: v501 = 5) data are available but never married (women: v501 = 0) data are not available. I need also this data of v501=0.
Regards.
Re: DHS Pakistan 2017-2018 [message #26865 is a reply to message #26844] Thu, 11 May 2023 03:58 Go to previous message
Khawar is currently offline  Khawar
Messages: 19
Registered: April 2023
Member

Dear Tom Pullum,

I am waiting for your response.

Kind regards.
Previous Topic: How to merge KR and IR
Next Topic: Disability in HR with characteristics from IR and MR
Goto Forum:
  


Current Time: Tue Nov 26 06:45:19 Coordinated Universal Time 2024