The DHS Program User Forum: IPUMS Demographic and Health Surveys (IPUMS-DHS)

Home » Data » IPUMS Demographic and Health Surveys (IPUMS-DHS) » Complex samples

Show: Today's Messages :: Show Polls :: Message Navigator

Complex samples [message #14836]

Tue, 08 May 2018 17:30

Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa

Senior Member

Dear all,I am starting to conduct my MPH Thesis on Individual and Community level factors associated with childhood nutritional status in Ethiopia Using 2016 EDHS Data set.I face many challenges on the following issues:-
1)Is there any difference between IPUMS-DHS Data and the original DHS Data sets?
2)How can I conduct Complex Samples Using SPSS version 24?
3) Which variables in IPUMS-DHS are equivalent to the variables V021(cluster),V022(Strata)and B16 (child line number) in the original DHS data set? If there is no equivalent,How can we merge the IPUMS-DHS to The original DHS Data set.
4) How can I conduct Multilevel analysis Using SPSS Version 24?
Respectfully,Hassen

Hassen Ali(Chief Public Health Professional Specialist)

Report message to a moderator

Re: Complex samples [message #14837 is a reply to message #14836]

Tue, 08 May 2018 17:51

Trevor-DHS
Messages: 803
Registered: January 2013

Senior Member

1) IPUMS DHS data are restructured and reformatted versions of the original DHS datasets. The datasets will look a little different, but the results that you can produce from each of the datasets would be the same. The major advantage of the IPUMS datasets is when working with multiple surveys across multiple countries and years as the variable coding has been harmonized so that the coding of categories are consistent across surveys. IPUMS also permits you to create a single customized dataset of data from multiple surveys, but only of the variables that you select. If you are only working with one dataset (Ethiopia DHS 2016) there is no real advantage of using the IPUMS version.
2) You will find a simple example of using complex samples here: https://userforum.dhsprogram.com/index.php?t=msg&goto=27 67#msg_2767 and a further note here: https://userforum.dhsprogram.com/index.php?t=msg&goto=96 6#msg_966
3) On the IPUMS DHS website, there is a toggle that allows you to switch between IDHS variable names and original DHS variable names. Most of the variables you mentioned can be found under the Topic Technical, Identifiers.
4) Sorry, I've no experience with Multilevel Analysis in SPSS.

[Updated on: Tue, 08 May 2018 17:51]

Report message to a moderator

Re: Complex samples [message #14847 is a reply to message #14837]

Wed, 09 May 2018 02:56

Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa

Senior Member

Dear Trevor-DHS Thank you very much for your Delicious reply!!
If my Unit of analysis is children using IPUMS-DHS Data,Can I copy then paste from KR file of the original DHS Data set when some of my interested variables like maternal BMI,Maternal Anemia etc not exist in IPUMS-DHS Data? Or I must merge them???
Even if I have used a single data set (2016 EDHS Data set),Can I use IPUMS-DHS Data for my analysis? Could the result differs?
Respectfully,Hassen

Hassen Ali(Chief Public Health Professional Specialist)

Report message to a moderator

Re: Complex samples [message #14853 is a reply to message #14847]

Wed, 09 May 2018 09:53

Trevor-DHS
Messages: 803
Registered: January 2013

Senior Member

You don't want to copy-paste from file to file - it is quite likely that they would mis-match. You should merge based on the identifiers variables.
You can use the IPUMS DHS dataset, but given that some of your variables are not in that dataset I don't see the point. You are going to more trouble trying to merge the datasets then it is worth. I would just work with the original DHS dataset. However, if you code your logic properly you will get the same results whether you use the original DHS dataset or the IPUMS DHS dataset. But that is the key to ensure that you are coding your logic correctly.

Report message to a moderator

Re: Complex samples [message #14863 is a reply to message #14853]

Wed, 09 May 2018 13:44

Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa

Senior Member

Dear Trevor,Thank you for assistance and guidance!! It will improve my work as well as my future carer!!
I appreciate your commitment to reply our Questions.
With Best Wishes,Hassen

Hassen Ali(Chief Public Health Professional Specialist)

Report message to a moderator

Re: Complex samples [message #14906 is a reply to message #14863]

Mon, 14 May 2018 11:09

Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa

Senior Member

Dear all,I am Using 2016 Ethiopia IPUMS-DHS Data set(Children as Unit of Analysis). Thus Can I merge IPUMS-DHS data set with the Original 2016 Ethiopia DHS data set(KR file) in order to add some predictor variables like maternal Anemia and maternal BMI etc?
I planned the following steps:
1)open the KR file Data set
2)Sort V001,V002,B16
3)save as Temporary name
4)Open IPUMS-DHS Data set(Children as unit of analysis)
5)sort V001,V002,B16
6)Then,Data-merge files-Add variable. KR file As External data file
7) Move V001,V002,B16 into Key variable box-Ok.
What are your recommendations to do this using SPSS Version 24?
Dear my Hero,I need your urgent reply, due to Our submission date of our final MPH thesis will be soon.
With best wishes,Hassen

Hassen Ali(Chief Public Health Professional Specialist)

[Updated on: Mon, 14 May 2018 11:12]

Report message to a moderator

Re: Complex samples [message #14908 is a reply to message #14906]

Mon, 14 May 2018 11:34

Trevor-DHS
Messages: 803
Registered: January 2013

Senior Member

I don't understand why you want to do this. All of the data that are in the IPUMS data file are already in the KR data file - that is the file from which the IPUMS file is created. maternal anemia is in v457 (see also v453 to v456), and maternal bmi is v445.

Report message to a moderator

Re: Complex samples [message #14909 is a reply to message #14908]

Mon, 14 May 2018 12:08

Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa

Senior Member

Dear Trevor Thank you very much,I want to merge Data from KR file into IPUMS-DHS Data set (Children as unit of analysis). V457 and V445 from KR file into IPUMS-DHS data set.
Respectfully,Hassen

Hassen Ali(Chief Public Health Professional Specialist)

Report message to a moderator

Re: Complex samples [message #14910 is a reply to message #14909]

Mon, 14 May 2018 12:11

Trevor-DHS
Messages: 803
Registered: January 2013

Senior Member

Again, why? All of the data in the IPUMS file is in the KR file already.

Report message to a moderator

Re: Complex samples [message #14911 is a reply to message #14910]

Mon, 14 May 2018 12:27

kingx025
Messages: 95
Registered: August 2016
Location: Minneapolis. Minnesota

Senior Member

Like Trevor, I think there is little advantage to your using IPUMS-DHS for your project, since you are focusing on just one sample and some of the variables you need are not in IPUMS-DHS yet. You may use the online documentation for variable of interest from the IPUMS-DHS website, or use the IPUMS-DHS website to locate additional variables to include in your analysis, but still use the KR data file from The DHS Program to run your analyses.

If you can afford to wait 2-4 weeks to run your analysis, you could get the mother's anthropometric data from IPUMS-DHS, since we have harmonized those variables but have not made them public yet--but will do so soon.

If you change your research plan to working with all the Ethiopian samples, then there would be more of a competitive advantage to analyzing data from IPUMS-DHS.

There is a user note on linking between original DHS files and an IPUMS-DHS data extract, if you are determined to follow that path.
https://www.idhsdata.org/idhs/linking.shtml

Miriam King

Dr. Miriam King
IPUMS-DHS Project Manager (www.idhsdata.org)

Report message to a moderator

Previous Topic:	Household members as unit of analysis
Next Topic:	IPUMS-DHS Tip #1. Applying weights to IPUMS-DHS data files in Stata

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Nov 23 18:12:03 Coordinated Universal Time 2024