The DHS Program User Forum: Sampling

Home » Data » Sampling » Statistical Power (Statistical Power SADHS 2016)

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

Statistical Power [message #27773]

Mon, 02 October 2023 07:55

Miyambu Langutani
Messages: 5
Registered: September 2023

Member

What is the statistical power used in South African Demographic and Health Survey 2016

Report message to a moderator

Re: Statistical Power [message #27777 is a reply to message #27773]

Mon, 02 October 2023 09:52

Bridgette-DHS
Messages: 3199
Registered: February 2013

Senior Member

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

The report on this survey, with its appendices, includes some discussion of the sampling design. The power of statistical tests varies across indicators and hypotheses and should take into account the clustering and stratification in the sampling design. Hope this helps.

Report message to a moderator

Re: Statistical Power [message #27778 is a reply to message #27777]

Mon, 02 October 2023 10:44

Miyambu Langutani
Messages: 5
Registered: September 2023

Member

Thank you for the response. However, don't you have a Standard statistical power which you use since this is a national representative survey? For instance, a statistical power of 90% or 80% ?

Report message to a moderator

Re: Statistical Power [message #27783 is a reply to message #27778]

Tue, 03 October 2023 10:14

Bridgette-DHS
Messages: 3199
Registered: February 2013

Senior Member

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

We have little to add to the previous response. Statistical power is a characteristic of a test. It depends on the indicator and the test. DHS data are used more for estimation than for testing. The width of confidence intervals and the coefficient of variation are more relevant criteria, but they also vary from one indicator to another. The appendix on the sample design in the final reports includes a great deal of information about the standard errors and design effects of the main indicators.

Report message to a moderator

Re: Statistical Power [message #27784 is a reply to message #27783]

Tue, 03 October 2023 12:15

Miyambu Langutani
Messages: 5
Registered: September 2023

Member

Thank You for the response.

Is it possible to provide statistics of women with maternal multimorbidity (the coexistence of two or more chronic conditions e.g Obesity, hypertension, HIV/AIDS and etc.) and women without multimorbidity in SADHS 2016?

Thank You

Report message to a moderator

Re: Statistical Power [message #27785 is a reply to message #27784]

Tue, 03 October 2023 13:01

Bridgette-DHS
Messages: 3199
Registered: February 2013

Senior Member

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

That would be possible with the data files, which are publicly available. You can register here for access to the files, if you have not already done so:

https://www.dhsprogram.com/data/new-user-registration.cfm

You would have to do the analysis yourself.

Report message to a moderator

Re: Statistical Power [message #27786 is a reply to message #27785]

Tue, 03 October 2023 13:18

Miyambu Langutani
Messages: 5
Registered: September 2023

Member

I have registered for DHS program. I have the file. I just wanted to find out if maybe you've a variable which directly provide the women with maternal multimobirdity and without multimobirdity. Which best file can one access that information

Thank you

Report message to a moderator

Re: Statistical Power [message #27787 is a reply to message #27786]

Tue, 03 October 2023 16:10

Bridgette-DHS
Messages: 3199
Registered: February 2013

Senior Member

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

There is not a variable that describes co-morbidities, but you can construct one. Please look at the report and/or questionnaire and find which illnesses/morbidities were included in this survey. I can then help you locate them in the data files and construct variables to describe combinations.

Report message to a moderator

Re: Statistical Power [message #27788 is a reply to message #27787]

Tue, 03 October 2023 16:23

Miyambu Langutani
Messages: 5
Registered: September 2023

Member

Thank you very much. I am interested in High blood pressure, HIV, Obesity, diabetes and other illnesses since the definition of multimobirdity include the coexist of two or more health conditions. If you can be able to construct that for me so that I can get women with multimobirdity and women without multimobirdity together with their frequency and percentages, I can be able to calculate the statistical power.

Report message to a moderator

Previous Topic:	Madagascar 2021: representative at district level?
Next Topic:	Two or three stages?

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Thu Nov 21 11:22:32 Coordinated Universal Time 2024