The DHS Program User Forum: Sampling

Home » Data » Sampling » Design effect

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

Design effect [message #3427]

Wed, 10 December 2014 06:05

awah
Messages: 2
Registered: December 2014
Location: Sweden

Member

Hi,

I am a PhD student working in the field of female genital cutting (FGC).
We are planning a cluster randomized intervention study to investigate knowledge, attitudes and practice of FGC.

For the sample size calculation we need to estimate the design effect (or the intracluster correlation coefficient).
I have found that DHS publish design effect for some of the questions used in the questionnaire. But I have not found this reported for the questions in the Female Genital Cutting Model.
Is this information available somewhere?

Much grateful for an answer!

Report message to a moderator

Re: Design effect [message #3429 is a reply to message #3427]

Wed, 10 December 2014 10:20

Sarah-DHS
Messages: 54
Registered: February 2013

Senior Member

Dear awah,

Here is a response from Senior Sampling Statistician Dr. Rulin Ren:
It is true that we never reported the CI for the FGC in the DHS final report, so there is no estimate of the design effect. But I guess it is an indicator with high intra-cluster correlation because it is culture related. Our rule of thumb is that if you do not know the design effect of specific DHS indicator, take it as 1.5 because this is the average value for DHS indicators. Please note that we use Deft, which is the square root of Deff.

Thank you,
Sarah-DHS

Report message to a moderator