| Use of Pooled Data set for PHD [message #5562] |
Mon, 08 June 2015 05:48  |
Robert
Messages: 19
Registered: April 2015
Location: Dar Es Salaam, Tanzania
|
Member |
|
|
Dear all
I am analyzing the effect of fertility on child health and maternal health status in Tanzania using the data set for the years 1999, 2004 and 2010. I have decided to pool the three round up surveys but my supervisor seems not to agree with my approach cause she believes that since these are not panel data or time series then it would be wrong for me to pool data set and create year dummies. However literature allows people to pool cross section DHS.
I am I doing the right thing?
Robert
R.Lihawa
|
|
|
|
| Re: Use of Pooled Data set for PHD [message #5566 is a reply to message #5562] |
Mon, 08 June 2015 14:53  |
Reduced-For(u)m
Messages: 292
Registered: March 2013
|
Senior Member |
|
|
yes - you can certainly pool data from multiple rounds, you just have to adjust the weights (see "denormalizing weights") and the strata (by making strata survey-round specific, also discussed much on this forum in the weighting section). The more general point is that, even though this does not create a true, individual level panel, most panel methods are also applicable (with some adjustment) to this "repeated cross section" approach. So:
a) your adviser is right that this is not a panel;
b) you can do a lot by using multiple DHS rounds using repeated cross section methods
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
