Home » Data » Weighting data » Weighting Data for Pooled Indonesian DHS Dataset
| Re: Weighting Data for Pooled Indonesian DHS Dataset [message #10054 is a reply to message #10051] |
Mon, 20 June 2016 10:12   |
Bridgette-DHS
Messages: 3043
Registered: February 2013
|
Senior Member |
|
|
Following is another response from Senior DHS Stata Specialist, Tom Pullum:
As I understand it, you are putting the 2007 and 2012 surveys of Indonesia into a single data file in order to analyze changes and differences. This is convenient for data processing and I would do the same thing.
The question you are asking is only relevant if you want to calculate a mean for the two surveys, and I would definitely not recommend doing that. For example, I would be interested in the CPR in each survey and the change in CPR between the two surveys, but I hope you are not planning to calculate the mean CPR of the two surveys combined. Are you? If so, why? How could you interpret such a number? (I mention the CPR just as an example.)
If you want to do what I think you want to do, that is, to estimate a change from 2007 to 2012, and calculate a confidence interval for the difference, or test whether the change is statistically significant, then you will get exactly the same result if you leave the weights alone OR if you re-scale them so the total weight is the same in each survey OR if you re-scale them so that the total weight in each survey is proportional to the population size. I encourage you to compare these options. Let me know if they are NOT all the same, in terms of point estimates, standard errors, confidence intervals, and test statistics.
|
|
|
|
Current Time: Sun Apr 28 23:52:27 Coordinated Universal Time 2024
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
