Home » Data » Merging data files » Merging Individual and HH Member Recode Files
| Re: Merging Individual and HH Member Recode Files [message #10515 is a reply to message #10499] |
Mon, 01 August 2016 12:04   |
Bridgette-DHS
Messages: 3203
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Specialist, Shea Rutstein:
Since 2009, the sampling design of the Peru surveys consists of two nationally representative halves ("semesters"). For one semester, the same clusters are used from one of the previous year's semester, either the first or last, and a new set of clusters are selected for the other semester, which will be reused in the following year. A set of clusters is only used twice. Within both the reused and newly selected clusters, a complete household listing is done for each year and a new selection of households is done with a new numbering. Thus there is no guarantee that the same households are selected in neighboring years, and most likely they are not reselected. Therefore, only half of the clusters match but none of the households nor members.
For a single year, the household and individual data files can be merged by using the cluster and the household numbers. The semester is given in SHSEMES and SSEMES but I don't think it is necessary to use them for merging. I am not sure if I would follow a panel of clusters from one survey to the next. To do so, one would have to id each semester as to which clusters were in the previous year and also denormalize the weights. The reuse of half the clusters was done to reduce the variance for trends.
Be careful about the sampling weights when pooling the years. The weights have been normalized to the sum of total households or individual interviews for each year.
|
|
|
|
Current Time: Wed Dec 4 19:50:37 Coordinated Universal Time 2024
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
