How to merge KR and IR [message #26774] |
Sun, 30 April 2023 04:25 ![Go to next message Go to next message](/theme/default/images/down.png) |
masternaveen123
Messages: 1 Registered: April 2023
|
Member |
|
|
Hi,
Objective: To join children's records(IAKR7DFL.DTA) and individual records(IAIR74FL.DTA) datasets.
What I have tried so far:
The following is for NFHS_5:
For children's records, we have clusters from 101-93,000 and the individual record is 10,000-360,000.
So, only common data points from both sides will exist between clusters 10,000-93,000. So one of the questions I have is why are there missing cluster values on both sides of the dataset.
I looked up the range of cluster values for NFHS 4 dataset, for both childrens records and individual records. It turns out that both children and individual record are in the range 10,001-360,482. So the anomaly exists between the ranges of the clusters only in NFHS 5.
On the DHS Merging website, if you see the section where it says matching variables, we can see what are the unique identifiers required to join the tables.
Following the same logic, I joined the children's record and household record, matching cluster number and household number. This gave me valid child records with their household details.
Now when I follow the same DHS guide to join children and womens dataset, matching cluster number and household number, I only get around 2000 valid records, and that too I don't think they are merging correctly and it isn't the intended data.
|
|
|
Re: How to merge KR and IR [message #26811 is a reply to message #26774] |
Wed, 03 May 2023 07:57 ![Go to previous message Go to previous message](/theme/default/images/up.png) |
Bridgette-DHS
Messages: 3151 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS staff member, Tom Pullum:
India files with "74" as the 5th and 6th letters of the file name are NFHS-4 Files with "7D" are NFHS-5. These are completely different surveys, with different clusters, households, and individual. You cannot merge the KR file from one survey with the IR file from a different survey.
Moreover, within a survey, the records in the KR file (children) carry along almost all the variables from the IR file (the mothers). You almost never need to merge the KR and IR files from the SAME survey.
|
|
|