Issue with duplicate mothers [message #25205] |
Thu, 15 September 2022 07:50 ![Go to next message Go to next message](/theme/default/images/down.png) |
erik.cl
Messages: 1 Registered: September 2022
|
Member |
|
|
Hello,
I am a medical student doing a project analysing maternal factors associated with ORS-administration to children with diarrhoea in Pakistan (using Children recode DHS data from 2017-18). Unfortunately I have run in to an issue I can't seem to resolve on my own.
Problem Using children's recode as base dataset, some mothers might have several children, resulting in issues where one mothers factors associated with ORS-use might be amplified in the analysis if they have many children who are experiencing diarrhoea simultaneously. I can't figure out a way to produce descriptive statistics showing how many mothers that had more than one child with diarrhoea simultaneously, and thus I am unable to asses how big of an issue this might be. Furthermore I am wondering if there is some SPSS syntax I could use to select one child < age 5 per mother at random so I can work around this issue?
I have already read the thread https://userforum.dhsprogram.com/index.php?t=tree&goto=1 6448&&srch=selecting+one+child+per+mother#page_top but the syntax does not seem to work for SPSS.
Thanks!
|
|
|
Re: Issue with duplicate mothers [message #25207 is a reply to message #25205] |
Thu, 15 September 2022 11:14 ![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 DHS staff member, Tom Pullum:
I strongly recommend against subsampling of children, for one reason: your results will be specific to your subsample and will not be replicable by other researchers. Many of the items collected for children (for example, ANC, place of delivery, and breastfeeding) are limited to the most recent birth in the past 5 years. In the KR or BR files, this is the child with bidx=1. If you want to reduce to just one child, I recommend that you select that child. This may introduce some bias (see https://www.dhsprogram.com/pubs/pdf/MR14/MR14.pdf) but it's a bias that is pretty familiar in DHS analyses.
|
|
|