| How to deal with non response data for M55Z$1 First 3 days, given nothing [message #2375] |
Tue, 10 June 2014 10:55  |
aligazan
Messages: 24
Registered: June 2014
Location: UK
|
Member |
|
|
Is there any way of dealing with missing data for this variable (or any other)?
I am using the Mali Phase V 2006 datafile, and am particularly interested in M55Z$1 'first 3 days, given nothing' to do with initial breastfeeding behaviour. There is a high non response rate from mothers whose babies died during the first month after birth, and this group is of great interest to me. The way the responses are structured there is only a yes or no option, and maybe this wasn't adequate to capture the range of responses which might have included 'can't remember', or 'not applicable' if the baby died before this was possible.
A statistics teacher suggested that there might be some DHS 'non response weights' which could be used to deal with the missing data? He also suggested that if I were using STATA (which I'm not at the moment!) there might be an easier way of dealing with this than when using SPSS.
Any advice would be very much appreciated.
|
|
|
|
| Re: How to deal with non response data for M55Z$1 First 3 days, given nothing [message #2377 is a reply to message #2375] |
Tue, 10 June 2014 15:40  |
Bridgette-DHS
Messages: 3214
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior Data Processing Specialist, Noureddine Abderrahim:
In some statistical software, missing values and not applicable values are considered to be the same and noted as missing. The cases that don't apply for the case of M55Z are those children who were not breastfed. By excluding these children, you will reduce the number of missing cases. For the remaining cases of missing, there is no way to go back and get that data, since the information was lost during data collection/fieldwork.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
