| Indonesia DHS 2012 published factor weight file mislabelling [message #10063] |
Tue, 21 June 2016 12:45  |
KenzoFry
Messages: 21
Registered: December 2015
Location: UK
|
Member |
|
|
Hi,
while working through Indonesia DHS 2012's wealth index to help a colleague that had used it, I noticed that two variables relating to toilet type - 'Flush toilet to sewer' (mean = 0.5936) and 'Flush toilet to septic tank' (mean = 0.0627) - may have been mislabelled.
The variable HV205 (toilet type) does not contain these options. It contains the options 'Private - with Septic Tank' and 'Private - without Septic Tank'. These also match the questionnaire as in the report.
The unweighted proportion of people with 'Private with septic tank' is 59.4% and for Private without septic tank it is 6.3% - matching the means for the two mislabelled variables. In other words, if this is correct, I think the file should be corrected as follows:
'Flush toilet to sewer' changed to 'Private with septic tank'
'Flush toilet to septic tank' change to 'Private without septic tank'
Another issue is with the variable "Bedroom available for sleep". Scores are given for this variable for 'if has' and 'if doesn't have', as if this is a binary variable. However, since its mean is 2.2 it seems this is actually a continuous variable more accurately described as 'number of bedrooms available for sleep'.
Thanks,
Kenzo
EquityTool.org
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
