| Re: Child development module in Honduras 2011-2012 [message #14449 is a reply to message #14304] |
Fri, 06 April 2018 13:46  |
Bridgette-DHS
Messages: 3138
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
Sorry for the delay with this response. Questions like this should identify the specific table in which the problematic number (in this case 2710) appears. I scanned the report and found the number in table 17.1.
The relevant variables appear to be s562* through s566* in the KR file (HNKR61FL.dta). If a variable is coded with a blank or dot then the variable is not applicable, i.e. is NA. For most of the s56* variables, there are 3614 unweighted cases that are not NA. If I restrict to hw1>=36 & hw1<=59, then 2923 cases are not NA. (The questions were asked about a large number of children, about 700, whose age turned out to be below 36 months or above 59 month.) Of the 2923, 5 cases consistently have code 9, which is not a legal code. If you just removed those cases you would get 2923-5=2918. I am guessing that that is how you got 2918. However, this is an unweighted frequency and the n's in table 17.1 are weighted. If I do
tab s564a if hw1>=36 & hw1<=59 [iweight=v005/1000000]
then the weighted number of cases that are not NA on the s56* variables and are in the range 36-48 months is 2778.2934, or 2778. The way I would read the table, that's the number that should be 2710. My best guess, 2778, is too high by 68 cases.l
This was a special module and table 17.1 is a special table, i.e. non-standard. If I were you I would simply take the cases that are not NA on s564a and are in the age range 36-59 months as the eligible cases and proceed to analyze them. I cannot match the table.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
