Liberia - Not possible to replicate education indicators [message #12355] |
Thu, 04 May 2017 14:10 ![Go to next message Go to next message](/theme/default/images/down.png) |
rosavid
Messages: 1 Registered: May 2017 Location: Paris
|
Member |
|
|
Hi,
I am working with the 2013 DHS for Liberia for an upcoming reporte. I am trying to replicate the Table 2.13 "School attendance ratios", specifically the "Total Net Attendance Ratio" indicator. I have similar problems to reproduce the results reported for net attendance ration for 2007 in StatCompiler.
Here it is what I did to calculate net attendance ratio in the 2013 Liberia database:
clonevar age=hv105
recode hv121 (2=1) (9=.)
tab hv121 [iw=hv005] if age>=6 & age<=11 // For PRIMARY
tab hv121 [iw=hv005] if age>=12 & age<=17 // FOR SECONDARY
*I used the variable hv121 because hv129 is missing in all cases
And the results I got are are:
- For Total primary net attendance ratio 78.66 (instead of the 38.3 published)
- For Total secondary net attendance ratio 83.49 (instead of the 22.8 published)
Could you please help me figure out what is wrong with my code?
Thanks in advance for your help.
Rosa Vidarte
e XX published)
|
|
|
Re: Liberia - Not possible to replicate education indicators [message #12415 is a reply to message #12355] |
Fri, 12 May 2017 08:07 ![Go to previous message Go to previous message](/theme/default/images/up.png) |
Bridgette-DHS
Messages: 3152 Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Senior Research Associate, Lindsay Mallick:
There is one error in your code that immediately jumps out. Typically, you should use the variable hv122 as this will ensure that you are calculating NAR for primary school attendance versus secondary for the appropriate age groups as the table specifies.
For example, for primary school among 6-11 year olds, the code would look like this:
gen wgt=hv005/1000000
tab hv122 if inrange(hv105,6,11) [iw=wgt]
There was also an additional calculation done for this survey for this table. Because the survey ran across two school calendar years, adjustments were made using the CMC for age of the child based on the beginning of the country's school year. Further, data were imputed for children with missing CMC of birth. Thus, the code above would not yield a match with the final report.
Please make sure to also view our newly released video on matching tables in the final reports:
https://www.youtube.com/watch?v=tjuaPV4eC_E&feature=yout u.be&list=PLagqLv-gqpTMU3avlnBDodTWCazURy4CT
[Updated on: Fri, 12 May 2017 09:49] Report message to a moderator
|
|
|