The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » General » Malawi - school attendance rates - not matching report
Re: Malawi - school attendance rates - not matching report [message #15887 is a reply to message #15858] Thu, 04 October 2018 11:37 Go to previous messageGo to previous message
geoK is currently offline  geoK
Messages: 33
Registered: May 2014
Thank you for your kind reply. I have been using the guide to replicate the NAR indicator for secondary school (school attendance ratios page). I still encounter a few issues, as well as my final percentage for Secondary school NAR in Malawi is still around 15% (instead of 17.4%).

I would like to ask a few questions, please, related to the step by step process described in the guide.

1) when it comes to merging the PR file and the BR file quite a few line numbers (b16) are missing - although variable B3 is actually not missing. Therefore that person is either not merged into the new file, or it is but it has all missing values for HV122 and other PR variables. What should i do with them? I can see that some of them are dead (B6).
I suppose, when doing the merging, I should prioritize retaining information from the PR file (hv122), and impute b3 in case it is missing, correct?

2) At page 3 of the School attendance ratios section of the guide (see below), you describe 2 ways to calculate the age at the start of the school year, 1 is by the merging i was talking about at point 1) above, and 1 by imputation. can you please advice whether you usually use one among the 2, or both, and what has been used in Malawi 2015-16 please?

"To calculate the age at the start of the school year a century month code of date of birth (cmcdob) is needed for all persons age 5-24 using one of two approaches is used:

1) Merge the birth history data for b3 from the BR file onto the PR file using cluster (hv001/v001), household (hv002/v002), and line number in the household roster (hvidx/b16). If the person is listed in the birth history of any women (BR file), the century month code (CMC) of the date of birth of the person (b3) is used as cmcdob.

2) If the CMC date of birth of the person is not known, a century month code for the date of birth is randomly imputed using a uniform distribution between the possible bounds of the date of birth, calculated as the CMC date of interview minus 12 times the age in years as a maximum and 11 months earlier as a minimum:

max_cmcdob = hv008 - hv105*12

min_cmcdob = max_cmcdob - 11

cmcdob = random(min_cmcdob, max_cmcdob)."

Can you please advice on commands in STATA or SPSS to impute cmcdob = random(min_cmcdob, max_cmcdob).

3) Finally, it would be very useful to know the absolute number of people of secondary school age for this survey please (age 14 to 17), so that i can have a reference to check if i am doing things well. I couldn't find it anywhere in the final report.

thank you!

[Updated on: Thu, 04 October 2018 11:50]

Report message to a moderator

Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Same women recurring?
Next Topic: individual and community factors of PNC utilization among mother's using 2016 Edhs
Goto Forum:

Current Time: Sat Sep 23 11:12:22 Coordinated Universal Time 2023