The DHS Program User Forum: Unmet Need » Revised unmet need calculation for Tanzania 2010

Home » Topics » Unmet Need » Revised unmet need calculation for Tanzania 2010

Show: Today's Messages :: Show Polls :: Message Navigator

Revised unmet need calculation for Tanzania 2010 [message #2831]

Fri, 29 August 2014 09:36

eschelar
Messages: 2
Registered: August 2014

Member

Hello all,

I am trying to calculate revised unmet need for Tanzania 2010 using the survey specific code for Stata available from http://dhsprogram.com/topics/Unmet-Need.cfm. I have not been able to get a comparable estimate for married women to that found in the Analytical Study 25: Revising Unmet Need for Family Planning
by Bradley, Croft, Fishel, and Westoff, 2012, published by ICF International
measuredhs.com/pubs/pdf/AS25/AS25.pdf or StataCompiler (should be 25.3%). I get 22.3%. Furthermore, when I tabulate v626 (the previous version of unmet need which is available in the Tanzania 2010 dataset) for married women I get 21.3% which also does not match the published reports. According to the revising unmet need publication, the total unmet need among currently married women is 25.3% for both the revised and original definition of unmet need.

I see in the individual recode documentation for Tanzania it says that corrections were made to the 2nd release of the data, and that the values of V626 have changed slightly due to a change in the recode program. Does anyone know what the new values should be for this variable or have any suggestions what the problem may be with calculating the revised unmet need measure?

Thanks very much!
Erin

INDIVIDUAL RECODE DATA ALERT (VERSION 63)

Corrections have been made to the Individual Recode data file (TZIR62). The changes implemented in the new version of the data (TZIR63) are as follows:

V361 values slightly changed, due to a change in the recode program.
V362 values slightly changed, due to a change in the recode program.
V624 values slightly changed, due to a change in the recode program.
V625 values slightly changed, due to a change in the recode program.
V626 values slightly changed, due to a change in the recode program.

Attachment: Revised unmet need.do
(Size: 6.14KB, Downloaded 1108 times)

Report message to a moderator

Re: Revised unmet need calculation for Tanzania 2010 [message #3024 is a reply to message #2831]

Sun, 05 October 2014 13:43

Trevor-DHS
Messages: 803
Registered: January 2013

Senior Member

Hi Erin,

Your results are correct. The correction that was made to the recode file was also made in the code for calculating the revised definition of unmet need - Tanzania uses s309b instead of v302 for ever use of contraception. If you want to reproduce the original result published in AS25 you can comment out the line of survey specific code for Tanzania 2010 as follows:

* survey-specific code for Tanzania 2010
*cap replace infec=1 if v502==1 & v512>=5 & v512!=. & (tsinceb>59 | tsinceb==.) & s309b==0 & pregPPA24!=1 & v000=="TZ5" & (v007==2009 | v007==2010)

This produces the results as published, but the corrected version produces the correct results. Thanks also for pointing out that the STATcompiler still has the wrong result. We will correct this.

Report message to a moderator