The DHS Program User Forum      
Discussions regarding The DHS Program data and results
Home » Dissemination and Capacity Strengthening » STATcompiler/mobile app » STATcompiler and final report differences
STATcompiler and final report differences [message #10] Wed, 13 February 2013 14:15 Go to next message
DHS user is currently offline  DHS user
Messages: 91
Registered: February 2013
Senior Member
Why are the numbers in STATcompiler sometimes different from the numbers in the final report?
Re: STATcompiler and final report differences [message #13 is a reply to message #10] Wed, 13 February 2013 14:27 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 581
Registered: January 2013
Senior Member
STATcompiler presents standard indicators calculated the same way, whenever possible, for all surveys. The indicators reported in survey reports may include country-specific adaptations that differ from the standard calculation used in STATcompiler. For example, countries may define "trained health care provider" differently.
Re: STATcompiler and final report differences [message #533 is a reply to message #13] Mon, 10 June 2013 14:59 Go to previous messageGo to next message
estammer is currently offline  estammer
Messages: 9
Registered: March 2013
Member

Hi,

I'm trying to determine whether the statcompiler or report data would be better to use to track Underweight women (BMI<18.5). The report data matches what was published in the "Nutrition Update 2010" report by Kothari and Abderrahim, but Statcompiler has many more data points than are included in the final reports, so it would be beneficial if I could use statcompiler. Do you know why there is a discrepancy between the numbers? Would really appreciate any help with this! Thanks so much!

Emily
Re: STATcompiler and final report differences [message #534 is a reply to message #533] Mon, 10 June 2013 15:40 Go to previous messageGo to next message
Sarah-DHS is currently offline  Sarah-DHS
Messages: 46
Registered: February 2013
Member
Dear Emily,

Monica Kothari, one of the authors of the Nutrition Update is requesting more information to be able to answer your question:

Does you mean more data points in the STATcompiler than the Nutrition Update 2010 or final reports?

Difference in the numbers between STATcompiler and report numbers are usually due to the variation in the universe of women selected for the analysis, de jure de facto etc.

Further for underweight among women the numbers should not vary too much between the STATcompiler and the reports. Can you share with us the specific discrepancies in the numbers so that it is easier to respond.

Thanks,
Sarah at DHS
Re: STATcompiler and final report differences [message #536 is a reply to message #534] Tue, 11 June 2013 10:22 Go to previous messageGo to next message
estammer is currently offline  estammer
Messages: 9
Registered: March 2013
Member

Hi Sarah,

Please disregard my comment about more values being available in Statcompiler. I realize now that the person who compiled my spreadsheet just didn't pull data for all of the countries.

In terms of the discrepancies, I put together a spreadsheet comparing data points across the three sources I mentioned: the Nutrition 2010 update, DHS final report and Statcompiler and have attached it to this post. These were just the three countries I looked at first, and as you can see, the data points from the Nutrition 2010 report and the DHS final reports match one another. However, in all cases, the statcompiler value is different from both the Nutrition 2010 report value, where available, and from the DHS final report value.

In my work, I have to pull values for multiple countries and multiple survey years at a time, so I would prefer to use Statcompiler as a source for this indicator when possible instead of going through each final report, but in order to do that, I would have to understand why the Statcompiler data is different from values reported in the DHS final reports and the Nutrition 2010 report, so I can decide if it is worthwhile.

Thanks so much for your help with this. I really appreciate it!

Best,

Emily



Re: STATcompiler and final report differences [message #543 is a reply to message #536] Fri, 14 June 2013 09:01 Go to previous messageGo to next message
Sarah-DHS is currently offline  Sarah-DHS
Messages: 46
Registered: February 2013
Member
Dear Emily,

From our expert Monica Kothari :
STATcompiler presents the indicator for defacto interviewed women with a child in the 5 years preceding the survey, not currently pregnant and at least one month post partum, which is a different definition used for this indicator compared to the final reports.

I would suggest to refer to the final report/Nut Update numbers as a preferred option to get the prevalence of underweight among women. If you still want to use the STATcompiler numbers, you should be aware that the universe of women for which underweight is reported in the SC is different than the DHS final reports.

Hope this answers your question,
Thanks,
Sarah at DHS
Re: STATcompiler and final report differences [message #11055 is a reply to message #543] Fri, 21 October 2016 15:26 Go to previous messageGo to next message
Reduced-For(u)m
Messages: 290
Registered: March 2013
Senior Member

Sara,

Just for reference, do the StatCompiler statistics all use the same height/weight standards? What I mean is, does a survey from 1999 use the CDC standards and a survey from 2010 use the WHO ones? Or have these been harmonized in some way for all StatCompiler data?

Thanks!

Re: STATcompiler and final report differences [message #11056 is a reply to message #11055] Fri, 21 October 2016 15:28 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 581
Registered: January 2013
Senior Member
They have all been updated to use the WHO Growth Standards set in 2006.
Re: STATcompiler and final report differences [message #11057 is a reply to message #11056] Fri, 21 October 2016 15:38 Go to previous message
Reduced-For(u)m
Messages: 290
Registered: March 2013
Senior Member

Many thanks Trevor!
Previous Topic: numerator and denominator numbers for the percentages presented in STATcompiler
Next Topic: Recode File - Universe definition
Goto Forum:
  


Current Time: Mon May 21 22:21:22 Eastern Daylight Time 2018