The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » General Data Questions » Education variable inconsistent (I find inconsistencies between years of schooling and age.)
Education variable inconsistent [message #17999] Tue, 13 August 2019 05:07 Go to next message
Kerstin is currently offline  Kerstin
Messages: 1
Registered: August 2019
Member
I am using the DHS data of African countries. By cross-checking the educational variable (hv108 education completed in single years), I found inconsistencies with respect to the age of the respondent (hv105). For instance in the Angola PR dataset 2015-16, there are quite some observations where it is reported that the individual went longer to school than (s)he is old. Is this a calculation error of hv108 or a misreport? How should I deal with it?

Best regards
Kerstin
Re: Education variable inconsistent [message #18017 is a reply to message #17999] Sat, 17 August 2019 14:19 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3016
Registered: February 2013
Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

I see that there are 372 cases in which hv108 is coded "98", for "don't know". These are cases for which hv108>hv105, numerically, but they are not errors. However, I see 16 cases with hv108>hv105 and hv108 is in the range 5 to 13 and hv105 is in the range of 3 to 10. These are definitely coding errors. They should have been discovered during editing and given a value of hv108=97, "inconsistent".

Of course, years of schooling should not just be greater than hv105, it should be greater than hv105+5 (approximately). In your analysis, I suggest recoding hv108 to 97 if hv108>hv105+5.
Previous Topic: Children recode - same caseid
Next Topic: Education variable (PR file)
Goto Forum:
  


Current Time: Thu Mar 28 07:35:05 Coordinated Universal Time 2024