The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » General » Education in single years (Education attendance?)
Education in single years Sat, 25 June 2022 03:59
 Daniel Pérez Messages: 10Registered: May 2022 Member
Hello,

I'm currently working on a project with the Ethiopian Demographic and Health surveys (PR module) and I want to take as an outcome variable the total years of education attended by the observations. However, I'm not totally convinced about whether HV108 "Education in single years" reflects this. The construction of the variable is not quite clear to me.

Daniel.
Re: Education in single years [message #24720 is a reply to message #24705] Tue, 28 June 2022 09:39
 Janet-DHS Messages: 385Registered: April 2022 Senior Member
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

Probably the best way to check the interpretation is to go to the questionnaire. In the 2019 household questionnaire, which is included as an appendix to the final report, the question is "What is the highest level of school (NAME) has attended? What is the highest grade or year (NAME) completed at that level?" this is a two-part question that is coded into hv106 and hv107, respectively. Then hv108 is constructed during data processing from the combination of hv106 and hv107.

Here are the labels for these three variables:

hv106 highest educational level attained
hv107 highest year of education completed
hv108 education completed in single years

To see how they are related, that is, how the third variable is constructed from the first two, I do some "reverse engineering" with the following Stata command: "tab hv107 hv106, summarize(hv108) means" , which produces the following table of means of hv108 for each combination of hv106 and hv107:

Means of education completed in single years

highest |
year of | highest educational level
education | attained
completed | primary secondary higher | Total
-----------+---------------------------------+----------
0 | 0 8 12 | 1.928264
1 | 1 . 13 | 2.3707113
2 | 2 . 14 | 3.3890339
3 | 3 . 15 | 6.5174477
4 | 4 . 16 | 5.5426389
5 | 5 . 17 | 5.7106918
6 | 6 . 18 | 6.4073853
7 | 7 . 19 | 7.1210593
8 | 8 8 20 | 8.0169252
9 | . 9 21 | 9.0375783
10 | . 10 22 | 10.023841
11 | . 11 23 | 11.195652
12 | . 12 24 | 12.75
don't kno | 98 . 98 | 98
-----------+---------------------------------+----------
Total | 3.9182786 9.9825762 15.239357 | 5.9666494

Ignore the "Total" column and the "Total" row! The table shows that hv108 values 0 through 24 can be interpreted as "highest grade" completed, as the label states. It is not the number of years attended school, which you might have been thinking. The construction is appropriate if "secondary" begins with grade 8. I believe that in some settings, secondary begins with grade 7.

I suspect that the higher values are reporting errors for some cases. Some of them are incompatible with current age. When you work with this variable, understand that there is a code 98, which must be omitted from any calculations such as means.
Re: Education in single years [message #25011 is a reply to message #24705] Fri, 19 August 2022 08:51
 Luke07 Messages: 1Registered: August 2022 Member
This is an excellent question. Although it is a continuous variable, we measure it discretely. So, for example, we measure someone's age discretely, but it is a continuous variable. We don't say, "I am 29.75" (though I do occasionally), but rather, "I am 29." So, while time is a continuous variable with an infinite range of values, it is almost always measured discretely. I will also suggest to use a subscription management platform for eLearning to manage and streamline your Education process.

Yes, this has been discussed previously, but take a look at your data and see how it was recorded. But it's cool to remember that time - in any form - is always continuous, and it all comes down to how the person collecting the data has actually collected the data.

[Updated on: Mon, 27 March 2023 02:49]

Report message to a moderator

 Previous Topic: Interpolated surfaces Next Topic: Age difference_using women data only
Goto Forum:

Current Time: Tue May 30 13:14:40 Coordinated Universal Time 2023