ELFE for dummies
In this section you find explanations on Elfe data which aim to make easier the understanding of the data sets for starters.
1. How children are represented in the data?
In this survey, each birth represents one observation and each family member is identified with a number. Therefore, a household with twins will be present in two observations. For example:
Number 1 always represents the ELFE baby, number 2 always corresponds to the twin (if there is a twin) and number 3 always represents the mother in the mother’s questionnaire, however the mother can be in several different positions in the father’s questionnaire. Also, the rest of household members have different ids depending on the order of response. In order to better understand the household grid system, go to Who is who in ELFE?
In the case of twins, there is one observation per baby. In this case, you will find all the information concerning the respective twin in the id 2 and the same values in all variables concerning the rest of household’s members. Thus, you have “repeated” observations.
2. Who is who in ELFE?
3. Variables names
All variables are preceded by a common root depending on the survey of origin:
- M00M/M00P corresponds to the maternity survey (Mother/Father)
- M02M/M02P corresponds to the 2 months survey (Mother/Father)
- A01M/A01P corresponds to the 1 year survey (Mother/Father)
Some other roots to know:
- M00X corresponds to medical record variables from maternity survey.
- M00E/M00F corresponds to weights variables from maternity survey.
- M02X corresponds to participation variables from 2 months survey.
- M02E/M02F corresponds to weights variables from 2 months survey.
- A01X corresponds to participation variables from 1 year survey.
4. Where can we find the mother information?
When you work on ELFE, you often need information concerning the mother of the ELFE child. In general, you find this information on the variables from mother’s survey (M02M/A01M), where the mother answered questions related to herself and all her household’s members. In the household grid, the mother is always in the third position.
If the father participated, you will find information from the father’s answers (M02P/A01P) regarding all his household’s members.
If the mother declared to have a relationship with the father and live with him, then the father’s answers also relate to the child’s household. If he does not live with the child, his answers relate to his household.
In order to find information on the mother while using the variables reported by the father, you just need to use the variable M02P_LIENTYP_i=2 to identify this individual i as the mother.
Example on STATA (filling information concerning mother’s education):
replace mother_education=0 if M02P_DIPLOME_`i’==1 & missing(mother_education) & M02P_LIENTYP_`i’==2
* Information concerning the father of the ELFE child in the mother questionnaire can also be found in the same way, using M02M_LIENTYP_i=3
5. Before making a request…
When you make a data request, you must not request variables preceded by * due to confidentiality reasons.
The constructed variables proposed in this website are now available through the Pandora platform, within a specific file named EQR12.In your data request, you must add in a comment that you need this additional file.
If you want to reconstruct the variables by yourself, you must ask the variables you will need in order to use the code.