Raison d'être des variables construites

The Elfe team and its associate researchers wish to make these constructed variables available for the scientific community’s use. They’re stored in EQRs, and can be matched with your Elfe data selection. On this page, we explain some reasons why you might prefer to use these constructed variables. More information on individual EQRs is available on the EQR page.

The Elfe data collected from families contains a lot of information, and can make analyses quite complicated. It requires in-depth knowledge about how it was collected, and scientific skills to be analysed to their full potential. It may also turn out that some variables reveal participants’ identity, sometimes indirectly. Consequently, researchers can’t use them, even if it means missing out on a wealth of information. Here are different examples:

• For each wave, certain variables are multiplied by the number of people of the household (as many as 12 or 25 depending on the wave). This is the case when we describe characteristics for everyone living with the cohort member, such as their living situation, their age, their diplomas, etc. To save researchers time and to save them from having to do complicated data management, we make synthetic variables available, e.g. parents’ level of education.
  
  
• Some variables reveal participants’ identity (e.g. when answers are not frequent enough, or when questions are open-ended). These answers are coded on a case-by-case basis to be delivered without risking participants’ identification.
  
  
• Certain variables are only measured in earlier waves, e.g. we won’t ask about a person’s level of education again unless they’ve pursued more studies. To avoid data users having to ask for variables from every wave to make sure they have all of the necessary information, variables aggregating information from different waves are available.
  
  
• Similarly, some variables can be corrected longitudinally. For example weight and height can be corrected thanks to measurements over time. Corrected anthropometric variables are available. 
  
  
• Some sweeps contain complex variables. For example, regarding screentime, respondents could answer about how much time they spent on screens in minutes or in hours. They were also asked about screentime on weekdays Monday to Friday, and on weekend days Saturday and Sunday. Lastly they were asked about their use of different screens. To help with analyses, variables have been constructed to summarise the information, and to homogenise their measurement.
  
  
• Lastly, some calculations were needed for some variables (scores for example). These are also available.