30 Nov-2 Dec 2020 Online (France)
Question driven harmonisation of data - the variable cascade in practice
Jon Johnson  1@  , Bill Norley  1@  
1 : CLOSER, UCL Institute of Education

Harmonisation of data is largely a manual process. Although there are naming conventions that are embedded in many studies (especially panel studies), that can provide strong clues to the equivalence of variables, it is by no means perfect and requires a priori knowledge. 

The capture of questions (in DDI-Lifecycle instruments) linked to the output variables, allows generic methods to be developed to reliably identify and organise variables which may be candidates for harmonisation.

The presentation will discuss the approach CLOSER has taken to develop user facing tools that allows researchers to determine the criteria they wish to apply to harmonising variables, these include, the question, response domain, mode, and study characteristics.

Online user: 1