McEachern, Steven, Orten, Hilde, Perry, Ryan, Strand, Kristina
The Social Surveys Work Package (WP06) of the WorldFAIR project is focussed on the improvement of FAIR practices in the management of harmonised content in cross-national social surveys. The first report from the Work Package (Deliverable 6.1) provided an overview of the practices of comparative (cross-national) social surveys, through case studies of: (1) the European Social Survey (ESS) and (2) a satellite study, the Australian Social Survey International – European Social Survey (AUSSI-ESS). The focus of Deliverable 6.2 was on three recommendations (Rs) from that first report – the use of the DDI Lifecycle and variable cascade (R6.1 and R6.2), and requirements for formal registries of variables and reusable content (R6.5). In Deliverable 6.2, we outlined a proposed workflow for the processing of data harmonisation of social surveys that takes account of the practical steps required to bring diverse content together in a machine-actionable way, and that could best take advantage of external registered, persistent content. This workflow considers the core steps involved in the harmonisation process, key issues that occur in the processing of data during this process, and potential resolutions of these issues.
This third report then picks up from the previous two to test out proof-of-concept implementations of the workflows outlined in the second WP report, to trial the use of standardised workflows based on registry services available at the Australian Data Archive (ADA) and Sikt through their respective Colectica registries. The workflow steps are piloted with another comparative social survey, the International Social Survey Program (ISSP), to evaluate the Cross-Cultural Survey Harmonisation workflow as a suitable process for machine-to-machine based survey harmonisation.
The pilot demonstrated that the CCSH workflow established in Deliverable 6.2 and piloted here in Phase 3 of Work Package 6 appears to be a viable method for standardising and progressively automating the process of survey data harmonisation. The six-step workflow, based on well-established procedures in cross-cultural survey data, has been shown to be a suitable means of engaging with both human and machine-mediated processes. The development of the workflow based on Sikt and ADA processes for managing the harmonisation of ESS and AUSSI-ESS, and the piloting of the workflow with the similarly structured ISSP survey data. This suggests that the workflow itself may be well suited to projects of this type.
Having said this, the workflow pilot also shows that there is still a significant degree of human manual input required. To this end, the following recommendations are proposed coming out of this Phase 3 work:
- Establishment of standardised access controls both to data and metadata registries, to limit the need for less technical users to navigate access control systems
- Establishment of a code repository for interaction with social science metadata repositories.
- Establishment of mechanisms for reuse of conceptual variable and other reference metadata across the DDI standards ecosystem. (It was not clear for example how to use or reference a conceptual variable in the Sikt ESS metadata registry within the ADA harmonisation tool)
- Standardised practices and code libraries for the creation of DDI resource packages for external reuse (to facilitate the reuse in Recommendation 3).
The full report is available on Zenodo.
