Getting out of the swamp? Methodological reflections on using qualitative secondary analysis to develop research design

Abstract

In recent years, the possibilities and pitfalls of qualitative secondary analysis have been the subject of on-going academic debate, contextualised by the growing availability of qualitative data in digital archives and the increasing interest of funding councils in the value of data re-use. This article contributes to, and extends these methodological discussions, through a critical consideration of how the secondary analysis of thematically related qualitative longitudinal (QL) datasets might be utilised productively in qualitative research design. It outlines the re-use of two datasets available in the Timescapes Archive, that were analysed to develop a primary empirical project exploring processes of continuity and change in the context of men’s care responsibilities in low-income families. As well as outlining the process as an exemplar, key affordances and challenges of the approach are considered. Particular emphasis is placed on how a structured exploration of existing QL datasets can enhance research design in studies where there is limited published evidence.

Keywords:

• Qualitative secondary analysis (QSA)
• qualitative longitudinal data
• conceptual frameworks
• realist methodology
• Timescapes

Acknowledgements

Thank you to the primary research teams of FYF and IGE for being so generous with their time and data. Thanks also to Sarah Irwin and Kahryn Hughes for their comments on earlier drafts of this manuscript and to the anonymous reviewers.

Notes

1. The archive acts as a bespoke repository for primary researchers by facilitating their on-going analysis, and also brings related datasets, with shared substantive interests in family, relationships, identities and time, together, for enhanced analysis by both primary and secondary users (Neale & Bishop, 2012).

2. The bid was positioned specifically so that it could build out of, and extend, the existing secondary analysis work conducted as part of Timescapes. Choices of dataset were therefore limited to Timescapes in this instance, although there is a wealth of qualitative data archived elsewhere that might have relevance to this particular study. Qualidata UK is just one example where archived data might be accessed for similar purposes.

4. The heritage data was generated for methodological research carried out between 2003 and 2005, that sought to identify methodologies for accessing individuals and groups considered hard to reach (see Emmel & Hughes, 2012).

5. When a colleague questioned why Yow didn’t write about her participants’ race prejudice she wondered if it was because she liked them too much. In being asked to choose transcripts, it is likely that primary researchers will also be influenced by their interactions with participants and the emotions they associate with them. If working directly with primary research teams it is important for secondary analysts to be aware of this and to consider how this might influence their own sampling choices.

6. All names used throughout the article are pseudonyms assigned to participants by the primary research teams to protect their privacy and that of their families.

7. While there was some inevitable variation, for both studies metadata was available in the form of pen portraits for each wave for each participant, sample data, copies of interview schedules and field notes written by the primary research teams.