Research:Wikidata Use in Cultural Institutions (2019, Qualitative Research)
The Lead: To learn from institutions and be able to direct other people to information about their work, we did qualitative research on how and why cultural institutions use Wikidata. 16 people who use Wikidata in their work in cultural institutions told us about their motivations, activities and problems and explained how they work.
- Remote interviews and observations via videochat. A person who took notes participated, so the conversation could run more freely. If possible, people were asked to show and tell about their processes directly. The interviews were semi-structured and took about 40min-1h:20min.
- Jan did the analysis of data in Quirkos, partly in parallel to the data gathering to inform future interviews.
- Our perspective on our data was a critical-realist one (in contrast to e.g. a critical one), the analysis method followed Thematic Analysis
Research start May 2019, closing data gathering September 2019, Publishing on Commons and other Wikimedia projects November 2019
Policy, Ethics and Human Subjects Research
While the research did not touch any particularly sensitive topics, we did take care not to gather information which could identify people directly. In our records, people and institutions were pseudonymized. We invited the participants to member-check our preliminary results upon which several people helped to clarify or extend the information.
This is the executive summary. If you want to know more: The (→Title in the brackets) refers to the section in the report which gives more detailed explaination.
- Participants want to share data since they perceive their institutions and their context using more and more digital systems. They also see direct advantages of sharing data, like greater use of the institution’s data.
- Participants are motivated by the plan to re-import improved data from Wikidata back into local databases. Participants called this a “roundtrip” (→ People would like to “roundtrip” data).
- It is hard for participants to come up with data structures that (→ Getting data to Wikidata the “right way”)
- represent the collections and
- match the requirements of Wikidata
- What happens with imported data is hard to monitor. This makes it difficult for participants to spot new problems or improvements and to participate in the community (→ Improvements or degradation of data).
- Currently, there are few ways to metricize engagement on Wikidata.
- Many open data initiatives start with image imports to Wikimedia Commons and then transition (partly) to Wikidata (→ Intertwinement with Commons).
What supports Wikidata Use?
- Data is often already available in GLAM institutions
- Participants are enthusiastic about using Wikidata
- Umbrella organizations and big institutions pushing for open data (→ Kinds of institutions)
- Hopes for quality improvement via future roundtrip(→ People would like to “roundtrip” data).
- Support in Wikidata-appropriate modeling, reconciliation and import by service providers, Wikimedia Chapters or community members (→ Kinds of institutions).
- Participants are often “onboarded” to Wikidata via Commons (→ Intertwinement with Commons).
- In some areas, Wikidata has better usability or feature set than participant’s current software (→Institution’s data).
What hinders Wikidata Use?
- Data modeling is hard and needs both GLAM-skills and Wikidata skills (→ Practices of Modeling).
- Difficulties in understanding the culture and practices on Wikidata
- Tech-resources (→ Data Imports) and consulting needed (→ Data Imports, Kinds of institutions)
- Impact of Wikidata use hard to measure, but metrics are often needed in organization (→ Provide Metrics)
- Often no immediate and easy-to-demonstrate benefit by using Wikidata
- Problems of continuous engagement due to difficulties in monitoring data of interest. (→ Monitoring Data)
- Braun, V. and Clarke, V., 2006. Using thematic analysis in psychology. Qualitative research in psychology, 3(2), pp.77-101. http://eprints.uwe.ac.uk/11735/2/thematic_analysis_revised_-_final.pdf