Interested in adding more dataset metadata about your repository that gives users a greater sense of the data's trustworthiness?

[Julian Gautier]

Hi everyone,

The team at Harvard is working on changes to Dataverse so that repositories can publish machine-readable metadata, defined by research communities, that lets users know more about a dataset's fitness for reuse, provenance, sensitivity, or other domain-specific indicators of trust. We've been calling this work the Trusted Data project, and you can read more about it at https://github.com/IQSS/dataverse-pm/issues/425.

As part of this project, we're interested in working with folks who manage Dataverse repositories to learn if we should also include, in each dataset, more metadata about repositories themselves, to give users a greater sense of the trustworthiness of the data.

For example, would someone who finds a dataset feel more confident in the data if they saw that it was published in a repository or collection that had the CoreTrustSeal certification, or in a repository that met certain requirements of that certification, or that had certain characteristics from the NIH's "Desirable Characteristics for Data Repositories"? Could we work out a way to describe such repository characteristics so that they're included in each dataset's metadata?

If you're interested in working with us or have questions about this effort, please email me at julian...@g.harvard.edu by December 19, 2025. Then during the week of January 5 (after we get back from winter holidays), we'll reach out to everyone who's expressed interest to start brainstorming and testing ideas.

Cheers!
Julian

Julian Gautier (he/him)
Product Research Specialist, IQSS

[Julian Gautier]

Hi again everyone! Just sharing more information about the scope of this effort, mostly that the deliverable will be a brief white paper where we'll list a set of characteristics about repositories that repository managers, curators, and authors of relevant literature think signal trust, and where we'll discuss how these characteristics might be represented and shared in Dataverse.

The GitHub issue at https://github.com/IQSS/dataverse-pm/issues/463 describes more of the scope of the effort and tasks.

Thanks to everyone who's expressed interest so far.