FAIR data quality: the need for standard operating procedures in an institutional repository

Topic: Data access conditions Summary/considerations The discussion in this group centred around the questions who has access and/or ownership over the data after the researcher left the institution. There should be a procedure in placet o still gain access to the data, but what does this workflow look like? If a data support professional responsible for handling requests, how do we make sure that they can make a sensible decision? Fold: developing the question 1. Wie heeft alleen toegang tot het sleutel bestand 2. Alleen de hoofdonderzoeker heeft toegang tot sleutel bestand dus kan beter helemaal niet gepubliceerd worden. 3. Alleen de hoofdonderzoeker heeft toegang tot sleutel bestand dus kan beter helemaal niet gepubliceerd worden. 4. Wie beslist welke data gedeeld worden? 5. Wie heeft toegang / eigenaarschap nadat onderzoeker uit dienst is Stretch: Wie heeft toegang/eigenaarschap nadat een onderzoeker uit dienst is? Exaggeration Wie gaat zich ontfermen over de tonnen zielige weesdata Inversion Wanneer de onderzoeker uit dienst treedt, worden de data vernietigd Extension Wie is de eigenaar van data met persoonsgegevens tijdens en na dienstverband Extension Wie is de eigenaar van data zonder persoonsgegevens tijdens en na dienstverband Extension Hoe krijg je toegang tot data waarvan de “eigenaar” er niet meer is Addition Welke procedure/flowchart wordt gehanteerd bij afwezigheid van 1e verantwoordelijke Subtraction Wie heeft toegang/eigenaarschap tot data (ook tijdens onderzoek) Topic: Data re-appraisal Summary/considerations The group started by exploring what form could take guidelines for re-appraisal of datasets. The discussion then broadened towards the selection criteria to preserve data (environmental footprint, cost, quality of the datasets, integrity) and the role(s) bearing the responsibility for making the re-appraisal decision. Fold: developing the question 1. What are some guidelines to help determine whether data should be deleted or preserved? 2. Whether data is preserved depends on requirements, and afterwards old DANS guides provide advise 3. Whether data is preserved depends on requirements that can be uniformized and DANS provide advice as mediator. Stretch: What are guidelines to help determine whether data should be deleted or preserved? Addition What are environmental aspects to consider (footprints) Addition Who should determine whether data should be preserved? Inversion Should data be preserved? Inversion Are guidelines needed to determine whether data should be deleted or preserved? Exaggeration Should all data be preserved indefinitely (kostenbesparing, kwaliteit van dataset, digitale footprint gaat omlaag, integer data/bewijs) Additional question What are the guidelines used for reappraisal and who reappraises the data? Topic: Data deletion after retention period Summary/considerations This group focused on the agents (‘who’) involved in data deletion, with a focus on the decision making (as opposed to the person actually deleting the data). The main question was who should have the final say if data should be kept, reduced or deleted? This involves a discussion on the quality aspects, or value, of the data, which may be diWerent per domain or discipline. Also in determining what these quality aspects are, it is important to consider who is the person in charge of doing that. Fold: developing the question 1. Lack of clarity 2. Who decides when to delete data? 3. Who decides what quality is required to keep the data 4. Who decides what quality and purpose is required to keep the data? 5. Who has the final say on the question if data should be kept or deleted? Stretch: Who should have the final say if data should be kept, reduced, or deleted? Exaggeration Who should have the final say if all data should be kept forever Inversion Who should definitely not have the say about keeping the data Subtraction Should data be kept after the retention period? Extension What quality aspects should be present so that the data can be kept after retention period? Addition Who decides what the quality aspects are? Addition Should there be diWerent quality aspects for diWerent research fields? Topic: Richness of metadata Summary/considerations In their discussion on the richness of metadata, this group focused on the use of controlled vocabularies. They feel that recommending controlled vocabularies, whether on national, international, or institutional level, contributes to the richness of the metadata, and thus enhances the usefulness and clarity. The question remains whether requiring the use of controlled vocabularies is necessary to enforce their use, and thus make the adoption more normative. Fold: developing the question 1. Controlled vocabulary versus “open” readme files? Or both combined? 2. Controlled vocabulary on subjects but also for example on characteristics of the data, standardization and completeness combined with readme file makes rich metadata 3. Recommending controlled vocabularies on an institutional or national level and providing templates for README files and other information adds to richness and usefulness of metadata 4. Recommending controlled vocabularies on institutional and on national level and providing templates for Readme files and other information adds to richness, usefulness, and clarity of the metadata 5. Semantic and organizational interoperability leads to richer metadata and higher reuse possibilities Stretch: Recommending controlled vocabularies on national, international level and providing templates for Readme file and other information adds to richness, usefulness and clarity of metadata! Exaggeration Requiring controlled vocabularies on national, international level and forcing templates for Readme file and other information adds to richness, usefulness and clarity of metadata! Addition Recommending multiple controlled vocabularies on national, international level and providing templates for Readme file and other information adds to easier/ quicker adoption, richness, usefulness and clarity of metadata! This results in the restriction of freedom of expression of the individual researcher. Subtraction Recommending controlled vocabularies on national, international institutional level and providing templates for Readme file and other information adds to richness, usefulness and clarity of metadata! Inversion Would requiring multiple controlled vocabularies and templates for readme files result in better adoption, richness, usefulness and clarity of metadata or would it restrict the richness of the metadata? Topic: Data selection Summary/considerations The group envisioned data selection as hard choices to be made by data contributors as responsibility bearers. They focused on how to assist them in this process (support roles, guidelines, policies) which triggered a reflection on how to define what meaningful data is / should be and how it links to the needs of reusers. Fold: developing the question 1. The process full of hard choices 2. How to assist data contributors in making the hard choices? 3. How to define what meaningful data is in order to help data contributors in making the hard choices? 4. How to define meaningful data so that data contributors know which data to store and share? 5. What is necessary to know in order to reuse the data? Stretch: How to assist data contributors in making the hard choices? Inversion Who decides what hard choices are (context)? Subtraction How to make (as a data steward) the hard choices (via guide)? Exaggeration How to assist data contributors making the hard choices in life? Extension what is the necessary to know in order to reuse data? Addition How to assist data contributors and stewards in making hard choices? Subtraction How to reduce the number of hard choices? (policy, management) Extension How to assist data contributors in defining what is meaningful data in their research? Exaggeration Why not share everything and let reusers decide what is meaningful? Unreadable transformation How to assist data contributors in deciding which data to exclude?