Data Management and Data Sharing in Psychological Science: Revision of the DGPs Recommendations

Abstract

Providing access to research data collected as part of scientific publications and publicly funded research projects is now regarded as a central aspect of an open and transparent scientific practice and is increasingly being called for by funding institutions and scientific journals. To this end, researchers should strive to comply with the so-called FAIR principles (of scientific data management), that is, research data should be findable, accessible, interoperable, and reusable. Systematic data management supports these goals and, at the same time, makes it possible to achieve them efficiently. With these revised recommendations on data management and data sharing, which also draw on feedback from a 2018 survey of its members, the German Psychological Society (Deutsche Gesellschaft für Psychologie; DGPs) specifies important basic principles of data management in psychology. Initially, based on discipline-specific definitions of raw data, primary data, secondary data, and metadata, we provide recommendations on the degree of data processing necessary when publishing data. We then discuss data protection as well as aspects of copyright and data usage before defining the qualitative requirements for trustworthy research data repositories. This is followed by a detailed discussion of pragmatic aspects of data sharing, such as the differences between Type 1 and Type 2 data publications, restrictions on use (embargo period), the definition of "scientific use" by secondary users of shared data, and recommendations on how to resolve potential disputes. Particularly noteworthy is the new recommendation of distinct "access categories" for data, each with different requirements in terms of data protection or research ethics. These range from completely open data without usage restrictions ("access category 0") to data shared under a set of standardized conditions (e.g., reuse restricted to scientific purposes; "access category 1"), individualized usage agreements ("access category 2"), and secure data access under strictly controlled conditions (e.g., in a research data center; “access category 3"). The practical implementation of this important innovation, however, will require data repositories to provide the necessary technical functionalities. In summary, the revised recommendations aim to present pragmatic guidelines for researchers to handle psychological research data in an open and transparent manner, while addressing structural challenges to data sharing solutions that are beneficial for all involved parties.

Promoting scientific integrity through open science in health psychology: results of the Synergy Expert Meeting of the European health psychology society

“…The Synergy Expert Group put forward a set of recommendations for actions to be taken to promote scientific integrity through open science in health psychology for EHPS members and other stakeholders, including behavioural researchers, gatekeepers, and end-users (Table 1). Our recommendations largely align with the existing recommendations and statements on open data and data management in psychological science (British Psychological Society, 2020;Gollwitzer et al, 2020). Next, we summarize each action and outline their implications for implementation by key stakeholders.…”

An open developmental science will be more rigorous, robust, and impactful

“…This illustrates how storing data in a repository can make it easy to generate reproducible analytical workflows from early stages in a project discovery. Seeking "broad consent" for a wide range of future uses of data can pose challenges, especially in an evolving international regulatory environment (Gollwitzer et al, 2020), but securing the broadest possible consent maximizes the potential for secondary re-use and future discovery. Crucially, the presumption that sensitive or identifiable research data cannot or must not be shared in order to protect participants can be, in our opinion, a misconception.…”

Ten steps toward a better personality science - how quality may be rewarded more in research evaluation

“…https://psych-ds.github.io/). The Open Science committee of the German Psychological Society is now working on a revision of their data sharing recommendations (Schönbrodt, Gollwitzer & Abele-Brehm, 2017;Gollwitzer et al, 2020).…”

“…As the same data collection effort may spawn several papers, we propose that all of the data, code, and materials associated with a given project should be stored in a single place online. If not all of the data is supposed to be publicly available from the get-go, access to parts of the data may be embargoed for a limited and defined amount of time, during which the original authors may conduct their planned analyses on those parts of the data (Gollwitzer et al, 2020). The directory containing all the content related to a project should become associated Better Personality Science 27 with the first research article based on this content, and we should make it a custom that all subsequent papers based on the same data (by the same or different authors) should cite that first paper.…”