Deep Dive into Psychology Data

by Katherine Akers*

Continuing our research data education program for librarians at the University of Michigan Library, Susan Turkel (Psychology & Sociology Librarian) and I (eScience Librarian, CLIR Postdoctoral Fellow, and psychologist by training) recently led a Deep Dive into Psychology Data workshop that was attended by a total of 31 librarians. Similar to the Deep Dive into Ecology Data workshop led by Scott Martin (Biological Sciences Librarian), our workshop explored how librarians can become more familiar with the research data management landscape of a particular discipline, using psychology as an example.

Susan and I demonstrated how librarians can deepen their understanding of issues and current trends in research data management within the context of particular disciplines by investigating the data policies of three different entities: funding agencies, academic societies, and journals. Our deep dive into psychology data revealed several interesting findings that could be useful to psychology librarians who want to provide greater support for psychology research data management.

• Funding agencies: At the University of Michigan, many psychology faculty receive funding from the National Science Foundation and the National Institutes of Health, which have clear policies on research data sharing. Many others, however, receive funding from the Department of Defense and smaller public or private funding agencies, which may have unclear or no data sharing policies.

• Academic societies: The American Psychological Association (APA) expects psychologists to share their data with other competent professionals who wish to replicate or verify the claims made in publications, but only for that purpose. The American Psychological Society (APS) does not seem to have a formal data sharing policy.

• Journals: Most psychology journals do not expect, encourage, or require authors to share the data underlying their articles, but there are some interesting exceptions. For instance, the Archives of Scientific Psychology is a new journal (published by APA) that requires authors to make the data underlying their analyses accessible to others. However, as potential data users must complete an extensive application for data access, some have questioned how much this journal actually supports the concept or practice of open data. Also, Psychological Science (published by APS) recently implemented a new system whereby authors can earn a digital badge affixed to their article if they make their data publicly available. Furthermore, the new Journal of Open Psychology Data is a data journal that publishes papers describing particular research datasets housed in data repositories.

We also found a paucity of repositories specifically designed for psychology research data. Some repositories that may be relevant to certain psychologists include the National Database for Autism Research, Interuniversity Consortium for Political and Social Research (ICPSR), National Archive of Computerized Data on Aging (part of ICPSR), and OpenfMRI. Therefore, as psychologists have few disciplinary data repositories to turn to, they may need to rely more on institutional repositories and general repositories such as Harvard Dataverse Network or figshare.

Finally, a major issue that complicates data sharing and preservation in this discipline is that much psychology research involves human subjects, meaning that research data are likely to contain personal and/or sensitive information that cannot be openly shared. I’ve often heard researchers say that they can’t share their data because the Institutional Review Board (IRB) requires data to be destroyed at a certain time point after collection. But the University of Michigan’s IRB states that identifiable information, not all data, should be deleted or destroyed. This means that psychology datasets can potentially be shared and openly preserved if personal information can be removed, such as by following the US Department of Health & Human Services’ guidelines for de-identifying datasets. Alternatively, psychology datasets containing personal and/or sensitive information could be placed in repositories, such as ICPSR and Harvard Dataverse Network, that can restrict access to certain users.

In summary, the research data management landscape of psychology is probably not altogether different from many other disciplines that do not have well-established cultures of data sharing. Although there may not be strict data sharing policies or a wealth of resources that support data sharing in psychology, our deep dive certainly enhanced our knowledge of current trends in psychology research data management and will help us raise awareness of options for the storage and dissemination of research data among psychology researchers.

* Katherine Akers is now a Biomedical Research Data Specialist at Wayne State University’s Shiffman Medical Library.