Data Sharing in Psychology

Narrowly defined, data sharing is the practice of making scientific research data available to other researchers. However, the term is often used to include a variety of open-science practices, including making data, methodology (e.g., coding scheme), analytic syntax, and other research materials available to other researchers, as well as the reuse of those resources by others. There are multiple avenues for data sharing, for example data repositories (either subscription-based or free) or direct request to the researcher. Data sharing is a fairly common practice in the life and earth sciences. Excepting a handful of longitudinal projects, psychology lacks this robust historical precedent for sharing data. In fact, in the not-so-distant past, institutional review boards typically required that data be destroyed after a preset period in order to protect participants’ privacy—and some still do. And many researchers still do not take the first step—modifying their informed consent procedures to include explicit consent to share. Although still not frequent, data sharing in psychology is becoming more common. In part, this trend is being driven by the requirements set by publications and funding agencies. For publications, data sharing is intrinsic to transparency and replication of study findings. For funders, data sharing ensures greater return on investment—that expensive and time-consuming primary data collection does not wind up sitting on a dusty shelf, but rather can be reused for secondary data analysis to answer new questions. In psychology as in other fields, technological improvements in storage capacity and computing power have also facilitated data sharing and reuse. While many psychologists are still concerned that data sharing will result in being “scooped” or found in error, there is increasing recognition of the benefits of data sharing. First, data repositories ensure that data are archived, and that the burden of preservation does not fall on the researcher or the researcher’s institution. Sharing also increases the pace of scientific progress, as researchers can build on each other’s work. For example, researchers can learn how other experts approached measurement or coding of a given outcome. In replication studies, inconsistent findings can point to contextual variations in the construct under study, rather than researcher error. And in a field where null findings are often difficult to publish, sharing allows these data to be included in meta-analyses across studies to examine broader impacts. Most importantly, data sharing enhances transparency, a key ingredient in the scientific process.