The time is ripe for universities, journals, scientific societies and individual scientists to invest their resources in data archiving, according to a new report.
The report, "Data Archiving for Animal Cognition Research: Report of an NIMH Workshop," is scheduled to appear in Animal Learning & Behavior (Vol. 30, No. 4). It summarizes a July 2001 workshop that brought together behavioral scientists and archivists to discuss the future of data archiving for animal cognition research. Their main conclusion: Technology has advanced to the point where the benefits of archiving far outweigh the costs.
"Our typical practice has been to collect enormous amounts of data, distill it to its essential features and publish that," says psychologist Russell Church, PhD, of Brown University, an organizer of the workshop and one of the report's authors. "The data might be preserved in the original lab, but not in a form where they could be understood by others. Now, with the Internet, that restriction has disappeared."
The report suggests that most new data should be archived as soon as they are published, and that archiving should be strongly encouraged, but not mandated. It also states that a critical first step is creating standards for "metadata"--information that describes how an archive's contents are organized. And it encourages a variety of organizations to play active roles in promoting and funding archiving.
The report's authors are not alone in concluding that better systems for data archiving would benefit the behavioral sciences, says Merry Bullock, PhD, APA's associate direc-tor for science. "We have many, many different kinds of rich data sources," she says, "and we've just begun thinking about whether we as a discipline ought to do something coordinated and coherent to preserve these data."
Howard Kurtzman, PhD, of the National Institute of Mental Health, who collaborated with Church on the report, says he hopes it will inspire other communities within the behavioral sciences to establish their own guidelines. "We thought that picking one particular well-defined area with its own community, its own tradition and types of data, and working out the issues there might be a way to start the process for the rest of psychology and behavioral science," he says.