Feature

In 1961, NASA launched Enos the chimpanzee into orbit. The experiment—part of preparations for sending the first human into space—showed that the psychological demands of space travel were tolerable. But for researcher Jonathon D. Crystal, PhD, an associate psychology professor at the University of Georgia in Athens, it also shows the power of long-term data storage: Scientists don't have to have their own space programs to answer new questions about life in orbit; they can simply draw upon the data NASA collected on Enos and stored in its Life Sciences Data Archive (http://lsda.jsc.nasa.gov).

Although other disciplines have long been committed to data archiving, psychology has been slow to adopt it as a standard. Crystal and other proponents say that should change: Data archiving gives research greater impact, saves time and money and offers many other benefits.

"For the researcher, there are two sides to data archiving," says Howard Kurtzman, PhD, deputy executive director for science at APA. "One is to put their own data up. The other is to make use of archived data."

And since the 2001 National Institute of Mental Health workshop and subsequent report calling for archiving in animal cognition research (see "Further reading"), the data-archiving movement has gained ground. Both individual researchers and journal publishers, such as APA and the Psychonomic Society, are now beginning to make data, instruments and other supplemental materials available.

Archiving's advantages

Other disciplines in the physical and social sciences, such as astronomy, oceanography, economics and political science, have long had a culture of making data publicly available, says Steven J. Breckler, PhD, executive director for science at APA (see "Dealing with data" in the February Monitor on Psychology).

But psychologists have been reluctant to archive, he says, citing concerns about research participants' privacy. APA's Ethics Code only requires researchers to share data with other professionals who want to verify their claims. The National Science Foundation expects data sharing but recognizes individual researchers' needs. Even the National Institutes of Health, which requires data sharing for grants of more than $500,000, excludes data that could threaten confidentiality.

Confidentiality isn't the only concern. Crystal acknowledges that there are some situations where psychologists probably shouldn't archive their data right away. Developmental psychologists doing longitudinal research, for example, might want to wait before making data publicly available. "You don't want to reveal your data before you've published your papers," he says.

For the most part, Crystal encourages his peers to get over their qualms and share their research materials. Ask him about the advantages of making data publicly accessible, and he easily lists half a dozen:

• Wide dissemination of ideas. Public archives allow researchers to share everything they've got rather than just the small amount of data they can squeeze into a journal article.

• New research directions. By making all the data available, other researchers can use them to answer other research questions—even if those questions are unrelated to an original researcher's goals in collecting the data.

• Cost savings. "It costs a lot to produce data," Crystal emphasizes. Although archiving your data does take a little extra time, he says, "recycling" research materials can save both money and time for the field overall.

• Ethical considerations. Drawing on archived data from previous animal research, for instance, can minimize the number of animals used in research.

• Quality control. Access to raw data can allow others to replicate experiments or identify errors in research conclusions.

• A teaching tool. Data archives give professors a handy way to train undergraduates and graduates without all the hassle and expense that come with collecting original data.

Some examples

Most researchers are probably already "self-archiving" their data in spreadsheets on their own computers, says Crystal. But with the advent of easy-to-create Web sites, he and others now see a move toward posting data in publicly accessible archives. Crystal makes his raw data available at www.uga.edu/animal-cognition-lab/web-pages/data-archive.htm, for example.

There has also been growth in institutional archiving, says Brown University psychology professor Russell Church, PhD, who occasionally draws on other people's data in addition to archiving his own at www.brown.edu/Research/Timelab. Many journal publishers, including APA, now give authors the opportunity to add data or other supplementary materials to the electronic versions of their articles.

The Psychonomic Society exemplifies that trend, says Jonathan Vaughan, PhD, the society's archivist and a psychology professor at Hamilton College in Clinton, N.Y. The society, which publishes six journals, recently launched a new system for the archive it has been amassing since 2004. Instead of maintaining a separate archive, the society now invites contributors to append data, norms, stimuli or computer program code to their articles. That material is then integrated with the online versions of the articles.

To find the materials they need, users visit the site at www.psychonomic-journals.org/search, choose the advanced search option and enter keywords plus the word "supplemental." A search for "word norms supplemental" in the text keyword field, for example, produces 44 recent articles with supplemental material that features word norms.

The archive includes a wide range of materials, including raw data from reaction-time studies, pictorial and word stimuli in several languages and computer code for statistical analysis and data acquisition.

By tying its archive to published articles, the society hopes to assuage would-be users' concerns about the quality of data in the archive. "The article that the material is attached to has gone through a peer review," explains Vaughan. "It's not a self-archiving site where anyone can just put up anything."

Judith F. Kroll, PhD, distinguished professor of psychology, linguistics and women's studies at Penn State, is one researcher who draws upon the Psychonomic Society and other archival materials.

In their research on bilingualism, Kroll and her students often rely on archived word and picture norms. Instead of reinventing the wheel, she says, they begin their studies by checking the archive to see if other researchers have already developed materials they can use or adapt for their own studies.

Kroll and her students also use archived data from the Psychonomic Society and other sources, such as the English Lexicon Project at Washington University in St. Louis (http://elexicon.wustl.edu).

"One of the practical issues in my research is that we often have to compare results from bilinguals with results from monolinguals," she explains. "We want to know if the effects we're getting in these studies are really effects of bilingualism and not just some peculiar thing about the way we're doing the experiment or the materials or whatever." The archive, she says, facilitates that kind of checking.

Archived data also reduce the need for travel, adds Kroll, citing the case of a graduate student who needed data on monolingual speakers of Japanese. "We're in the middle of Pennsylvania—not a haven for bilingualism," she says. While she and her students do travel and collaborate with international colleagues, archives allow them to collect data without even getting up from their desks. "Having databases like this is just a wonderful resource," she says.

Future directions

Several factors would help spur the further development of data archiving, say Church and others. First, says Church, archiving should be voluntary rather than coercive. "If it's voluntary, people think of it as a great privilege to be able to get the extra exposure," he explains. "And one really does get extra exposure if someone uses the data."

A common format for data submissions would also be helpful, he says, so that researchers don't need to learn a different system every time they contribute. And the data need to be "raw" and easily extracted, adds Crystal. "To be useful to someone else, it has to be available in a way that you could download it in a spreadsheet or database," he explains. "If it's in a PDF—a static image of data—I'm not going to go and do all the work to input that into a usable form."

A centralized data archive at APA or some other professional society would be ideal, says Crystal. Centralization would help make it easier for researchers to find data. It would also help ensure long-term preservation, he says.

"We're at our institutions for a limited period of time," he says. "You'd like to think your data would outlast you."


Rebecca A. Clay is a writer in Washington, D.C.

Further reading

  • Crystal, J.D. (2004). "Data archiving in animal experimentation: Merits, challenges, and a case study." Behavior Research Methods, Instruments, & Computers 36, 656–660.

  • Kurtzman, H.S., Church, R.M., and Crystal, J.D. (2002). "Data archiving for animal cognition research: Report of an NIMH workshop." Animal Learning & Behavior, 30, 405–413.

  • Vaughan, J. (2004). "Editorial: A Web-based archive of norms, stimuli, and data." Behavior Research Methods, Instruments, & Computers 36, 363–370.