Once upon a time, psychologists considered it poor form to use data collected by other scientists. But the way researchers gather data is changing. "With the zeitgeist of open science, people are more open to the idea of sharing data—and using what’s already out there," says Amber Story, PhD, associate executive director for scientific affairs in APA’s Science Directorate. "There is so much data available that can answer the questions we ask as behavioral scientists."
Today’s researchers are finding that existing data previously collected by others can answer new and important questions—without the big price tag that comes with launching a study from scratch.
"There’s just not as much money as there used to be for research," says Kathy Hirsh-Pasek, PhD, a developmental psychologist at Temple University and data-sharing advocate who uses large public data sets to teach research methods. "Especially for junior researchers, secondary data analysis can allow you to break in earlier—and allows our field to flourish and move our science forward."
Existing data sets have advantages beyond saving money, Hirsh-Pasek says. Many are large and well-established, with rigorous methodology that has stood up to scrutiny. In recent years, the so-called replication crisis in psychology has underscored the importance of having sample sizes large enough to have statistical power, she says. "Secondary data analysis helps with that, because you have some tried-and-true data sets out there that are highly trusted and have stood the test of time."
Repurposing data doesn’t mean rehashing what others have done, adds Karen Adolph, PhD, a professor of psychology and neuroscience at New York University and director of Databrary, a repository for research video from behavioral studies. "Video captures so much of the richness and complexity of behavior and the surrounding context that it is uniquely suited for research reuse," she says. For example, she studies motor development, and often watches videos of infant participants with the sound turned off. "Someone else could take exactly the same videos and listen to the language or study infants’ emotional expressions or social interactions," she says. "It’s time-consuming and expensive to collect such data with infants and children. Now researchers’ efforts have a new life—something one researcher doesn’t care about can be a treasure trove for another lab."
Adding diversity
Tapping into other data sets also allows scientists to more easily expand their research to other geographic regions and cultures, Adolph adds, addressing the criticism that psychology research too often focuses on Western, educated, industrialized, rich and democratic (or "WEIRD") societies.
Of course, in order to reuse data, someone has to share it first. Luckily, it’s fast becoming the norm to share data publicly. U.S. federal agencies require that publications and data from taxpayer-funded research be made available publicly. And funders are increasingly wary of paying for research that has already been done. When researchers submit a grant to the United Kingdom’s Economic and Social Research Council, for example, they must certify that no data already in existence can answer their research questions.
Many large public databases also accept data from researchers, providing a place for scientists to store and share their own findings with students and colleagues—including potential collaborators they might not have met yet. Another perk of large, trusted public data repositories: They take precautions to protect the privacy of research participants, making them a secure place to store data without jeopardizing confidentiality.
In other words, there are good reasons to share as well as to borrow. "People have already spent millions of bucks collecting this data, and I think it’s really important to explore what’s there to see if you can address your questions," says Hirsh-Pasek. "Secondary data analysis is the way of the future."
Where to find good data
The more that data sharing becomes the norm, the easier it’s becoming to find high-quality data available for secondary reuse. Unfortunately, there’s no central repository that lists all of the large-scale data sets available to researchers. But these resources—including single-source data sets and repositories of data from multiple studies—are a good place for behavioral and social scientists to start:
Logitudinal data sets
■ Add Health (www.cpc.unc.edu/projects/addhealth/documentationopens in new window). Beginning with a survey of adolescents in 1994, the National Longitudinal Study of Adolescent to Adult Health (Add Health) has followed participants to study development and health trajectories across adolescence and into adulthood.
■ Behavioral Risk Factor Surveillance System (cdc.gov/brfss/data_documentation/index.htmopens in new window). Since 1984, the Centers for Disease Control and Prevention’s BRFSS telephone survey has collected data on adults’ risk behaviors and preventive health practices.
■ General Social Survey (gss.norc.orgopens in new window). Since 1972, the GSS has gathered data on American attitudes, behaviors and attributes, covering topics such as stress, morality, crime and violence, intergroup tolerance and psychological well-being.
■ Health and Retirement Study (hrs.isr.umich.edu/data-productsopens in new window). The HRS is a multidisciplinary, longitudinal study that surveys a representative sample of American adults to address questions related to aging.
■ MIDUS (midus.wisc.eduopens in new window). Since 1995, the Midlife in the United States (MIDUS) study has investigated the role of behavioral, psychological and social factors in age-related variations in health and well-being in a national sample of Americans.
■ Monitoring the Future (monitoringthefuture.orgopens in new window). Launched in 1975, this ongoing study investigates the behaviors, attitudes and values of American secondary school students, college students and young adults.
■ Neighborhood Atlas (neighborhoodatlas.medicine.wisc.eduopens in new window). The Neighborhood Atlas is a tool to help researchers visualize socioeconomic measures such as income, education, employment and housing quality by mapping those factors at the neighborhood level.
■ NHANES (cdc.gov/nchs/nhanes/index.htmopens in new window). The National Health and Nutrition Examination Survey (NHANES) combines interviews and physical examinations to assess the health and nutritional status of adults and children in the United States.
■ Panel Study of Income Dynamics (psidonline.isr.umich.eduopens in new window). Since 1968, the PSID has followed participants and their descendants, collecting data on employment, income, spending, health, marriage, childbearing, child development, education and more.
■ U.S. Census (www.census.gov/data.htmlopens in new window). Data from the decennial U.S. Census contains information on factors including demographics, employment, housing, income and family composition.
Data repositories
■ Allen Brain Atlases (portal.brain-map.orgopens in new window). The Allen Institute for Brain Science provides high-quality reference atlases covering themes such as cell types, gene expression, brain development and neural connectivity, from both mice and humans.
■ Databrary (databrary.orgopens in new window). Databrary is a digital library for behavioral scientists to store, share and reuse research videos, audio data and related metadata.
■ Data.gov. The federal government’s data.govopens in new window is a searchable repository for federal, state, local and tribal government data in areas including education, health and economic indicators.
■ Google Dataset Search (toolbox.google.com/datasetsearchopens in new window). Still in beta form, Google Dataset Search enables users to use a simple keyword search to find data sets stored in thousands of repositories across the internet.
■ International Neuroimaging Data-Sharing Initiative (fcon_1000.projects.nitrc.orgopens in new window). INDI was launched to promote sharing of brain imaging data among the neuroimaging community, and includes the 1000 Functional Connectomes Project, which contains more than 1,200 resting fMRI data sets from dozens of sites.
■ Inter-university Consortium for Political and Social Research (icpsr.umich.eduopens in new window). The ICPSR includes more than 750 academic institutions and research organizations and maintains an archive of data in the social and behavioral sciences, with specialized collections in fields such as education, aging, terrorism, criminal justice and substance use.
■ LearnSphere (learnsphere.orgopens in new window). LearnSphere provides a repository for learning science researchers to store their data and access data sets across a wide variety of educational data.
■ NICHD Datasets (nichd.nih.gov/research/resources/indexopens in new window). The Eunice Kennedy Shriver National Institute of Child Health and Human Development shares data from its wide-ranging research, with data sets available on topics such as genomics, brain development and intervention trials.
■ NIMH Data Archive (data-archive.nimh.nih.govopens in new window). The National Institute of Mental Health Data Archive contains data from hundreds of research projects with human subjects, across many scientific domains, in six NIMH repositories:
- National Database for Autism Research
- National Database for Clinical Trials Related to Mental Illness
- Research Domain Criteria Database
- Adolescent Brain Cognitive Development (ABCD) Study
- Connectome Coordination Facility
- Osteoarthritis Initiative
■ OpenNeuro (openneuro.orgopens in new window). This platform is designed for sharing neuroimaging data.
■ Open Science Framework (osf.ioopens in new window). A free, open- source service of the Center for Open Science, the OSF provides a place for researchers to store and share data and request access to others’ data sets.
■ Stanford Education Data Archive (cepa.stanford.edu/seda/overviewopens in new window). This archive from the Stanford Center for Education Policy Analysis includes detailed data on educational conditions, contexts and outcomes in school districts and counties across the United States.
