Today, people generate and store more data thanever before as they interact with both real and virtual environ-ments. These digital traces of behavior and cognition offercognitive scientists and psychologists an unprecedented op-portunity to test theories outside the laboratory. Despite gen-eral excitement about big data and naturally occurring datasetsamong researchers, threeBgaps{\^{}}stand in the way of theirwider adoption in theory-driven research: theimaginationgap, theskillsgap, and theculturegap. We outline an ap-proach to bridging these three gaps while respecting our re-sponsibilities to the public as participants in and consumers ofthe resulting research. To that end, we introduce Data on theMind (http://www.dataonthemind.org), a community-focusedinitiative aimed at meeting the unprecedented challenges andopportunities of theory-driven research with big data and nat-urally occurring datasets. We argue that big data and naturallyoccurring datasets are most powerfully used tosupplement—not supplant—traditional experimental para-digms in order to understand human behavior and cognition,and we highlight emerging ethical issues related to the collec-tion, sharing, and use of these powerful datasets.