Introduction

Figure 1: Graphic summary of the main aims of the project and their relationship to the phases of scientific research.

About

This site provides supporting documentation for a project funded by the U.S. National Science Foundation “Collaborative Research: HNDS-I: Enhancing Infrastructure For Discovery About Human Behavior”, NSF 2444730. The project period is March 1, 2025 to February 28, 2027.

We are grateful to U.S. taxpayers for their support of scientific research.

Principal Investigators

Karen E. Adolph, Ph.D.
Julius Silver Professor of Psychology and Neural Science
Professor of Applied Psychology and Child and Adolescent Psychiatry
New York University

Rick O. Gilmore, Ph.D.
Professor of Psychology
The Pennsylvania State University

Abstract

Video is a uniquely powerful source of information about human behavior. Video provides data about what people do, documentation about how research is carried out, and provides demonstrations that inform scientists and the public about research results. Researchers across the sciences routinely use video as data, documentation, and demonstration. One such resource is Databrary, the world’s only known large-scale repository specializing in storing and sharing research videos in behavioral sciences. Housed at New York University, Databrary was launched with NSF support. Databrary has removed the most significant barriers to video reuse while reinforcing core ethical principles of informed consent and restricted access to sensitive or identifiable data. Databrary now supports the research and teaching of thousands of scientists across the globe.

This project updates and enhances Databrary’s software to make it an even more powerful and useful platform for research and teaching about human behavior. This updated version accelerates the reuse of shared data by making it easier to find target video clips or specific videos. The updates include tools for users to create their own custom collections for research or teaching. Upgraded software make it easier for scientists to organize their data before it is shared widely. Expert staff provide professional curation assistance to make shared data maximally useful to the widest possible audience. Researchers also create software libraries in R and Python that empower users to write their own code to access Databrary. Through substantial improvements to Databrary, the project enables novel, innovative, and data-intensive research about the characteristics and consequences of human behavior using powerful, flexible, affordable tools available in a web browser. The enhancements enrich datasets already shared on Databrary–many funded by taxpayers–thereby increasing the value of prior public investments in research.

Aims

Aim 1: We will deliver on the potential of search as a tool for discovery by seeding an upgraded search engine with time-locked annotations and code books from PLAY and other big-data projects while expanding filtering to include a broad range of demographic variables.

Aim 2: We will create infrastructure to allow users to create custom “virtual” collections of recordings and annotations that automatically track the provenance of each contributing data element to ensure transparency.

Aim 3: We will make Databrary a more flexible tool for collecting, storing, and analyzing data from in-progress projects—prior to sharing with the research community—by creating private, temporary workspaces that can be easily imported into Databrary proper when it is time to share.

Aim 4: We will make video-based behavioral research a leader in research transparency and reproducibility by enhancing access to Databrary API functions via the publication of free open-source scripting libraries in R and Python, alongside sample scripts and other training materials.

Aim 5: We will make existing shared datasets more readily discoverable by improving searchable metadata. We will expand the number and diversity of shared datasets by providing expert curation services to the owners of private datasets and sharing them.

--- title: Introduction --- ![Graphic summary of the main aims of the project and their relationship to the phases of scientific research.](include/img/Fig-1-Aims-v2.png){#fig-aims} ## About This site provides supporting documentation for a project funded by the U.S. National Science Foundation "Collaborative Research: HNDS-I: Enhancing Infrastructure For Discovery About Human Behavior", [NSF 2444730](https://www.nsf.gov/awardsearch/showAward?AWD_ID=2444730&HistoricalAwards=false). The project period is March 1, 2025 to February 28, 2027. We are grateful to U.S. taxpayers for their support of scientific research. ## Principal Investigators **Karen E. Adolph, Ph.D.** \ Julius Silver Professor of Psychology and Neural Science \ Professor of Applied Psychology and Child and Adolescent Psychiatry \ New York University **Rick O. Gilmore, Ph.D.** \ Professor of Psychology \ The Pennsylvania State University ## Abstract Video is a uniquely powerful source of information about human behavior. Video provides data about what people do, documentation about how research is carried out, and provides demonstrations that inform scientists and the public about research results. Researchers across the sciences routinely use video as data, documentation, and demonstration. One such resource is [Databrary](https://databrary.org), the world's only known large-scale repository specializing in storing and sharing research videos in behavioral sciences. Housed at New York University, Databrary was launched with [NSF support](https://www.nsf.gov/awardsearch/showAward?AWD_ID=1238599&HistoricalAwards=false). Databrary has removed the most significant barriers to video reuse while reinforcing core ethical principles of informed consent and restricted access to sensitive or identifiable data. Databrary now supports the research and teaching of thousands of scientists across the globe. This project updates and enhances Databrary's software to make it an even more powerful and useful platform for research and teaching about human behavior. This updated version accelerates the reuse of shared data by making it easier to find target video clips or specific videos. The updates include tools for users to create their own custom collections for research or teaching. Upgraded software make it easier for scientists to organize their data before it is shared widely. Expert staff provide professional curation assistance to make shared data maximally useful to the widest possible audience. Researchers also create software libraries in R and Python that empower users to write their own code to access Databrary. Through substantial improvements to Databrary, the project enables novel, innovative, and data-intensive research about the characteristics and consequences of human behavior using powerful, flexible, affordable tools available in a web browser. The enhancements enrich datasets already shared on Databrary--many funded by taxpayers--thereby increasing the value of prior public investments in research. ## Aims **Aim 1**: We will deliver on the potential of search as a tool for discovery by seeding an upgraded search engine with time-locked annotations and code books from PLAY and other big-data projects while expanding filtering to include a broad range of demographic variables. **Aim 2**: We will create infrastructure to allow users to create custom “virtual” collections of recordings and annotations that automatically track the provenance of each contributing data element to ensure transparency. **Aim 3**: We will make Databrary a more flexible tool for collecting, storing, and analyzing data from in-progress projects—prior to sharing with the research community—by creating private, temporary workspaces that can be easily imported into Databrary proper when it is time to share. **Aim 4**: We will make video-based behavioral research a leader in research transparency and reproducibility by enhancing access to Databrary API functions via the publication of free open-source scripting libraries in R and Python, alongside sample scripts and other training materials. **Aim 5**: We will make existing shared datasets more readily discoverable by improving searchable metadata. We will expand the number and diversity of shared datasets by providing expert curation services to the owners of private datasets and sharing them.