Skip to content

About Psych-DS

Psych-DS is a community data standard for research in psychology and other behavioral sciences, which provides a flexible set of conventions for formatting and documenting scientific datasets. It is heavily inspired by the Brain Image Data Structure (BIDS) standard for fMRI data.

What is Psych-DS?

Psych-DS provides a simple and easy-to-adopt standard for organizing data in the psychological and behavioral sciences, which aims to help researchers satisfy FAIR principles for data sharing.

Key Goals

  1. To promote the adoption of good, consistent practices in the management of behavorial data
  2. To create a machine-readable format for these datasets that can support tools for their analysis, discovery, and preparation

Why do I need Psych-DS?

In the social and behavioral sciences:

  • Datasets can be arranged in many different ways and use various file formats
  • No consensus exists about how to organize and share project data
  • Even researchers within the same lab may arrange data differently
  • Lack of standardization leads to:
    • Miscommunications
    • Time wasted on reformatting/rearranging data
    • Difficulties indexing datasets for search tools
    • Challenges in writing reusable analysis scripts

Getting Started with Psych-DS

Documentation Resources

Core Components

To be compliant with Psych-DS, focus on two key aspects:

1. Metadata

Understanding Metadata

Metadata provides rich contextual information about a dataset, including:

  • Summary of contents
  • Creation and modification records
  • Essential context regarding the provenance/design of the study

Without Metadata:

Without standardized metadata:

  • Context is provided on an ad-hoc basis
  • Information lives in email communication or separate documentation
  • Sharing requires re-explaining context
  • Machine readability is impossible

Traditional Email-based Sharing

Email example showing ad-hoc data sharing

With Metadata

Standardized metadata provides:

  • Permanent attachment to data
  • Consistent information structure
  • Machine readability
  • Integration with semantic web standards

Key features:

  • Uses JSON-LD formatting
  • Integrates with Schema.org vocabulary
  • Supports both minimal and comprehensive documentation

Structured Metadata Example

{
    "@context": "https://schema.org",
    "@type": "Dataset",
    "name": "X Experiment",
    "author": {
        "@type": "Person",
        "name": "Test Researcher",
        "@id": "https://orcid.org/0022-0002-3833-3472"
    },
    "description": "A self-paced reading study with N participants...",
    "funding": {
        "@type": "Grant",
        "@id": "https://dx.doi.org/10.1080/02626667.2018.1560449",
        "name": "Y Grant"
    },
    "locationCreated": {
        "@type": "Place",
        "name": "Z facility",
        "address": "123 Main St..."
    }
}
2. File Organization

The Challenge of Unstructured Data

Without standardization:

  • Files may be scattered across directories
  • Naming conventions vary widely
  • Mixed formats and processing states create confusion

Unstructured Dataset Example

Example of unstructured data organization

Psych-DS File Structure

Key requirements:

  • Dedicated data/ subdirectory
  • CSV format for data files
  • _data suffix in filenames
  • "Keyword" formatting for file properties

Structured Dataset Example

Example of Psych-DS structured organization

Validation Tools

The Psych-DS team provides validation tools across multiple platforms:

  • Browser-based (best option for most researchers)
  • npm package (best option for developers)
  • Python library (coming soon)
  • R package (coming soon)

Features

  • Binary VALID/INVALID output
  • Detailed error reporting
  • Optional warning flags
  • Client-side processing for privacy

Privacy Commitment

All validation is performed locally. No data is uploaded or stored during validation. The browser-based tool uses client-side JavaScript exclusively, with no server interaction or database storage.

Our Repositories

Repositories