SweetPea

SweetPea specifies and synthesizes randomized experimental designs to generate unbiased trial sequences for rigorous, reproducible empirical research.

Key Features:

• Domain-specific language: A domain-specific programming language for declarative specification of experimental designs and constraints.
• Declarative factor and constraint specification: Allows researchers to define factors, levels, and inter-factor constraints to express experimental requirements.
• Randomized sequence synthesis: Samples and synthesizes randomized trial sequences while minimizing bias and potential confounds.
• Python implementation: Implemented in Python to perform specification parsing and sequence generation.
• Support for complex designs: Capable of representing and handling complex experimental structures and constraints common in empirical studies.
• Executable design translation: Translates declarative specifications into executable trial sequences for deployment in experiments.
• Known limitations: Current challenges include scalability and adaptability to highly specialized or unconventional experimental designs.

Scientific Applications:

• Psychology experiments: Designs studies requiring precise control over stimuli and trial sequences to avoid confounds.
• Neuroscience: Applies to experimental setups that demand complex counterbalancing and constraint specification.
• Machine learning experiments: Supports empirical evaluations that require randomized, unbiased trial sequences and controlled factor manipulations.

Methodology:

Declarative specification of factors and constraints in a domain-specific language and unbiased sampling/synthesis of randomized trial sequences implemented in Python.