Blindkit was written to help the bench neuroscientist perform blinded, random assignment of experimental and control groups for preclinical studies, a step towards more rigorous science.
It is a code framework which solves three operational gaps that are often overlooked or informally implemented in practice:
- Reproducible, programmatic random assignment into experimental and control groups, either with a 50:50 split or a custom bias
- Generation of idempotent, unique alphanumeric blinding codes which eliminates duplication and reduces human error
- Full end-to-end auditability post-unblinding with timestamps of each step.
The core requirements for the experimentalist are:
- blindkit.py
- two separate github repositories, one for the experimentalist and one for an assistant blinder
- a human blinding assistant, who executes the random assignment, generates the blinded alphanumeric codes, and applies the labels onto the physical syringes or aliquots.
A detailed quickstart guide is written for the blinding assistant and is not reproduced here. It can be found at blindkit_quickstart_blinder.md. A general guide and one specifically for the experimentalist is forthcoming.
Blindkit is still under active development, and the code is presented as is, with no warranty or guarantee of accuracy. Any user should verify the source performs what they want before using this code for production or real experiments.
Copyright (c) 2026 Windsor Kwan-Chun Ting, PhD.
BlindKit is released under the GNU GPLv3.
The author retains the right to offer BlindKit under alternative licenses, including proprietary licenses.
Outputs generated by BlindKit (e.g., randomization tables, assignment files, logs) are not covered by the GPL.
ChatGPT was used in the initial generation of template code for this software. The code was then debugged and vetted for accuracy, and modified to accomplish the stated goals.