All free tools
Free · no signup

Random group generator

Paste your participants (one per line), pick the number of groups, and get a balanced random allocation — optionally seeded so the assignment is reproducible. Free, in your browser.

Let the AI do this — sign up freeUpload your data, get the write-up back.

Participants

Any word or number. The same participants + seed always give the same allocation — record it for reproducibility.

Runs in your browser — names never leave this page.

When to use this

Use this whenever participants must be assigned to conditions by chance rather than by convenience: treatment and control arms of an experiment, counterbalancing task orders, splitting a class into discussion groups, or dividing a sample for split-half validation. Randomization is what licenses causal language — assigning 'whoever signed up first' to treatment builds selection bias directly into the design. The seed matters for your write-up: 'participants were randomly assigned' is more credible when the assignment is reproducible.

What you get

  • Unbiased complete randomization via Fisher–Yates — no allocation is more likely than any other.
  • Balanced sizes: groups differ by at most one participant.
  • A recorded seed, so the allocation can be reproduced and audited.
  • Duplicate names are allowed (e.g. twins with the same name) — each line is one participant.

Common mistakes

  • Re-rolling until the groups "look right" — that quietly destroys the randomization. Decide the seed before looking.
  • Randomizing once for a study run in waves — randomize within each wave (stratify) instead.
  • Confusing random assignment (this tool, for causality) with random sampling (who gets into the study, for generalizability).

Frequently asked questions

How does the randomization work?
A Fisher–Yates shuffle (the standard unbiased shuffle algorithm) reorders your list, then participants are dealt round-robin into the groups. Group sizes never differ by more than one, and every allocation of those sizes is equally likely.
What is the seed for?
Reproducibility. The same participant list with the same seed always produces the identical allocation, so you can record the seed in your protocol or lab notebook and anyone can verify the assignment later. Leave it blank and the tool draws a fresh random seed (and shows you which one it used).
Is this proper randomization for a clinical trial?
It implements simple/complete randomization, which is fine for lab experiments and classroom studies. Regulated clinical trials usually need more: allocation concealment, stratification by site or prognostic factors, and an audit trail — typically via a dedicated randomization service. For stratified assignment, run this tool once per stratum.
Are the names sent to a server?
No. The shuffle runs in your browser; your participant list never leaves the page — worth knowing when the list contains identifiable information.

Related free tools

Sample size for two groupsIndependent samples t-testWhich statistical test?Effect size calculator

Designing the whole experiment?

In Kahubi, the AI helps you plan the design, power the sample, run the survey or experiment, and analyze the results — grounded in your own materials.

Kahubi AI
Hi! Ask me anything you’d normally google mid-study — study design, statistics, literature, writing.

Free accounts include AI chat, literature search, and every tool on this site.

From allocation to analysis in one place

Kahubi runs your surveys and experiments end to end: recruit, collect, clean, analyze with real statistics, and draft the write-up — with the AI grounded in your data.

Try Kahubi freeMore free tools