Random assignment generator
Assign people to the groups of a study at random, in the order they join. Balanced assignment keeps the groups the same size. Simple random assignment gives each person an equal chance of each group, so the sizes can differ.
The generator uses the code the ConductExperience study builder assigns people with, and runs in your browser.
Assign people to groups
Type the groups and how many people, then Assign.
Balanced or random
| Assignment | What it does |
|---|---|
| Balanced | Each new person goes to a group with the fewest people so far, picked at random when groups tie. The groups never differ by more than one person. It is block randomization with blocks the size of the number of groups. |
| Random | Each person has an equal chance of each group, whatever came before. Also called simple random assignment. Over many people the groups come out about even, but a small study can come out uneven. |
How often simple random assignment splits two groups 60 to 40 or worse:
| People | Chance |
|---|---|
| 20 | 50% |
| 40 | 27% |
| 100 | 6% |
Random assignment in ConductExperience
- On the study builder's Groups and order tab, add up to 12 groups and choose Balanced, Random or Weighted.
- Each person is assigned when they join the study. Balanced counts the people in each group at that moment, so the groups stay even as people join.
- A weight gives a group a bigger share, such as 2 to 1, and a limit closes a group when it is full.
- The Participants table and participants.csv show each person's group.
Limits
- The generator takes 2 to 12 groups, the most a study takes, and up to 1,000 people.
- In a balanced list, the last person of each round goes to the one group left. Keep the list from the people who enroll participants, so no one can choose who gets which group.
- The list comes from your browser's own random source, and nothing is sent anywhere.
Related
Between and within subjects
When to compare groups of people, and when each person does every condition.
Counterbalancing
Balanced orders for conditions, with a Latin square generator.
Study builder
Groups, sections and parts in fixed, shuffled or counterbalanced orders.
Online studies
One link for a panel or a mailing list.
Planning a study?
Start one now, or tell us what you want to measure.