Support Home Experiment Builder Tutorials Tutorial: Randomiser and Allocator Nodes

Tutorial: Randomiser and Allocator Nodes

Overview

This tutorial demonstrates how to randomly assign participants to different activities (e.g., tasks or questionnaires) using either the Randomiser or Allocator nodes.

Randomiser Node:

If you don't know exactly how many participants you want to complete each activity, but you do know the probability with which participants should be assigned to each activity (e.g., 50% to Task A and 50% to Task B), you can use the Randomiser node.

Setting up the Randomiser node:

  1. Add the relevant participant activities to your experiment.
  2. Add a Randomiser node before these participant activities.
  3. Set up the randomisation ratios for each group.
  4. Choose between the two available Randomisation Modes: Balanced or Random.
  5. Connect the Randomiser node to the experiment flow.

Allocator Node:

If you know exactly how many participants you want to randomly assign to each activity (e.g., 10 participants assigned to Task A and 10 participants assigned to Task B), you can use the Allocator node.

Setting up the Allocator node:

  1. Add the relevant participant activities to your experiment.
  2. Add an Allocator node before these participant activities.
  3. Set up the 'Max number of participants' for each group's allocations, e.g., enter the number of participants you want to allocate to that group.
  4. An additional white dot will appear on the Allocator node for the Overflow group. This group includes participants who arrive after all other groups reach their maximum capacity. You will need to decide how to handle these participants. For example, you could connect the Overflow group to a Reject node to exclude them from the experiment.

Open Randomiser Node Example in Gorilla

Open Allocator Node Example in Gorilla



Info

This tutorial shows the new Experiment Builder interface. If you’re still using the classic interface, it will appear slightly different.

Pro Tip

Balanced or Random Randomisation?

To maximise the chances of ending up with balanced numbers across your groups, use Balanced mode. To assign participants to groups completely at random, use Random mode.

Participant dropout

The Randomiser node is not attrition-sensitive, meaning it does not account for participants dropping out of the experiment. If you need to account for participant dropout, you can use the Allocator node instead, which is attrition sensitive.

If you're running an experiment and notice that your conditions are imbalanced, check out our Troubleshooting Guide for further advice.