18  Assign Experimental Units

Assignment of individuals to appropriate groups is an essential part of experimental design. The NC3Rs EDA provides two nodes to help us describe how this assignment is performed in an experiment:

• Group, which describes a group of individuals
• Allocation, which describes how individuals are taken from one group, and allocated to another

Note

For example, if we have an initial pool of all our experimental subjects, this will be one Group node. Should we want to assign evert individual from this group to one of two new groups, one receiving the treatment and the other receiving a control, then we would need two new Group nodes (one for control, one for treatment), and an Allocation node to represent our assignment of individuals from one group to another.

We will carry out this process for our experimental design:

• all individuals initially belong to a single pool
• individuals are assigned from this pool either to a treatment or a control group
• a complete randomisation allocation process is used to assign the subjects to these groups

18.1 Create the experimental subject group structure

18.1.1 Add the initial pool

1. Click and hold the Group node in the sidebar and drag-and-drop it to the canvas (Figure 18.1). This will represent the complete pool of animals involved in the experiment.

Figure 18.1: A new Group node, added to the EDA canvas

2. Double click on the Group node to change the label. Call this Group “Pool Group” (Figure 18.2)

Figure 18.2: The Group node representing the complete pool of animals, with the label field as highlighted by double-clicking.

18.1.2 Add an Allocation node

1. The icon panel to the right of the Group node has more options than that for the Experiment node (Figure 18.3). Move the pointer over the blue circle and then click on it, to link a new Allocation node (Figure 18.4).

Figure 18.3: The icon panel to the right of the Group node has more options than that for the Experiment node, as more kinds of node can be linked from a Group.

Figure 18.4: A new Allocation node, linked to the pool of individuals that will be allocated.

2. Open the Allocation node properties and set the label to read “Complete randomisation”, and the randomisation strategy (a drop-down) to be “complete randomisation” (Figure 18.5). Click on the Close button to return to the canvas.

Figure 18.5: Setting properties for the Allocation node. The label is changed by typing directly, and the choice of randomisation scheme is made in a drop-down menu.

18.1.3 Add Treatment and Control groups

1. Click the blue square (Group node) icon to the right of the Allocation group twice, to generate two new Group nodes (Figure 18.6).

Figure 18.6: Two new groups are generated by the randomisation, one for control individuals receiving no drug, and one for the treatment group, who receive the drug.

2. Click and hold the mouse on each group, and move it until the nodes are positioned separately from each other (Figure 18.7).

Figure 18.7: By clicking on a group and holding the mouse button down, you can move a node into a more aesthetically-pleasing location.

3. Double-click on each new group and name as “Control” and “Treatment” accordingly (Figure 18.8)

Figure 18.8: It can be helpdul for understanding to name each group according to its role in the experiment. Here the two post-randomisation groups are called “Control” and “Treatment”.

18.2 Summary

By following the process above we have defined our initial pool of experimental subjects, and described how they are randomly assigned to either a control or a treatment group. In the next steps, we will add the nodes that describe our experimental interventions, and the measurements we intend to record.