Once researchers have determined what they want to research, they will need to decide which experimental design they are going to use. What this means is that they will need to decide how their participants are allocated to the conditions in the experiment.
In order to be able to explain this better, let’s use the following scenario of some imaginary research to see how the design might develop:
Researchers want to find out if children can complete a jigsaw puzzle faster when in a silent classroom or in a noisy playground.
If they decide that they are going to ask all of the children to complete the jigsaw both in the classroom and in the playground, this means that this is a repeated measures design as the children have been subject to both of the conditions – classroom and playground.
The researchers may decide that they are going to randomly allocate the children to either the playground or the classroom condition – this would be an independent measures design as the children are only going to complete the task once, in one of the conditions. In this instance, each group will have a different set of children.
It gets a bit more complicated if the researchers decide that they want to do a matched pairs design. Here, each child would be matched with another child by a factor that is significant to the research, so in this case, it could be the intelligence level of the child or the closeness of their age. One of each pair in the ‘match’ would go to the playground condition and the other one to the classroom condition and their results would be assessed compared to each other.
Finding the appropriate design for an experiment can be very difficult because each has its relative strengths and weaknesses, which are outlined in the table below:
| Type of design | Strengths of this design | Weaknesses of this design |
| Independent measures | · No order effects – participants tend not to get tired or bored because they are only doing one of the conditions. They also don’t ‘learn’ how to complete the task as they are only doing it once. | · Does not account for individual differences because different participants take part in different tasks. So, in the previous example, the children completing the jigsaw in the classroom may have done it faster simply because they are better at jigsaws, not because it is quieter.· Requires more participants than repeated measures and this can become expensive and time consuming. |
| Repeated measures | ·Fewer participants are needed as each does all of the conditions.·Researchers are comparing the same participants in all of the conditions, so it is easier to see if results are caused by the independent variable than by individual differences. | ·Order effects – participants can get tired, bored or they can learn how to complete the task before doing it again in another condition. |
| Matched pairs | · Reduces the possibility of individual differences affecting the outcome, as participants have been paired on similar ability.· Avoids order effects as the participants are only doing one of the conditions. | · Very time consuming to organise.· It is impossible to match people exactly – even identical twins have some differences.· If one participant withdraws from the research, this means that the pair has to leave too and so two sets of results are lost. |
Activity – Which research design?
Given the example of the children and the jigsaw puzzle, which research design would you choose and why? It often helps with questions like this to think why you wouldn’t choose a design, as this helps you to narrow it down.
