When designing an experiment, we need to consider the experiments design. This is the way in which participants are organised to allow us to conduct our study and compare our results. There are three main ones we use in psychology, each with their advantages and disadvantages. The three designs are
- Independent group designs
- Repeated measures
- Matched pair design
An independent group design involves groups of participants only experiencing one of the conditions being tested. For example, if we have two groups A and B and we are investigating the effects on chocolate and memory. Half the participants would be in group A who eat the chocolate and group B who eat nothing. Group B would be our control group and Group A would be in the experimental condition. We would then gather our results and then compare the results between the groups.
A concern with this experimental design is that each group is simply not the same. By chance everyone in group A may have a better memory. This means when we compare our results we may end up with a false conclusion. Therefore, it is important for repeat experiments, compare with other studies and peer review.
However, an advantage of using an independent group design is it removes the chance of order effects affecting the results.
A repeated measures design involves participants taking part in all conditions of the experiment. So, they would have their memory tested both after eating chocolate and another time when they have not consumed the chocolate. We can then compare all the participants results for condition 1 and 2.
A repeated measures design has an advantage as you are comparing “like for like”. Therefore, in theory, we can get a fair comparison between the different experimental and control conditions. It also requires less participants than an independent groups design.
Repeated measures suffer from the issue of order effects. Results could be affected as they have repeated the task and got better, boredom or they may have worked out the aim of the experiment and this changes their behaviour, demand characteristics. To combat this, we can use counterbalancing. Half of the participants start off in group A and the other in group B and then they swap. This way everyone participated in each condition but in a different order. This may not be suitable for experiments that investigate the use of drugs and therapies.
A matches pairs design involves trying to match participants as close as we can considering the variable being tested. Twin studies can be a great tool for this but there may not be enough, or they may not be suitable. So, with our chocolate and memory experiment we may decide to match participants on IQ, which is said to be a good indicator for memory. We then put on of the participants in group A and the other in group B. Again, we can then compare the results.
As participants only take part in one condition, order effects and demand characteristics are less of an issue and they participants are more like each other than an individual groups design.
Sometimes we may even use twins for a matched pairs design and although they would be very similar, we can never get an exact match. Another limitation of this design is the time and cost it takes to organise and select participants using matched pairs.