What are the design types in experiments?

Travis DixonInternal Assessment (IB), Research Methodology

Avoid making the common mistake of misunderstanding the term "research design."

Choosing the best research design for your experiment is an important part of the planning process. When conducting an experiment for the IB Psychology IA, you must think very carefully about which design is best for your purposes. 

Experimental Designs (MP, IS, RM)

There are three design types of experiments:

  • Independent Samples
  • Repeated Measures
  • Matched Pairs

Independent Samples is when the sample is divided into different groups. This means that not everyone will experience all conditions in the experiment.

For example, in an experiment that tested the effects of cortisol on memory, Buchanan and Lovallo randomly allocated their participants into two conditions. One group received a pill containing cortisol and the other group took a placebo. They did a memory experiment on them to see how the cortisol influenced their memories. The key benefit is that it controls for order effects. 

Repeated Measures is when the treatment is repeated for all participants. In other words, the participants experience all conditions of the experiment.

An example of a repeated measures design was Passamonti et al.’s experiment on reducing serotonin and its effects on the prefrontal cortex. In this study, participants came into the lab and consumed a drink that either reduced their serotonin levels (treatment condition) or had no effect whatsoever (control condition). They were then put in a lab and exposed to different images of faces to see how their prefrontal cortex would react. The visits to the lab were separated by one week to control for order effects. Feder et al.’s study on ketamine and PTSD is another example.

Matched Pairs design is quite similar to independent samples in one regard: the sample is divided. Participants are paired up with another participant based on some criteria and then they are separated so they are in a different group from their partner.

For example, Maguire matched her participants in the study of London Bus drivers and Taxi drivers to control for confounding variables like handedness, years of driving experience, age, education, etc.

In his  bobo doll experiments to see if watching violence can cause children to be violent, Bandura used a matched pairs design by gathering lots of data on the children’s (i.e. participants’) pre-existing aggressive tendencies. They asked their parents and teachers for information and then used a scoring system to rate their aggressiveness. The children were then paired (not physically) with another child who had a similar rating and then put in different groups.

Notes: Repeated measures is sometimes called “within subjects design” and independent samples is sometimes called “between subjects design” or “independent groups.”