An experimental design is how you test whether one thing actually causes another. You change one factor on purpose (the treatment), keep everything else as steady as you can, and compare a group that got the treatment with a group that did not (the control group). Because you control the conditions, an experiment is the strongest design for claiming cause and effect.
When to Use It
- Your research question is about cause and effect — does X cause a change in Y?
- You can split participants into a treatment group and a control group.
- You can apply the treatment and measure the result fairly.
- You can hold other conditions reasonably constant.
When Not to Use It
- You only want to describe a situation — use a survey.
- It would be unethical or impossible to assign the treatment (you cannot, for example, force students into worse conditions).
- You cannot control the setting, so a true experiment is not realistic. A quasi-experiment may be more honest.
Nigerian Project Example
"The effect of computer-assisted instruction on students' performance in mathematics in [School]." One class is taught with the computer-assisted method (treatment), another with the usual method (control). You give both a test before and after, then compare the gains.
Undergraduate vs Postgraduate
Undergraduate experiments in Nigeria are often quasi-experimental using intact classes, with a pre-test and post-test. Postgraduates are expected to control more variables, justify their design choice against threats to validity, and use more advanced analysis (such as ANCOVA) to account for differences between groups at the start.
Common Mistakes
- Calling a study experimental when there was no control group at all.
- Ignoring obvious factors that could explain the result instead of the treatment.
- Claiming true random assignment when you used existing classes (it is quasi-experimental).
- Too small a sample for the difference to mean anything.
Document your design and controls in your methodology — see how to write Chapter Three and the full research methodology guide. Project Lab can help you set up a clean treatment-and-control structure.
Frequently Asked Questions
What is the difference between experimental and quasi-experimental?
In a true experiment you randomly assign participants to groups. In a quasi-experiment you use existing groups (like intact classes) because random assignment is not possible. Most Nigerian classroom studies are quasi-experimental — and that is fine if you say so.
Do I always need a control group?
For a true experiment, yes — the control group is what lets you say the treatment, not chance, caused the change. Without a comparison group, you cannot fairly claim cause and effect.