In an experimental study, 3 things need to occur.
Control
- as many aspects of the experiment as possible so that there is an effect
Randomization
- When groups are chosen, subjects need to be randomized so that no biases can occur
Replication
- Even though the groups are random, when researchers repeat an experiment, it should be similar in makeup so that changes from one group to another are easier to detect.
- You should not be changing the experiment when given to other groups