What is the difference between experimental and correlational research designs?
Correlational designs and experimental designs yield information which can look similar but are fundamentally and drastically different.
A correlation show the relationship between two factors in terms of how often they are associated with each other. For example, there is probably a very high correlation between height and shoe size. Taller people tend to have larger feet than short people. Or there may be no correlation between two factors, such as the amount of cereal people buy and their shoe size.
The key to understanding correlation is this: it shows whether there is a relationship between two factors, but it says nothing about cause, whether one factor causes the other. For instance, we can safely say that all bank robbers have drunk water, so there is a high correlation there. But drinking water does not cause a person to rob a bank.
Experimental research designs focus on finding out how one variable affects another; in other words, because of its design, it can determine cause. Experimental research uses the scientific method which controls all variables except one to see how that one variable affects something. For example, suppose you want to find out if watching violence causes children to act violently, you would set up two groups of children: the control group and the experimental group. The control group would experience everything the experimental group except for one thing: the experimental group would see an adult punching a toy while the control group would watch something neutral. Children's behavior would then be observed. In fact, Albert Bandura carried out this research and found that observing violence caused children to act violently.
The essential difference, then, between correlational design and experimental design lies in the determination of cause. Correlational design cannot determine the cause of the relationship between two factors. Experimental design, because of its control over variables, can determine cause.