True experiments tend to be difficult to set up and complete properly, but in many cases they are relatively straightforward to analyze and interpret once the results are collected.
They can, however, become fairly complex in cases where more than one treatment or treatment level are being tested, and there can also be the temptation to do what is called "data fishing", or post hoc analysis. For now, tuck away in the back of your mind that there are ways to do post hoc analysis and ways - not - to do this. You can learn more about post hoc analyses once you learn the basics.
The designs discussed starting on this page are the most common experimental designs. There are others, and experiments can actually be constructed in "building block" or "factorial" fashion, but this is a rather complex enterprise. Trust me on this one. If you want to learn more about factorial design, here is a reference.
First and foremost, learn to think of the people who participate in your experiment, control group(s) as well as experimental group(s) as representatives. They represent persons who may receive the new treatment or procedure in the future, if your experiment yields promising results. Persons in the control group as well as the experimental group must be equally likely candidates, and, to be really correct, the persons who participate in your experiment should be drawn randomly from all persons who could conceivably be candidates for the new treatment or procedure. (This last condition is rarely feasible, but you need to be aware of it.) Drawing from the total population of all possible candidates is random selection. Once a pool of participants is randomly selected, they are randomly assigned to experimental or control group(s).
This design can be diagrammed as follows:
We shall start our discussion of the true experimental designs with the Posttest-Only Control Group Design, as this is the simplest possible form of the true experiment.
This design can be diagrammed as follows:
Randomization is used in the Posttest-Only Control Group design to create a situation of initial statistical equivalence of the experimental and control groups. Under some conditions, however, it is desirable to actually measure certain participant characteristics before commencing the experiment. The simplest means of accomplishing this is to use the Pretest-Posttest Control Group design.
This design can be diagrammed as follows:
The Solomon Four-Group Design is a very tight experimental design, controlling well for both internal and external potential sources of error or ambiguity.