This tutorial will introduce you to the basics of experiments, including:
An experiment is a different type of study from an observational study. We'll talk about the differences in a second, but the researchers are allowed to impose treatments on the participants. Treatments are administered and response to those treatments is measured. Because the researchers are the ones implementing the treatments and measuring the response, a cause-and-effect relationship between variables can be determined.
Conversely, in an observational study, the researcher observes the individuals but does not administer treatment. The researcher just has to allow what would normally happen to happen. And again, they can record variables of interest, but not affect it in any way. The researcher is not necessarily an active participant in study, other than observing and recording.
An experiment is a lot more active on the part of the researcher. They are creating the differences between the two groups, then determining whether or not there is a cause-and-effect relationship.
When talking about experiments, there is some very common terminology that you should be aware of. For example, subjects and participants are used interchangeably, and describes people involved in an experiment.
If animals or things are used in an experiment, you’re going to call them experimental units. While it may seem a bit cold to some people, it’s universal terminology in the field of experiments.
On the occasion that you have a study that you'd like to do but you can't perform it due to ethical or practical concerns, or it takes too much time. or it takes too much money, sometimes you can avoid those concerns or get around them by doing an observational study.
When trying to determine if cigarette smoking causes cancer, several observational studies have been conducted, but never a true experiment. Why would that be?
Well, it would be unethical to break people into groups and administer cigarettes to a group of people when trying to determine if it causes terminal illness.
The same goes for alcohol consumption. There are many instances where an experiment would involve privacy issues and it’s unlikely people will divulge that information.
There are certain times when an observational study will be preferred over an experiment due to things like time, money, and privacy.
In an experiment, a researcher can directly influence the subjects by applying treatments. Terminology such as subjects and participants is important to know since it identifies individuals directly involved in the experiment. Animals may be directly involved in an experiment, but they are referred to as experimental units rather than subjects or participants.
Sometimes an experiment may be unethical, expensive, or too lengthy. In those cases, observational studies may be used, which allows a researcher to study occurrences in a natural setting without administering treatment of any kind.
Good luck!
Source: This work is adapted from sophia author jonathan Osters.
