Observational studies are experiments which are based on analyzing pre-existing variables which a researcher cannot manipulate. These studies are structured beyond the control of a researcher because investigators research already existing variables and do not assign treatments to the subjects of study. Rather, the subjects naturally follow particular treatments.
For example, pretend you are doing a study about the influence of running on the blood pressure ratings of adults 40 years old and over. If this study required 50 individuals, the researchers would find 25 individuals aged 40 or over who have been running regularly over the past 5 years and 25 individuals who have not been running. Then blood pressure ratings would be taken from these two groups, various analyses would be conducted and conclusions could be drawn.
Observational studies are based completely on natural conditions. No manipulations are employed on the pre-existing, real-life conditions of the variables of interest. Thus, observational studies are more useful in gathering information about a certain subject or topic of interest, rather than studying a cause and effect relationship. An observational study can be thought of as a survey almost.
Furthermore, when doing observational studies it is important to try and eliminate bias. In many cases, it is impossible to sample the entire population and finding a representative sample is important for drawing generalized conclusions. This is a similarity that observational studies share with multiple types of experiments and studies in the subject of statistics.
Image Credit: Wikipedia Commons