Observational research is not hypothesis-driven, but simply begins by observing people or animals “doing stuff,” as Chicago researcher Walt Dickie describes it. Initially, a researcher may use observation to determine one or more research questions for further refined study. Because the researcher is observing subjects at random and observing whatever behavior occurs, he is freed of the constraints of typical structured, hypothesis-driven research. As an example, a researcher might observe a group of smokers and non-smokers over a period of years. While not interfering with their habit or lack thereof, he will be able to make observations and contrasting outcomes for smokers and nonsmokers.
Characteristic disadvantages of observational research include researcher bias, a disparity between subjects and a study group that does not reflect the general population. Researcher bias can unintentionally slip into the study or that the sample is not a typical representation of the population, making any generalization from the results suspect. The reliability of observational research rests on the likelihood that it can be replicated and it is critical that the research can be applied to other groups in other times and places – generalized – and still arrive at the same result. If these goals cannot be achieved, then the research applies only to the specific population originally observed.
In conducting direct or reactive observations, the subjects often are aware that they are being watched. Here the concern is that the subjects will react to the observer and consequently change their behavior, thereby distorting their actions. While this is a problem, some researchers feel that reactive behavior simply reveals other behavioral aspects that shed new and different light on the subject. Additionally, behavior that is not typical of the subject may disappear after a time as the subject returns to his or her normal patterns.
Two distinct types of direct observation are prevalent. The first is continuous monitoring, which entails observing and recording as much of the subjects’ behavior as possible. This type of direct observation is often employed in group settings, for example in a company’s assessment of people’s performance. However, continuous monitoring can easily lead to the Hawthorne Effect, which sees marked productivity increases among workers who know that they are being observed.
The second type of direct observation is time allocation research, which consists of selecting a random place and time and recording people from when they first arrive at the site until they spot the observer. While time allocation research poses several problems, it is useful in observing how people spend their time. For example, a researcher might watch adults with their children at a park to determine how much time they spend actually playing with the children or watching people in a cafeteria to see how they eat and how long it takes them to finish. However, getting a good representative sample with time allocation requires spending considerable time observing and, too, questions arise as to how often you observe, at what times and where.