Designing an Observation Study
There are many different ways to design an observation study, depending on the objective of your study, the type of data you are trying to collect, and the resources you have available for your study. Following are five different features that you should consider when designing the ideal observation study for your project:
Natural vs. Contrived Settings:
Conducting the study in a natural setting essentially means that you are simply observing your subjects in their "real life" environments. Because you have no way of influencing what your subjects are doing, this method can be time consuming to gather the information that you are specifically trying to obtain for your project. Alternatively, the data that is collected in a natural setting does have more accuracy in reflecting "real life" behavior rather than "contrived" behavior.
A contrived setting is one where the specific situation being studied is created by the observer. The contrived setting offers you, the observer, greater control over the gathering of data and specifically will enable you to gather the information more quickly and efficiently. However, it may be questionable as to whether or not the data collected does truly reflect a "real life" situation.
Disguised vs. Non-disguised Observation:
When subjects do not know they are being observed, this is called a disguised observation. Subjects in disguised observations tend to act more naturally and the data collected tends to reflect their true reactions. The primary concern with disguised observation is the ethical concern over recording behavioral information that would normally be private or not voluntarily revealed to a researcher. However, if you are simply observing a subject's behavior in a public setting then by definition, their behavior is no longer private.
When subjects know they are being observed, this is called a non-disguised observation. Using the non-disguised observation technique alone alleviates ethical concerns, however, since the subjects are aware that they are being watched, the advantages of using the observational technique are neutralized and a survey technique would be equally effective. There is one exception: the non-disguised approach offers the advantage of allowing the researcher to follow up the observations with a questionnaire in order to get deeper information about a subject's behavior.
Human vs. Mechanical Observation:
Human observation is self explanatory, using human observers to collect data in the study. Mechanical observation involves using various types of machines to collect the data, which is then interpreted by researchers. With continuing improvements in technology, there are many "mechanical" ways of capturing data in observation studies, however, these new "gadgets" tend to be extremely expensive. The most commonly used and least expensive means of mechanically gathering data in an observation study is a video camera. A video camera offers a much more precise means of collecting data than what can simply be recorded by a human observer.
Direct vs. Indirect Observation:
Direct observations involve looking at the actual behavior or occurrence rather than a result of that occurrence, which would be an indirect observation. For example, if you were interested in seeing how much candy was purchased by a particular neighborhood, you could gather the information in one of the two following ways:
Direct observation: observe customers in a store and count how many bags of candy they purchase.
Indirect observation: look through trash cans on garbage day to see how many empty candy bags are in each trash bin
Indirect observation tends to be used when the data cannot be gathered through direct means, or when gathering the data through direct observation tends to be too expensive.
Structured vs. Non-structured Observation:
Structured observations are made when the data that is being collected can be organized into clear categories or groups so that the observer can record the data by simply marking off or checking a category on an observation form. Non-structured observations are not looking for specific facts or actions, but rather are capturing everything that occures. For example, if the US Postal Service were interested in knowing the gender and racial profile of the people that use a particular post office, they could post an observer at the front door and simply record the data as people entered the post office. This would be a structured observation, where the observer would simply be marking off boxes on an observation form. However, if the US Postal Service were interested in knowing the general level of satisfaction with service in a particular post office, they could post an observer in that office to capture more general data such as the length of the line during various times of day, the general change in customer demeanor as the line grows longer, the change in customer demeanor when there are one, two, or three windows open, etc....