Observational Research

The observation of human and animal behavior has been referred to as the sine qua non of science, and indeed, any research concerning behavior ultimately is based on observation. A more specific term, naturalistic observation, traditionally has referred to a set of research methods wherein the emphasis is on capturing the dynamic or temporal nature of behavior in the environment where it naturally occurs, rather than in a laboratory where it is experimentally induced or manipulated. What is unique about the more general notion of observational research, however, and what has made it so valuable to science is the fact that the process of direct systematic observation (that is, the what, when, where, and how of observation) can be controlled to varying degrees, as necessary, while still permitting behavior to occur naturally and over time. Indeed, the control of what Roger Barker referred to as “the stream of behavior,” in his 1962 book by that title, may range from a simple specification of certain aspects of the context for comparative purposes (e.g., diurnal vs. nocturnal behaviors) to a full experimental design involving the random assignment of participants to strictly specified conditions.

Even the most casual observations have been included among these research methods, but they typically involve, at a minimum, a systematic process of specifying, selecting, and sampling behaviors for observation. The behaviors considered might be maximally inclusive, such as in the case of the ethogram, which attempts to provide a comprehensive description of all of the characteristic behavior patterns of a species, or they might be restricted to a much smaller set of behaviors, such as the social behaviors of jackdaws, as studied by the Nobel Prize—winning ethologist Konrad Lorenz, or the facial expressions of emotion in humans, as studied by the psychologist Paul Ekman. Thus, the versatile set of measurement methods referred to as observational research emphasizes temporally dynamic behaviors as they naturally occur, although the conditions of observation and the breadth of behaviors observed will vary with the research question(s) at hand.

Because of the nature of observational research, it is often better suited to hypothesis generation than to hypothesis testing. When hypothesis testing does occur, it is limited to the study of the relationship(s) between/among behaviors, rather than to the causal links between them, as is the focus of experimental methods with single or limited behavioral observations and fully randomized designs. This entry discusses several aspects of observational research: its origins, the approaches, special considerations, and the future of observational research.