Observations

Specifying the Behavior of Interest

The first step in collecting observations is to specify the behavior(s) of interest. This often consists of formulating an operational definition, or precisely describing what constitutes an occurrence of each type of behavior. For instance, physical aggression may be operationally defined as hitting, kicking, or biting another person. Thus, when any of these behaviors occur, the researcher would record an instance of physical aggression. Researchers often create coding manuals, which include operational definitions and examples of the behaviors of interest, to use as a reference guide when observing complex behaviors.

Recording Observations

Researchers use different methods for recording observations dependent on the behavior of interest and the setting. One dimension that may differ is whether observations are made live or while watching a recording of the episode. Researchers often choose to record simple behaviors live and in real time as they occur. In situations characterized by complexity, researchers often elect to make a DVD/video recording of the episode. This gives researchers more flexibility in recording a variety of behaviors, most notably the ability to progress at their own speed, or to watch behaviors again if needed.

Similarly, researchers may choose to record behavior by making hand tallies or using computers. If the behavior interest is simple, researchers may choose to record each time a behavior occurs. Many researchers, however, choose to use a computer to facilitate observational research. Several computer programs are available for recording observations. Computer entry allows for exact timing of behavior so that researchers can determine time lags between particular instances of behavior.

Another choice in recording behavior is whether to use time or event sampling. Time sampling involves dividing the observational time session into short time periods and recording any occurrences of the behavior of interest. For instance, researchers studying classroom participation might divide a 1-hour class into twelve 5-minute intervals. They could then record if students participated in each interval and then determine the percentage of intervals that included student participation. A second option is to use event sampling, recording each behavior of interest as it occurs. Researchers using the event sampling technique in the classroom participation example would record each time a student participated in class and would ultimately calculate the frequency of student participation across the class period.

...