Data collection techniques: observation.

Use of observation as a measurement procedure, assigning numerals to human behavioral acts, is discussed. Observation has important advantages which makes it best suited for certain kinds of studies, and some limitations which preclude its use in others. The central problems in the use of observation are: (1) the effect of the observer on the observed, which is usually not severe and can be minimized; (2) observer inference, which is a crucial strength and a crucial weakness; and (3) the unit of behavior to be used, which involves the molar-molecular problem. The considerations in planning both unstructured and structured observation studies are discussed, including what to observe, how to record it, how to maximize validity and reliability, and how to handle the relationship between the observer and the observed. Behavior is usually sampled using event sampling or time sampling. The uses and weaknesses of rating scales to assess perceived behavior are summarized.