The Evaluation of Physical Settings: To Measure Attitudes, Behavior, or Both?
A recurring problem in the evaluation of physical settings is that the dependent variables (what is measured) are usually insufficient to provide valid and generalizable information. Numerous evaluation studies are reviewed, and shown to utilize either attitudinal (e.g., satisfaction) or behavioral (e.g., use) measures. This paper explicates the threats to validity in using either attitudes or behavior individually. Further, both behavioral and attitudinal measures are defined in the context of evaluating designs, and their strengths and weaknesses are presented. Finally, a multi-method technique of measurement is proposed, where behavioral measures are used to validate attitudinal measures. Only in this way can sound and valid data be obtained. Many evaluations of physical settings fail to provide data of value to the designer for a simple reason:. the dependent variables, that is, those things that are measured about human behavior in the setting, are often insufficient to assess the designs effect. The purpose of this paper is to show how both attitudes and behavior must be measured in order to provide useful data.