Interaction, Reported and Observed: The Case of One Local Union

Abstract

The observation is neither profound nor original that, over the past two to three decades, behavioral scientists have placed increasingly heavy reliance on the questionnaire survey as a principal means of data collection. The reasons for this are obvious and compelling. The method is relatively inexpensive; once the questionnaire is constructed, no great skills are involved in its administration; by means of questionnaires, data may be rapidly, efficiently, and uniformly gathered from large and representative populations. And, most important, the combination of the standard stimulus, accurate sampling procedures, and ever more sophisticated techniques for analysis of quantitative data, offers behavioral scientists their most reasonable approximation, for large numbers of respondents, to the classic experimental model, without which, no researcher may lay claim to hard and fast "proof" of any scientific proposition.