Using scale reduction techniques for improved quality of survey information

Surveys have been used since time immemorial to collect raw data for the production of information in a variety of contexts. The main tool for conducting surveys has been questionnaires, mainly because they are relatively easy to administer and can be used to efficiently gather sufficient data at a relatively low cost. The ease of being able to administer questionnaires using Internet protocols has motivated many researchers to integrate questionnaires into their research design over the past decade. Additionally it is common practice for surveys to be used by managers as a source of information for decision making. Given the commonplace use of survey design in both the academic and practitioner environments, it is important that sound survey design techniques are adhered to. The authors undertook an analysis of survey design techniques in articles published over a two-year period in the South African Journal of Information Management, and of a sample of completed post-graduate research. It was found that an area that requires greater attention concerns scale development. In particular it was found that researchers are generally not implementing pilot tests of questionnaires in order to improve the scale reliability thereof. The primary objective of this article, therefore, is to highlight this problem and to advocate improved practice in survey design. In light of the foregoing, the use of statistical techniques to refine questionnaire instruments through a piloting process is described. The article provides empirical evidence to demonstrate how, by applying item-analysis techniques, questionnaires can be refined and prepared through a pilot process. Four case studies are used to show that had item analysis been applied by the researchers, the length of the scales would have been reduced, and the reliability of the results may have been improved. This therefore has implications for improving the quality of information gained from the survey data.