Instrument Design: Decisions and Procedures

摘要

While questionnaires must be custom-built to the specifications of given survey situations, there are guidelines that inform instrument design. Even so there are still unresolved issues and basic methodological questions about the instrument design process. A few issues are identified in this conclusion that stem from issues that arose previously in this chapter. While this chapter attempts to accommodate both self-administered and interviewer-mediated questionnaires, it is apparent that the task of questionnaire construction is very much tied to the survey mode. Furthermore, research issues related to questionnaire construction are also mode-specific. For example, with Internet surveys, researchers are attempting to understand principles that underlie visual processing of information, such as the effects of relative size, colour, brightness, or the location of information on questionnaire pages (Dillman, 2002). On the other hand, with CATI questionnaires, researchers are more concerned with interviewer performance, because it is a crucial part of the data quality (Czada and Blair, 1996). Researchers are focusing on how to reduce the level of interviewer effort. Future research should focus on identifying these mode-specific instrument problems and their potential solutions. This latter idea is important within the context of mixed-mode designs, and underlies the additional level of effort required for their application. The instrument designer needs to be aware of the decisions and problems associated with each individual mode and to design instruments accordingly. Ensuring the conceptual equivalence of the mixed-mode instruments is challenging at best. Also, empirical research needs to be done on evaluating the potential mode affects specific to personal travel surveys and their impact on data quality.