ABSTRACT
Social scientists have long been aware that minor changes in question wording, question format, or question order may profoundly affect the answers that research participants provide. In the survey literature, these influences of the research instrument are usually referred to as response effects or context effects. Over the last decade, researchers have made considerable progress in understanding the cognitive and communicative processes underlying the emergence of context effects, and this chapter provides an introduction to what has been learned. It draws on the collaborative work of cognitive psychologists, social psychologists, and survey methodologists, which was initiated in the early 1980s and has grown into an active field of basic and applied cognitive research (for a comprehensive review, see Sudman, Bradburn, & Schwarz, 1996; for research examples, see the contributions in Hippler, Schwarz, & Sudman, 1987; Jabine, Straf, Tanur, & Tourangeau, 1984; Jobe & Loftus, 1991; Schwarz & Sudman, 1992,1994,1996; Sirken, Hermann, Schechter, Schwarz, Tanur, & Tourangeau, in press; Tanur, 1992). Initially, this research was based on general information processing models developed in cognitive psychology (cf. Lachman, Lachman, & Butterfield, 1979). As the work progressed, however, it became increasingly clear that the information processing paradigm could only capture some of the key aspects of survey interviews. Most importantly, mainstream information processing models share an exclusive focus on individual thought processes. As many critics noted, this concentration on individuals as isolated information processors fostered a neglect of the social context in which human judgment occurs (cf. Forgas, 1981). The survey interview, however, is best considered as an ongoing conversation in which respondents conduct their own share of thinking and question answering in a specific social and conversational context. Hence, conceptualizations of the question-answering process need to consider conversational as well as cognitive processes and need to pay close attention to the complex interplay of social communication and individual thought processes (for more detailed discussions, see Clark & Schober, 1992; Schober, in press; Schwarz, 1994, 1996; Strack, 1992; Strack & Schwarz, 1992).
