THE INFORMATIVE FUNCTIONS OF RESEARCH PROCEDURES - BIAS AND THE LOGIC OF CONVERSATION

Conversational rules of everyday communication are applied to the interaction between experimenters and subjects. According to these rules, contributions to a communication should be informative, relevant, true, and unambiguous. It is assumed that subjects determine the pragmatic meaning of instructions and questions on the basis of these rules and the provided context. In contrast to most natural settings, standardized experimental procedures rarely allow for an interactive determination of pragmatic meaning and often preclude feedback as a corrective device. As a consequence, subjects are required to rely heavily on general rules, and even subtle cues may become informationally loaded. The information extracted from context cues may often not be intended by the experimenter. Thus subjects may infer more than they are supposed to, resulting in discrepancies between the experimenter's intended and subjects' inferred meaning of the instructions. If researchers are not sensitive to the information provided by verbal and non-verbal context cues, their interpretation of research results may be based on biased data. Evidence from different research domains is reported to support the presented assumptions and their implications for bias avoiding strategies are discussed.