Context Effects as an Illustration of Response Uncertainty –A Cautionary Tale

Many years of experience make it difficult to believe that survey answers are mostly unambiguous. For example, by going through accessible studies, a lot of problems concerning memory can be identified in the field of living conditions and in other domains. Consequently, it may be claimed that facts are seldom stored in an easily accessible manner. It is probably not the case that people – when answering questions – respond promptly and automatically. Rather, some sort of thinking activity is initiated. Accordingly, the answers may be dependent on the occasion when the questions are asked, which background information is given, how familiar the respondents are with the phenomenon in question, how much time there is for reflection, and so on. We have had the opportunity to compare answers to the same questions asked in different contexts. The results clearly indicate the existence of context effects. At the same time, the results give some clues to how these effects can be interpreted, and even avoided. Problems may be minimized if questions are more precisely formulated, if their focus is on strictly limited aspects, and if they are only about things to which people have devoted a lot of interest and thought.