ABSTRACT
We cannot see (i.e., directly observe) a person’s confi dence, likes, beliefs, preferences, or attitudes; he must tell us in some way. A variety of devices, such as question naires, interviews, and choice situations, may be used to determine subject’s cognitive behavior, but the usual prob lem is that he knows he is giving such information. If the subject realizes that he is in some sort of evaluative situa tion, such as a survey or an experiment, his answers may be very biased. People do not want to reveal personal beliefs and feelings that they feel might make them appear unintelligent, foolish, or bigoted. To overcome such biases experimenters frequently concentrate on designing situa tions to keep subjects unaware of the exact nature of the information being sought. Sometimes, for example, experi menters are interested in only a few answers, but they bury the questions in long questionnaires. Sometimes informa tion is sought under false pretenses.
