THE RESPONDENT IS POSED QUESTIONS IN PAIRS - ONE SENSITIVE AND ONE INNOCUOUS. WHICH QUESTION THE RESPONDENT ANSWERS IS DETERMINED BY A RANDOMIZING TECHNIQUE, SUCH AS THE TOSS OF A COIN, OF WHICH ONLY HE IS AWARE OF THE OUTCOME. THE RESEARCHER USES PROBABILITY THEORY TO DETERMINE AGGREGATE RESPONSES TO THE SENSITIVE QUESTIONS. THE RANDOMIZED RESEARCH DESIGN ELIMINATES RELUCTANCE TO ANSWER INCRIMINATING QUESTIONS BY ASSURING RESPONDENTS THAT THEIR ANSWERS CANNOT BE ATTRIBUTED TO SENSITIVE QUESTIONS. IN THIS ARTICLE THE AUTHOR REVIEWS THE TECHNIQUE AND USES IT TO INVESTIGATE SHOPLIFTING, DESTRUCTION, PRICE ALTERING, AND OTHER DEVIANT CONSUMER BEHAVIOR THAT OCCURS IN RETAIL STORES. (AUTHOR ABSTRACT MODIFIED)