The fabrication of data is seemingly the most serious form of fraud since it involves the outright production of fictitious information. In the natural and experimental sciences, fraud may be detected through replication, but in the social sciences alternative findings do not necessarily suggest fraud. In some types of survey research involving studies of self-reported deviant behavior, fraud has occurred through the fabrication of data by dishonest interviewers. In an attempt to detect such data fabrication, an experiment was performed comparing valid and fictitious data, resulting in statistical differences between the two types. (Author abstract)