T HE RANDOMIZED response technique first proposed by Warner (1965) has now undergone more than a decade of development. Prompting this effort has been the desire to obtain more reliable information when dealing with sensitive issues on surveys. As researchers in this area are fully aware, the two most frequently encountered problems when asking sensitive questions are refusals to respond and intentionally misleading responses designed to conceal socially undesirable behavior and attitudes. The Warner technique uses two mutually exclusive statements dealing with a sensitive issue. Greater respondent cooperation is elicited through use of a device which selects, by chance, one of two statements to which the respondent is to reply, without revealing this selection to the interviewer. As an example, consider the following statements: I am a member off Group A. I am not a member of Group A. The respondent is to answer "Yes" or "No" to the statement selected by the randomizing device. The interviewer is not told which statement was selected and hence does not know to which statement the reply referred. This technique was subsequently modified by Abul-Ela et al. (1967), who provided for obtaining information on more than one sensitive issue through use of a randomizing device. This method could deal with three or more categories, at least one of which had to be of a nonsensitive nature. In addition to obtaining information on a greater number of sensitive topics, it was felt that a respondent might be more willing to reply