The relative reliability of reported induced abortion obtained through a conventional survey technique (direct questioning [DQ]) as compared with an estimated proportion of women with induced abortion obtained through the randomized response technique [RRT] was tested. Two independent, nationally representative samples were used, one for each approach. A total of 1521 women were interviewd in the DQ sample, providing a 70.1% response rate. In the RRT sample, 1674 women participated in the study, representing a 72.9% response rate; 1044 or 62.4% of these women provided usable answers. There were substantial differentials in the successful use among the subcategories of independent variables. Of the 1044 women in the RRT sample, 33.1% had at least 1 induced abortion during their reproductive lives.sbd.a much higher rate than the 13.9% obtained from the DQ sample. The factors found to be of relevance in relation to induced abortion included age, education, place of residence, developmental status of the province where the woman resided, number of pregnancies and living children, occupation and family type.