A randomized response will enable you to obtain accurate answers to your sensitive question. A carefully selected sample, obtained by personal interviews, can avoid the non-response problem. I would not be surprised that the cost would turn out to be much the same as what you are now doing especially if you project the realistic costs into the further and allow for the problems that incorrect information causes.
-- Bob Wheeler --- http://www.bobwheeler.com/ ECHIP, Inc. --- Randomness comes in bunches. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
