Ulf Böckenholt, Sema Barlas, and Peter G.M. van der Heijden, "Do
Randomized-Response Designs Eliminate Response Biases?  An Empirical Study
of Non-Compliance Behavior", Journal of Applied Econometrics, Vol. 24, 
No. 3, 2009, pp. 377-392.

This paper uses the frequency data provided in Table 1 of the article.  The
data were obtained from two health surveys conducted in the years 2000 and
2002, which used the Kuk and Forced-Choice randomized response methods,
respectively.

The first four columns of this table contain the binary responses to four
health items.  A "1" indicates a negative (e.g., "No"/"Black") response.
The remaining two columns contain the frequency counts for each of the 
observed response patterns obtained in the two surveys.

The table is provided in the file bbh-data.txt.