330 – Making Sense of Network Data and Randomized Response
Estimation of Odds Ratio, Attributable Risk, Relative Risk, Correlation Coefficient, and Other Parameters Using Randomized Response Techniques
Cheon-Sig Lee
San Jacinto College
In this paper, we first define odds ratio, attributable risk, relative risk, correlation coefficient, membership in at least one group, difference between two proportions, a new estimator of single proportion when the proportion of second sensitive variable is know, while considering investigating two sensitive attributes in real practice. Then we define two estimators in each one of these cases based on simple model or crossed model proposed by Lee, Sedory and Singh (2013). We derive expressions for biases and variances of the resultant estimators. We investigate the performance of estimators based on the crossed model over those based on the simple model under the same choice of parameters, as discussed in Lee et al. (2013). Also the values of the various statistics such as odds ratio, attributable risk, relative risk, correlation coefficient, membership in one variable and the difference between two proportions are estimated based on a real data set.