Logarithmic confidence intervals for the cross-product ratio of binomial proportions under different sampling schemes

We consider the problem of logarithmic interval estimation for a cross-product ratio rho=p(1)(1-p(2))/p(2)(1-p(1)) with data from two independent Bernoulli samples. Each sample may be obtained in the framework of direct or inverse Binomial sampling schemes. Asymptotic logarithmic confidence intervals are constructed under different types of sampling schemes, with parameter estimators demonstrating exponentially decreasing bias. Our goal is to investigate the cases when the relatively simple normal approximations for estimators of the cross-product ratio are reliable for constructing logarithmic confidence intervals. We use the closeness of the confidence coefficient to the nominal confidence level as our main evaluation criterion, and use the Monte-Carlo method to investigate the key probability characteristics of intervals corresponding to all possible combinations of sampling schemes. We present estimations of the coverage probability, expectation and standard deviation of interval widths in tables. Also, we provide some recommendations for applying each logarithmic interval obtained.

Automated summary

We investigated the problem of logarithmic interval estimation for a cross-product ratio with data from two independent Bernoulli samples. Asymptotic logarithmic confidence intervals were constructed under different types of sampling schemes, and the parameter estimators showed exponentially decreasing bias. Our findings suggest that the relatively simple normal approximations are reliable for constructing logarithmic confidence intervals.