 
Summary: GRADUATE SEMINAR
Kezia Baidoo
Confidence Intervals for Ratios of Binomial
Proportions Based on
Unbiased Estimators
MSc Student supervised by A. Volodin and A. Sardarli
Monday, April 26, 2010
10:00 a.m.
College West 307.20 (CW307.20)
Abstract: A general statement of the interval estimation problem of a
proportion of two ratios according to data from two independent samples is
considered. Each sample may be obtained in the framework of direct or inverse
binomial sampling. Four asymptotic confidence intervals are constructed in
accordance with different types of sampling schemes. The main probability
characteristics of the intervals are investigated by the MonteCarlo method:
coverage probability, median, mean, and standard deviation of interval length.
The results of the simulations show that we can arrange the new confidence
intervals in descending order according to the size of the regions of values where
their application is recommended as follows: directdirect, directinverse when
the asymptotic variance is used, directinverse when the number of successes in
