Estimation of Confidence Intervals for the Log-Normal Mean

• Activity Number: 670
• Type: Contributed
• Date/Time: Thursday, August 5, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Lingmin Zeng*+ and Lanju Zhang and Harry Yang
• Companies: MedImmune and MedImmune and MedImmune
• Keywords: lognormal ; confidence interval ; Cox's method ; bootstrap method ; Generalized confidence interval

Abstract

The lognormal distribution is a widely used model in bio-medical field. The statistical inference concerning its mean is a problem of interest. Through simulations, Zhou and Gao compared four main methods for construction of confidence intervals (CIs) of log-normal means: the naive method; Cox' method; conservative method and a parametric bootstrap method. They concluded that Cox' method is the best for moderate and large sample size while the bootstrap method is the best for small sample size. Later on, a generalized confidence interval (GCI) prosed by Krishnamoorthy and Mathew(2003) was claimed to be more satisfactory for small sample size. In this talk, we propose a new method to construct the CI for the lognormal mean. This method is easy to implement, and take less CPU time. Simulations are conducted to compare performance of all these method.