Posted on Jun 1, 2001

Josef Schmee, the Kenneth B. Sharpe Professor of Management, has published a paper, “Likelihood Confidence Intervals for Proportions in Finite Populations” in the February issue of the American Statistician. The paper was co-authored by John P. Wendell, professor of accounting at the University of Hawaii, and a Ph.D. student at Union. The paper shows a method to obtain the shortest possible confidence intervals for the number or proportion of items in a finite population that possess a particular characteristic. Such intervals are of great importance in applications such as auditing or in sample surveys from smaller populations. The proposed method is based on the likelihood of the observed sample and although computationally a bit more complex, it is superior to existing methods.