Advances and Applications in Statistics
Volume 9, Issue 1, Pages 1 - 12
(June 2008)
|
|
AN IMPROVED APPROXIMATE UNBIASED ESTIMATOR OF LOG-ODDS RATIO FOR CONTINGENCY TABLES
Kouji Tahata (Japan), Ryota Tomisato (Japan) and Sadao Tomizawa (Japan)
|
Abstract: For a
contingency table, the estimated log-odds ratio based on the sample proportions is asymptotically unbiased estimator of the true log-odds ratio. This paper proposes the improved approximate unbiased estimator of log-odds ratio using the second-order Taylor expansion of estimated log-odds ratio about the sample proportions being equal to the true probabilities. By the simulation studies, it is shown that the proposed estimator approaches the true log-odds ratio faster than the usual estimator as the sample size becomes larger and also as the value of true log-odds ratio becomes larger or smaller than zero. |
Keywords and phrases: delta method, log-odds ratio, odds ratio, contingency table, unbiased estimator. |
|
Number of Downloads: 378 | Number of Views: 1280 |
|