Simultaneous Confidence Intervals Based on Logarithm Transformations in Multi-Sample Models with Bernoulli Responses

• Shiraishi Taka-aki

  International College of Arts and Sciences, Yokohama-City University

Bibliographic Information

Other Title

• 多群２項モデルにおける対数変換による同時信頼区間
• タグン 2コウ モデル ニ オケル タイスウ ヘンカン ニ ヨル ドウジ シンライ クカン

Abstract

We consider simultaneous confidence intervals for the differences, ratios, odds ratios among propotions in k binomial populations. Although the simultaneous confidence intervals for all the pairwise differences among the propotions are expressed in Hochberg & Tamhane(1987), the intervals sometimes cause the contradiction including -1 or 1. To solve this contradiction, we construct the simultaneous confidence intervals based on logarithm transformations such as logit. Especially, we derive the upper and lower bounds for the asymptotic distribution of the statistic deriving the simultaneous cofidence intervals for all the pairwise differences. By using the inequalities, it is shown that the conservative is small. For multiple comparisons with a control, the procedures based on the Bonferroni inequality is discussed.

Journal

• Ouyou toukeigaku

  Ouyou toukeigaku 38 (3), 131-150, 2009

  Japanese Society of Applied Statistics

Keywords

• binomial distribution
• single-step procedures
• odds ratio
• asymptotics

Details

• CRID: 1390282679418980864
• NII Article ID: 10026049065
• NII Book ID: AN00330942
• DOI: 10.5023/jappstat.38.131
• ISSN: 18838081
• ISSN: 02850370
• NDL BIB ID: 10540088
• Web Site: http://id.ndl.go.jp/bib/10540088
• Web Site: https://ndlsearch.ndl.go.jp/books/R000000004-I10540088
• Text Lang: ja
• Data Source:
  • JaLC
  • NDL
  • Crossref
  • CiNii Articles
  • KAKEN
• Abstract License Flag: Disallowed