F tests with random sample sizes. Theory and applications

Elsa E. Moreira, João T. Mexia, Christoph E. Minder

Research output: Contribution to journalArticlepeer-review

10 Citations (Scopus)
27 Downloads (Pure)

Abstract

Given a fixed time span for collecting observations in a study comparing, for example, the pathologies of patients entering a hospital sequentially, it is advisable to consider the sample sizes of the ANOVA levels as random variables. Using this approach, more powerful tests are developed, leading to lower critical values. The approach is used to obtain the minimum duration of data collection to ensure a pre-fixed power for the F test.

Original languageEnglish
Pages (from-to)1520-1526
Number of pages7
JournalStatistics and Probability Letters
Volume83
Issue number6
DOIs
Publication statusPublished - 1 Jun 2013

Keywords

  • ANOVA
  • F distribution
  • Pathologies comparison
  • Poisson distribution
  • Power analysis

Fingerprint

Dive into the research topics of 'F tests with random sample sizes. Theory and applications'. Together they form a unique fingerprint.

Cite this