[Confidence interval calculation for small numbers of observations or no observations at all].

Gil Harari, Shimrit Herbst

Research output: Contribution to journalArticlepeer-review

Abstract

Confidence interval calculation is a common statistics measure, which is frequently used in the statistical analysis of studies in medicine and life sciences. A confidence interval specifies a range of values within which the unknown population parameter may lie. In most situations, especially those involving normally-distributed data or large samples of data from other distributions, the normal approximation may be used to calculate the confidence interval. But, if the number of observed cases is small or zero, we recommend that the confidence interval be calculated in more appropriate ways. In such cases, for example, in clinical trials where the number of observed adverse events is small, the criterion for approximate normality is calculated. Confidence intervals are calculated with the use of the approximated normal distribution if this criterion is met, and with the use of the exact binomial distribution if not. This article, accompanied by examples, describes the criteria in which the common and known method cannot be used as well as the stages and methods required to calculate confidence intervals in studies with a small number of observations.

Original languageEnglish
Pages (from-to)289-291, 304, 303
JournalHarefuah
Volume153
Issue number5
StatePublished - May 2014

ASJC Scopus subject areas

  • General Medicine

Fingerprint

Dive into the research topics of '[Confidence interval calculation for small numbers of observations or no observations at all].'. Together they form a unique fingerprint.

Cite this