A note on the use of the variance inflation factor for determining sample size in cluster randomized trials

O C Ukoumunne, M C Gulliford, S Chinn

Research output: Contribution to journalArticlepeer-review

11 Citations (Scopus)

Abstract

The correct application of the variance inflation factor approach for sample size calculations in trials that randomize intact clusters of individuals rests on knowledge of the variance of the outcome under simple random allocation of individuals and the intraclass correlation coefficient at the given level of clustering., Each quantity may be derived from the between and within-cluster variance components, but often only the latter component is available to investigators. Substitution of the outcome variance by the within-cluster variance may result in an underestimation of the sample size required. The paper presents a modified variance inflation factor formula which is appropriate for use when the within-cluster variance component is known, or can be estimated, and an assumption is possible for the intraclass correlation coefficient.
Original languageEnglish
Pages (from-to)479 - 484
Number of pages6
JournalStatistician
Volume51
Issue number4
Publication statusPublished - 2002

Fingerprint

Dive into the research topics of 'A note on the use of the variance inflation factor for determining sample size in cluster randomized trials'. Together they form a unique fingerprint.

Cite this