| Literature DB >> 12392862 |
Abstract
A formula to estimate the required sample size for a study with repeated measurements was constructed based on the test of an interaction term in a mixed model. It covers both random effects and serial correlation and allows for missing data. This formula indicates that the method suggested by Dawson is conservative. A simulation study further verified the accuracy of the formula. Factors that influence the required sample size, such as the number of repeated measurements, the structure of the within-subject correlation, and their interaction, were investigated in detail. Copyright 2002 Elsevier Science Inc.Mesh:
Year: 2002 PMID: 12392862 DOI: 10.1016/s0197-2456(02)00223-4
Source DB: PubMed Journal: Control Clin Trials ISSN: 0197-2456