The two-stage design involves sample size recalculation using an interim variance estimate. Stein proposed the design in 1945; biostatisticians recently have shown renewed interest in it. Wittes and Brittain proposed a modiΓΏcation aimed at greater e ciency; Gould and Shih proposed a similar procedur
Some Aspects of the Application of Internal Pilot Studies
β Scribed by Peter A. Lachenbruch; Janet Wittes
- Publisher
- John Wiley and Sons
- Year
- 2006
- Tongue
- English
- Weight
- 46 KB
- Volume
- 48
- Category
- Article
- ISSN
- 0323-3847
No coin nor oath required. For personal study only.
β¦ Synopsis
Abstract
This is a discussion of the paper' Sample size recalculation in internal pilot study designs' by Tim Friede and Meinhard Kieser, appearing in this special issue on adaptive designs. (Β© 2006 WILEYβVCH Verlag GmbH & Co. KGaA, Weinheim)
π SIMILAR VOLUMES
The assessment of sample size in clinical trials comparing population means requires a variance estimate of the main efficacy variable. When this variance estimate has a low precision, it may be appropriate to use the data from the first patients entered in the trial ('internal pilot study') to esti
Some aspects of the carbonization of polyimides Actr Polymeriu 39 (1988) Nr. 8 431 Copolymer1) see Scheme 2 Table 2. Properties of the poly(amidobenzimidazoles) Sulfur-containing monomer monomers showed that solubility increases with greater molar share of t h e amide p a r t s and, as a rule, t h e
less of the type of information created, there is a cost of creation which can involve one's time and other resources necessary to formulate ideas all the way to the billions ($) spent by the government. There must be incentives to expend the resources necessary to create the information. Sometimes
Molecular orbital calculations of sixfold barriers in nitromethane, methyl boron difluoride, and trifluoronitromethane were performed by various HartreeαFock and electron correlation methods. In those calculations, staggered and eclipsed conformations are of primary concern. These results indicated
When sample size is recalculated using unblinded interim data, use of the usual t-test at the end of a study may lead to an elevated type I error rate. This paper describes a numerical quadrature investigation to calculate the true probability of rejection as a function of the time of the recalculat