Bootstrap hypothesis testing and power analysis at low dose levels
Machelle D. Wilson *
The University of Georgia. Savannah River Ecology Lab, P.O. Drawer
E, Aiken, SC 29801, United States
Received 16 May 2004; received in revised form 8 November 2004; accepted
17 November 2004
Available online 29 January 2005
Abstract
This study demonstrates the variability in dose estimates using the nonparametric
bootstrap to estimate the variability in the mean dose when mean values
from environmental data are used in the dose calculation. Bootstrap hypothesis
testing and power analysis are demonstrated. For the data set shown here,
the normal assumption works well if the environmental data can be considered
fixed, known constants. However, when there exists a good deal of variability
in the environmental data, as is most often the case, or where scarce
data are available, making a nomlal assumption leads to gross underestimation
of the variability in the mean dose.
@ 2004 Elsevier B. V. All rights reserved.
Keywords: Bootstrap; Power analysis; Hypothesis testing; Incurred
radioactive dose
SREL Reprint #2875
Wilson, M. D. 2005. Bootstrap hypothesis testing and power analysis at low dose levels. Science of the Total Environment 346:38-47.
