samplesize: Sample Size Estimation For Calibration Study

In CVcalibration: Estimation of the Calibration Equation with Error-in Observations

Description

Compute the sample size needed for a calibration study assuming that the coefficients of the variation of measurements are constants.

Usage

samplesize(x0, d0, x=seq(20, 200, length=1000), CVx, CVy)

Arguments

x0	The x-value you plan to calibrate with the estimated calibration equation
d0	The required length of the 95% confidence interval of the calibrated x-value based on the estimated calibration equation. The narrow confidence interval require a large sample size.
x	The empirical observations of the target distribution from which $x$-values will be drawn in the planned study.
CVx	The underlying coefficient of variation of measurement $x$
CVy	The underlying coefficient of variation of measurement $y$

Value

size	The sample size needed for estimating the calibration equation based on (1) CVx only; (2) CVy only; (3) both CVx and CVy; and (4) the ratio of CVy^2/CVx^2.

Author(s)

Lu Tian, He Qi

Examples

samplesize(40, 10, x=seq(10, 200, length=1000), CVx=0.1, CVy=0.05)

Example output

$size
     CV of X only CV of Y only CV of both X and Y CVy/CVx
[1,]           23           18                 18      18

CVcalibration documentation built on May 2, 2019, 2:36 p.m.