Home
Scholarly Works
A geometric confidence ellipse approach to the...
Journal article

A geometric confidence ellipse approach to the estimation of the ratio of two variables. Statistics in Medicine 2008; 27:5956–5974.

Abstract

We are grateful to Dr Etienne Kaelin for pointing out four typographic errors in the equations for this paper. These are as follows: 1. The term xy in equation (1) should be replaced by: \documentclass{article}\usepackage{amssymb}\usepackage{amsbsy}\usepackage[mathscr]{euscript}\footskip=0pc\pagestyle{empty}\begin{document}\begin{eqnarray*}(x-\mu_{x})(y-\mu_{x})\end{eqnarray*}\end{document} 2. Equation (3) should be \documentclass{article}\usepackage{amssymb}\usepackage{amsbsy}\usepackage[mathscr]{euscript}\footskip=0pc\pagestyle{empty}\begin{document}\begin{eqnarray*}t=\frac{\rho\sigma_x\sigma_yc^2-\mu_x\mu_y\pm[(\mu_x\mu_y-\rho\sigma_x\sigma_y c^2)^2-(\sigma_x^2c^2-\mu_x^2)(\sigma_y^2c^2-\mu_y^2)]^{1/2}}{\sigma_x^2c^2-\mu_x^2}\end{eqnarray*}\end{document} 3. The first term of the unnumbered equation after equation (3) should be µy rather than µx. 4. The correct version of equation (4) is \documentclass{article}\usepackage{amssymb}\usepackage{amsbsy}\usepackage[mathscr]{euscript}\footskip=0pc\pagestyle{empty}\begin{document}\begin{eqnarray*}x=\frac{-y\sigma_x(\sigma_x\mu_y-\sigma_y\mu_x\rho)+v-u}{\sigma_y(\sigma_y\mu_x-\sigma_x\mu_y \rho)}\end{eqnarray*}\end{document} The numerical results of this paper as shown in the various tables are not affected by these errors because correct versions of the equations were used in their calculation. The term xy in equation (1) should be replaced by: \documentclass{article}\usepackage{amssymb}\usepackage{amsbsy}\usepackage[mathscr]{euscript}\footskip=0pc\pagestyle{empty}\begin{document}\begin{eqnarray*}(x-\mu_{x})(y-\mu_{x})\end{eqnarray*}\end{document} Equation (3) should be \documentclass{article}\usepackage{amssymb}\usepackage{amsbsy}\usepackage[mathscr]{euscript}\footskip=0pc\pagestyle{empty}\begin{document}\begin{eqnarray*}t=\frac{\rho\sigma_x\sigma_yc^2-\mu_x\mu_y\pm[(\mu_x\mu_y-\rho\sigma_x\sigma_y c^2)^2-(\sigma_x^2c^2-\mu_x^2)(\sigma_y^2c^2-\mu_y^2)]^{1/2}}{\sigma_x^2c^2-\mu_x^2}\end{eqnarray*}\end{document} The first term of the unnumbered equation after equation (3) should be µy rather than µx. The correct version of equation (4) is \documentclass{article}\usepackage{amssymb}\usepackage{amsbsy}\usepackage[mathscr]{euscript}\footskip=0pc\pagestyle{empty}\begin{document}\begin{eqnarray*}x=\frac{-y\sigma_x(\sigma_x\mu_y-\sigma_y\mu_x\rho)+v-u}{\sigma_y(\sigma_y\mu_x-\sigma_x\mu_y \rho)}\end{eqnarray*}\end{document} The numerical results of this paper as shown in the various tables are not affected by these errors because correct versions of the equations were used in their calculation. Dr Kaelin has offered to make available an Excel spreadsheet that will carry out the calculations for the method described in this paper. It is available to interested readers by contacting him at Etienne.Kaelin@pmintl.com.

Authors

Walter SD; Gafni A; Birch S

Journal

Statistics in Medicine, Vol. 28, No. 21, pp. 2723–2723

Publisher

Wiley

Publication Date

September 20, 2009

DOI

10.1002/sim.3642

ISSN

0277-6715

Contact the Experts team