Message-ID: <Pine.LNX.4.43.0903240118590.11646@hymn11.u.washington.edu>
Date: 2009-03-24T08:18:59Z
From: Thomas Lumley
Subject: confidence interval or error of x intercept of a linear regression
In-Reply-To: <1237865492.586760.alphamail@mailapps1.uoregon.edu>
I should probably also add the health warning that the delta-method approach to ratios of coefficients is REALLY BAD unless the denominator is well away from zero.
When the denominator is near zero nothing really works (Fieller's method is valid but only because 'valid' means less than you might like for a confidence interval method, and the bootstrap is useful but not necessarily valid)
-thomas
Thomas Lumley Assoc. Professor, Biostatistics
tlumley at u.washington.edu University of Washington, Seattle