varRatio: First order approximation to variance of y-ordinate to slope...

In jhmaindonald/qra-R-package: Quantal Response Analysis for Dose-Mortality Data

First order approximation to variance of y-ordinate to slope ratio

Description

In contexts where an LD99 estimate will be used as a data value in a further analysis step, the inverse of the variance may be used as a weight. The y-ordinate is for the link function transformed value of a specified mortality proportion, commonly 0.50, or 0.90, or 0.99

Usage

varRatio(phat = 0.99, b, vv, link = "cloglog")

Arguments

phat	Mortality proportion
b	Length 2 vector of intercept and slope
vv	Variance-covariance matrix for intercept and slope
link	Link function that transforms expected mortalities to the scale of the linear predictor

Details

This function should only be used, in order to speed up calculations that use the function fieller (call fieller with (type="Delta")), in a context where it is to be used many times, and where a check has been made that its use leads to confidence intervals that are a close approximation to those given with the default argument (type="Fieller").

Value

A vector, with elements

xhat	Estimate
var	Variance, calculated using the Delta method, See the help page for fieller for further details and references.

Examples

redDel <- subset(qra::codling1988, Cultivar=="Red Delicious")
redDel.glm <- glm(cbind(dead,total-dead)~ct, data=redDel,
                  family=quasibinomial(link='cloglog'))
vv <- summary(redDel.glm)$cov.scaled
qra::varRatio(0.99, b=coef(redDel.glm), vv=vv, link="cloglog")

jhmaindonald/qra-R-package documentation built on Nov. 16, 2023, 2:39 p.m.