prLogistic (version 1.2)

downer: Downer Data

Description

Data are related to dairy cows with a serious illness that may lead them to death. The dairy cows are also called downers because they sometimes lay down. A study was conducted with over 400 downer cows at the Ruakura New Zealand Animal Health Laboratory during 1983-1984. In this study a variety of blood tests were performed. The main goal of the study was to evaluate if the survival of the cows can be predicted from the blood measurements. For the illustration considered here, we use a subset of the original dataset. The variables on this dataset are serum asparate amino transferase, calving, serum creatine phosphokinase, Myopathy and Survival. These data are originally from Clark, Henderson, Hoggard, Ellison, and Young (1987). There were missing for some variables. We consider here only the complete cases.

Usage

data(downer)

Format

A data frame with 216 observations on the following 5 variables:

AST

Serum asparate amino transferase, in U/l at 30C, recoded to 1 if equal or less than 492, 0= greater than 492.

CK

Serum creatine phosphokinase, in U/l at 30C, recoded to 1 if equal or less than 5467, 0= greater than 5467.

Calving

Calving (0 if measured before calving, 1 if after).

Myopathy

Muscle disorder indicator (1=yes, 0=no).

Survival

Survival indicator (1=survived, 0=died or killed).

Source

The data come from Weisberg, S. (2005) Applied Linear Regression. Wiley-Interscience. These data are copyrighted by John Wiley & Sons Inc. We acknowledged it and use the data to illustrate the methodology included in this library.

References

Clark R, Henderson HV, Hoggard GK, Ellison R and Young B (1987) The ability of biochemical and haematological tests to predict recovery in periparturient recumbent cows. New Zealand Vet. Journal, 32, 126-133

Examples

data("downer", package = "prLogistic")
attach(downer)
# prop.table(table(Survival))
fit.logist1=glm(Survival ~ Calving + CK + AST + Myopathy, 
family=binomial, data=downer)
# summary(fit.logist1)
exp(fit.logist1$coef)