hiv: hiv

In LOGIT: Functions, Data and Code for Binary and Binomial Data

Description

A study of 47 patients with similar profiles. Measurements on cd4 and cd8 levels are displayed for each panel of patients having identical predictor profiles. Both cd4 and cd8 have three levels. The data should be modeled as a grouped logistic model, but may also be modeled as a count model depending on what a research wishes to determine from the data.

Usage

data(hiv)

Format

A data frame with 11 observations of grouped data with 4 variables.

infec

1=Patient diagnosed with HIV; 0=not diagnosed with HIV

cases

number of patients for each patient profile

cd4

3 levels: 0, 1, and 2

cd8

3 levels: 0, 1, and 2

Details

hiv is saved as a data frame.

Source

Hilbe, Practical Guide to Logistic Regression, Chapman & Hall/CRC.

References

Hilbe, Joseph M (2015), Practical Guide to Logistic Regression, Chapman & Hall/CRC.

Examples

# Not run
#data(hiv)
#table(hiv); hiv
#noinfec <- hiv$cases -  hiv$infec
#myhiv<- glm(cbind(infec, noinfec) ~  factor(cd4)  + factor#(cd8), family=binomial, data=hiv)
#summary(myhiv)
#mymodq <- glm( cbind(infec, noinfec) ~  factor(cd4)  + factor(cd8), family=quasibinomial, data=hiv)
#summary(mymodq)
#toOR(myhiv)
#End(Not run)

library(LOGIT)
data(hiv)
table(hiv); hiv
noinfec <- hiv$cases -  hiv$infec
response <- cbind(hiv$infec, noinfec)
myhiv<- glm(response ~  factor(cd4)  + factor(cd8), family=binomial, data=hiv)
summary(myhiv)

mymodq <- glm(response ~  factor(cd4)  + factor(cd8), family=quasibinomial, data=hiv)
summary(mymodq)

#library(sandwich)
#sqrt(diag(vcovHC(myhiv, type="HC0")))
toOR(myhiv)

LOGIT documentation built on May 29, 2017, 10:26 a.m.