Displaying 1 result from an estimated 1 matches for "augmenteddat".
2007 May 10
1
Follow-up about ordinal logit with mixtures: how about 'continuation ratio' strategy?
...ented data frame as recommended by
### STEPHEN R. COLE, PAUL D. ALLISON, AND CANDE V. ANANTH,
### Estimation of Cumulative Odds Ratios
###
### Ann Epidemiol 2004;14:172?178.
### This new data frame has 3 lines for each original case,
### and a new dichotomous "response" variable called D
augmentedDat <- NULL
newdf <- dat
for (i in 2: length(levels(dat$y))) {
D <- ifelse ( dat$y >= i , 1, 0)
newone <- data.frame(newdf, i, D)
augmentedDat <- rbind(augmentedDat, newone)
}
### Compare previous POLR output to this model
summary(glm(D~ -1 + factor(i) + x + afactor , data=aug...