I need to run a rank-ordered logit where some individuals ranked a maximum number of alternatives, and some ranked fewer than that maximum.
mlogit::mlogit.data doesn't like when the input data has differing numbers of ranks per individual.
Here's a small example, using data from the mlogit package.
In the original data, all individuals rank 6 alternatives, but this example modifies that by setting a random number of each individual's ranks to NA, then to 6, the maximum possible rank.
set.seed(62617)
data("Game2", package = "mlogit")
library(dplyr)
library(mlogit)
df1 <- Game2 %>% group_by(chid) %>%
mutate(ch = ifelse(ch > sample(2:6, size = 1), NA, ch))
dat <- df1 %>% mutate(ch = replace(ch, is.na(ch), 6)) %>% as.data.frame()
G <- mlogit.data(dat, shape = "long",
choice = "ch",
alt.var = "platform",
ranked = TRUE)
summary(mlogit(ch ~ own | hours + age, G, reflevel = "PC"))
This results in the following error:
Error in if (abs(x - oldx) < ftol) { :
missing value where TRUE/FALSE needed
How should I be coding unranked alternatives to make the model work?
