I am wondering whether there is a direct way to get nested relative frequencies in a 3-way flat (!) contingency table.
N <- 10
(sex <- factor(sample(c("f", "m"), N, replace=TRUE)))
(work <- factor(sample(c("home", "office"), N, replace=TRUE)))
(satisfaction <- factor(sample(c("excellent", "ok","bad"), N, replace=TRUE)))
ftable(work, sex, satisfaction, row.vars=c("sex", "satisfaction"))
This code produces this output:
work home office
sex satisfaction
f bad 1 1
excellent 0 1
ok 1 1
m bad 1 0
excellent 1 0
ok 1 2
What I would like to get is
work home office
sex satisfaction
f bad 0.2 0.2
excellent 0 0.2
ok 0.2 0.2
m bad 0.2 0
excellent 0.2 0
ok 0.2 0.4
I am aware that i can calculate tables with prop.table but haven't figured out how this could be combined with flat contingency tables as produced by ftable. There is certainly a way to calculate the relative frequencies separately for each sex - group, but I am interested in a more direct way. So far I haven't found any way how to get relative frequencies into flat contingency tables.
Note: In this example male and female groups have the same size (5); however, the solution I am looking for should be flexible enough so that groups can also have different sizes.
Many thanks!
Perhaps you can consider starting with table, then using prop.table wrapped in ftable:
y <- table(sex, satisfaction, work)
y
# , , work = home
#
# satisfaction
# sex bad excellent ok
# f 1 1 0
# m 1 1 1
#
# , , work = office
#
# satisfaction
# sex bad excellent ok
# f 0 1 1
# m 0 1 2
ftable(prop.table(y, margin=3))
# work home office
# sex satisfaction
# f bad 0.2 0.0
# excellent 0.2 0.2
# ok 0.0 0.2
# m bad 0.2 0.0
# excellent 0.2 0.2
# ok 0.2 0.4
This is assuming we start with something like this:
set.seed(1)
N <- 10
sex <- factor(sample(c("f", "m"), N, replace=TRUE))
work <- factor(sample(c("home", "office"), N, replace=TRUE))
satisfaction <- factor(sample(c("excellent", "ok","bad"), N, replace=TRUE))
Similarly, if you already have an ftable (let's call it "x"), you can try:
ftable(prop.table(as.table(x), margin=3))