Question

Possible Duplicate:
How to create a prop.table() for a three dimension table?

I am new to R, and I have been working on a three way table and need some help. Here is the table description via dput:

mytable <- structure(c(42L, 151L, 313L, 69L, 22L, 46L, 47L, 24L, 17L, 36L, 108L, 16L), .Dim = c(2L, 2L, 3L), .Dimnames = list(c("0", "1" ), c("female", "male"), c("adult", "child", "unknown")), class = "table") 

I would like to access elements of this table by using the attribute names. For example:
mytable["0"]["female"]["adult"]Is it possible to do this ?

Also, I am trying to calculate proportions. Here is my table….

, , adult

      female male
    0     42  313
    1    151   69

    , , child

      female male
    0     22   47
    1     46   24

    , , unknown

      female male
    0     17  108
    1     36   16

I would like to calculate proportions for age group-sex-0|1. When I run prop.table(mytable)it is calculating proportions of the cells, but it is considering all the three age groups.

, , adult

          female       male
    0 0.04713805 0.35129068
    1 0.16947250 0.07744108

    , , child

          female       male
    0 0.02469136 0.05274972
    1 0.05162738 0.02693603

    , , unknown

          female       male
    0 0.01907969 0.12121212
    1 0.04040404 0.01795735

I want proportions for each age group and sex. For example: Of all the adults, females with 1 are 78 %. How do we do this ?

Answer 1

Is this what you are chasing?

The data:

test <- structure(c(42L, 151L, 313L, 69L, 22L, 46L, 47L, 24L, 17L, 36L, 108L, 16L), .Dim = c(2L, 2L, 3L), .Dimnames = list(c("0", "1" ), c("female", "male"), c("adult", "child", "unknown")), class = "table") 

Get the column percentages (2) within each sub-table (3):

prop.table(test,c(2,3))

Results in:

, , adult

     female      male
0 0.2176166 0.8193717
1 0.7823834 0.1806283

, , child

     female      male
0 0.3235294 0.6619718
1 0.6764706 0.3380282

, , unknown

     female      male
0 0.3207547 0.8709677
1 0.6792453 0.1290323