multi.freq.table: Tabulates Columns From a 'data.frame' Containing...
In mrdwab/mrdwabmisc: Miscellaneous R functions, mostly for data processing

Description Usage Arguments Details Author(s) References Examples

The multi.freq.table function takes a data.frame containing Boolean responses to multiple response questions and tabulates the number of responses by the possible combinations of answers.

1 2	multi.freq.table(data, sep = "", boolean = TRUE, factors = NULL, NAto0 = TRUE, basic = FALSE, dropzero = TRUE, clean = TRUE)

`data`	The multiple responses that need to be tabulated.
`sep`	The desired separator for collapsing the combinations of options; defaults to `""` (collapsing with no space between each option name).
`boolean`	Are you tabulating boolean data (see `dat` Examples)? Defaults to `TRUE`.
`factors`	If you are trying to tabulate non-boolean data, and the data are not factors, you can specify the factors here (see `dat2` Examples). Defaults to `NULL` and is not used when `boolean = TRUE`.
`NAto0`	Should `NA` values be converted to `0`? Defaults to `TRUE`, in which case, the number of valid cases should be the same as the number of cases overall. If set to `FALSE`, any rows with `NA` values will be dropped as invalid cases. Only applies when `boolean = TRUE`.
`basic`	Should a basic table of each item, rather than combinations of items, be created? Defaults to `FALSE`.
`dropzero`	Should combinations with a frequency of zero be dropped from the final table? Defaults to `TRUE`. Does not apply when `boolean = TRUE`.
`clean`	Should the original tabulated data be retained or dropped from the final table? Defaults to `TRUE` (drop). Does not apply when `boolean = TRUE`.

In addition to tabulating the frequency (Freq), there are two other columns in the output: Percent of Responses (Pct.of.Resp) and Percent of Cases (Pct.of.Cases).

Percent of Responses is the frequency divided by the total number of answers provided; this column should sum to 100 combination table is generated and there are cases where a respondent did not select any option, the Percent of Responses value would be more than 100 cases; this column would most likely sum to more than 100 table is produced since each respondent (case) can select multiple answers, but should sum to 100

Ananda Mahto

apply shortcut for creating the Combn column in the output by Justin. See: http://stackoverflow.com/q/11348391/1270695 and http://stackoverflow.com/q/11622660/1270695

## ========================================= ##
## ============= BOOLEAN DATA ============== ##

# Make up some data
set.seed(1)
dat <- data.frame(A = sample(c(0, 1), 20, replace=TRUE),
            B = sample(c(0, 1, NA), 20,
                       prob=c(.3, .6, .1), replace=TRUE),
            C = sample(c(0, 1, NA), 20,
                       prob=c(.7, .2, .1), replace=TRUE),
            D = sample(c(0, 1, NA), 20,
                       prob=c(.3, .6, .1), replace=TRUE),
            E = sample(c(0, 1, NA), 20,
                       prob=c(.4, .4, .2), replace=TRUE))

# View your data
dat

# How many cases have "NA" values?
table(is.na(rowSums(dat)))

# Apply the function with all defaults accepted
multi.freq.table(dat)

# Tabulate only on variables "A", "B", and "D", with a different
# separator, keep any zero frequency values, and keeping the
# original tabulations. There are no solitary "D" responses.
multi.freq.table(dat[c(1, 2, 4)], sep="-", dropzero=FALSE, clean=FALSE)

# As above, but without converting "NA" to "0".
# Note the difference in the number of valid cases.
multi.freq.table(dat[c(1, 2, 4)], NAto0=FALSE,
            sep="-", dropzero=FALSE, clean=FALSE)

# View a basic table.
multi.freq.table(dat, basic=TRUE)

## ========================================= ##
## =========== NON-BOOLEAN DATA ============ ##

# Make up some data
dat2 <- structure(list(Reason.1 = c("one", "one", "two", "one", "two",
                              "three", "one", "one", NA, "two"),
                 Reason.2 = c("two", "three", "three", NA, NA,
                              "two", "three", "two", NA, NA),
                 Reason.3 = c("three", NA, NA, NA, NA,
                              NA, NA, "three", NA, NA)),
                 .Names = c("Reason.1", "Reason.2", "Reason.3"),
                 class = "data.frame",
                 row.names = c(NA, -10L))

# View your data
dat2

## Not run: # The following will not work.
# The data are not factored.
multi.freq.table(dat2, boolean=FALSE)
## End(Not run)

# Factor create the factors.
multi.freq.table(dat2, boolean=FALSE,
            factors = c("one", "two", "three"))

# And, a basic table.
multi.freq.table(dat2, boolean=FALSE,
            factors = c("one", "two", "three"),
            basic=TRUE)