mChoice: Methods for Storing and Analyzing Multiple Choice Variables

Description

mChoice is a function that is useful for grouping variables that represent individual choices on a multiple choice question. These choices are typically factor or character values but may be of any type. Levels of component factor variables need not be the same; all unique levels (or unique character values) are collected over all of the multiple variables. Then a new character vector is formed with integer choice numbers separated by semicolons. Optimally, a database system would have exported the semicolon-separated character strings with a levels attribute containing strings defining value labels corresponding to the integer choice numbers. mChoice is a function for creating a multiple-choice variable after the fact. mChoice variables are explicitly handed by the describe and summary.formula functions. NAs or blanks in input variables are ignored.

format.mChoice will convert the multiple choice representation to text form by substituting levels for integer codes. as.double.mChoice converts the mChoice object to a binary numeric matrix, one column per used level (or all levels of drop=FALSE. This is called by the user by invoking as.numeric. There is a print method and a summary method, and a print method for the summary.mChoice object. The summary method computes frequencies of all two-way choice combinations, the frequencies of the top 5 combinations, information about which other choices are present when each given choice is present, and the frequency distribution of the number of choices per observation. This summary output is used in the describe function.

in.mChoice creates a logical vector the same length as x whose elements are TRUE when the observation in x contains at least one of the codes or value labels in the second argument.

match.mChoice creats an integer vector of the indexes of all elements in table which contain any of the speicified levels

is.mChoice returns TRUE is the argument is a multiple choice variable.

Usage

mChoice(..., label='',
        sort.levels=c('original','alphabetic'), 
        add.none=FALSE, drop=TRUE)
# S3 method for mChoice
format(x, minlength=NULL, sep=";", ...)
# S3 method for mChoice
as.double(x, drop=FALSE, ...)
# S3 method for mChoice
print(x, quote=FALSE, max.levels=NULL,
       width=getOption("width"), ...)
# S3 method for mChoice
as.character(x, ...)
# S3 method for mChoice
summary(object, ncombos=5, minlength=NULL, drop=TRUE, ...)
# S3 method for summary.mChoice
print(x, prlabel=TRUE, ...)
# S3 method for mChoice
[(x, ..., drop=FALSE)
match.mChoice(x, table, nomatch=NA, incomparables=FALSE)
inmChoice(x, values)
is.mChoice(x)
# S3 method for mChoice
Summary(..., na.rm)

Value

mChoice returns a character vector of class "mChoice"

plus attributes "levels" and "label".

summary.mChoice returns an object of class

"summary.mChoice". inmChoice returns a logical vector.

format.mChoice returns a character vector, and

as.double.mChoice returns a binary numeric matrix.

Arguments

na.rm: Logical: remove NA's from data
table: a vector (mChoice) of values to be matched against.
nomatch: value to return if a value for x does not exist in table.
incomparables: logical whether incomparable values should be compaired.
...: a series of vectors
label: a character string label attribute to attach to the matrix created by mChoice
sort.levels: set sort.levels="alphabetic" to sort the columns of the matrix created by mChoice alphabetically by category rather than by the original order of levels in component factor variables (if there were any input variables that were factors)
add.none: Set add.none to TRUE to make a new category 'none' if it doesn't already exist and if there is an observations with no choices selected.
drop: set drop=FALSE to keep unused factor levels as columns of the matrix produced by mChoice
x: an object of class "mchoice" such as that created by mChoice. For is.mChoice is any object.
object: an object of class "mchoice" such as that created by mChoice
ncombos: maximum number of combos.
width: With of a line of text to be formated
quote: quote the output
max.levels: max levels to be displayed
minlength: By default no abbreviation of levels is done in format and summary. Specify a positive integer to use abbreviation in those functions. See abbreviate.
sep: character to use to separate levels when formatting
prlabel: set to FALSE to keep print.summary.mChoice from printing the variable label and number of unique values
values: a scalar or vector. If values is integer, it is the choice codes, and if it is a character vector, it is assumed to be value labels.

Author

Frank Harrell
Department of Biostatistics
Vanderbilt University
fh@fharrell.com

Examples

Run this code

options(digits=3)
set.seed(3)
n <- 20
sex <- factor(sample(c("m","f"), n, rep=TRUE))
age <- rnorm(n, 50, 5)
treatment <- factor(sample(c("Drug","Placebo"), n, rep=TRUE))


# Generate a 3-choice variable; each of 3 variables has 5 possible levels
symp <- c('Headache','Stomach Ache','Hangnail',
          'Muscle Ache','Depressed')
symptom1 <- sample(symp, n, TRUE)
symptom2 <- sample(symp, n, TRUE)
symptom3 <- sample(symp, n, TRUE)
cbind(symptom1, symptom2, symptom3)[1:5,]
Symptoms <- mChoice(symptom1, symptom2, symptom3, label='Primary Symptoms')
Symptoms
print(Symptoms, long=TRUE)
format(Symptoms[1:5])
inmChoice(Symptoms,'Headache')
levels(Symptoms)
inmChoice(Symptoms, 3)
inmChoice(Symptoms, c('Headache','Hangnail'))
# Note: In this example, some subjects have the same symptom checked
# multiple times; in practice these redundant selections would be NAs
# mChoice will ignore these redundant selections

meanage <- N <- numeric(5)
for(j in 1:5) {
 meanage[j] <- mean(age[inmChoice(Symptoms,j)])
 N[j] <- sum(inmChoice(Symptoms,j))
}
names(meanage) <- names(N) <- levels(Symptoms)
meanage
N

# Manually compute mean age for 2 symptoms
mean(age[symptom1=='Headache' | symptom2=='Headache' | symptom3=='Headache'])
mean(age[symptom1=='Hangnail' | symptom2=='Hangnail' | symptom3=='Hangnail'])

summary(Symptoms)

#Frequency table sex*treatment, sex*Symptoms
summary(sex ~ treatment + Symptoms, fun=table)
# Check:
ma <- inmChoice(Symptoms, 'Muscle Ache')
table(sex[ma])

# could also do:
# summary(sex ~ treatment + mChoice(symptom1,symptom2,symptom3), fun=table)

#Compute mean age, separately by 3 variables
summary(age ~ sex + treatment + Symptoms)


summary(age ~ sex + treatment + Symptoms, method="cross")

f <- summary(treatment ~ age + sex + Symptoms, method="reverse", test=TRUE)
f
# trio of numbers represent 25th, 50th, 75th percentile
print(f, long=TRUE)