percentages: Easy Creation of Tables of Percentages

In memisc: Tools for Management of Survey Data and the Presentation of Analysis Results

Description

The generic function percentages and its methods create one- or multidimensional tables of percentages. As such, the function percentages can be viewed as a convenience interface to prop.table. However, it also allows to obtain standard errors and confidence intervals.

Usage

percentages(obj, ...)
## S3 method for class 'table'
percentages(obj,
      by=NULL, which=NULL, se=FALSE, ci=FALSE, ci.level=.95, ...)
## S3 method for class 'formula'
percentages(obj,
      data=parent.frame(), weights=NULL, ...)
## S3 method for class 'percentage.table'
as.data.frame(x, ...)
## S3 method for class 'xpercentage.table'
as.data.frame(x, ...)

Arguments

obj	an object; a contingency table or a formula. If it is a formula, its left-hand side determines the factor or combination of factors for which percentages are computed while its right-hand side determines the factor or combination of factors that define the groups within which percentages are computed.
by	a character vector with the names of the factor variables that define the groups within which percentages are computed. Percentages sum to 100 within combination of levels of these factors.
which	a character vector with the names of the factor variables for which percentages are computed.
se	a logical value; determines whether standard errors are computed.
ci	a logicla value; determines whether confidence intervals are computed. Note that the confidence intervals are for infinite (or very large) populations.
ci.level	a numerical value, the required confidence level of the confidence intervals.
data	a contingency table (an object that inherits from "table") or a data frame or an object coercable into a data frame.
weights	an optional character string with the name of a variable containing weights.
x	an object coerced into a data frame.
...	Further arguments, ignored.

Value

An array that inherits classes "percentage.table" and "table". If percentages was called with se=TRUE or ci=TRUE then the result additionally inherits class "xpercentage.table".

Examples

# Two equivalent ways to create the same table
100*prop.table(UCBAdmissions)
percentages(UCBAdmissions)

# Three equivalent ways to create the same table
100*prop.table(UCBAdmissions,c(2,3))
(p0 <- percentages(UCBAdmissions,by=c("Gender","Dept")))
percentages(UCBAdmissions,which="Admit")
# Percentage table as data frame
as.data.frame(p0)

# Three equivalent ways to create the same table
100*prop.table(margin.table(UCBAdmissions,c(1,2)),2)
percentages(UCBAdmissions,which="Admit",by="Gender")
percentages(Admit~Gender,data=UCBAdmissions)

# Three equivalent ways to create the same table
100*prop.table(margin.table(UCBAdmissions,c(1,3)),2)
percentages(Admit~Dept,data=UCBAdmissions)
percentages(Admit~Dept,data=as.data.frame(UCBAdmissions),
            weights="Freq")

# Standard errors and confidence intervals
percentages(Admit~Dept,data=UCBAdmissions,se=TRUE)
percentages(Admit~Dept,data=UCBAdmissions,ci=TRUE)
(p<- percentages(Admit~Dept,data=UCBAdmissions,ci=TRUE,se=TRUE))

# An extended table of percentages as data frame
as.data.frame(p)

memisc documentation built on May 2, 2019, 5:45 p.m.