1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
|
\name{Means}
\alias{Means}
\alias{Means.data.frame}
\alias{Means.numeric}
\alias{Means.formula}
\alias{as.data.frame.means.table}
\alias{as.data.frame.xmeans.table}
\title{Means for groups of observations}
\description{The function \code{Means()} creates a table of group
means, optionally with standard errors, confidence intervals, and
numbers of valid observations.}
\usage{
Means(data, \dots)
\method{Means}{data.frame}(data,
by, weights=NULL, subset=NULL,
default=NA,
se=FALSE, ci=FALSE, ci.level=.95,
counts=FALSE, \dots)
\method{Means}{formula}(data, subset, weights, \dots)
\method{Means}{numeric}(data, \dots)
\method{as.data.frame}{means.table}(x, row.names=NULL, optional=TRUE, drop=TRUE, \dots)
\method{as.data.frame}{xmeans.table}(x, row.names=NULL, optional=TRUE, drop=TRUE, \dots)
}
\arguments{
\item{data}{an object usually containing data, or a formula.
If \code{data} is a numeric vector or an object that can be coerced
into a data frame, it is changed into a data frame and the data
frame method of \code{Means()} is applied to it.
If \code{data} is a formula, then a data frame is constructed from
the variables in the formula and \code{Means} is applied to this
data frame, while the formula is passed on as a \code{by=} argument.
}
\item{by}{a formula, a vector of variable names or a data frame or
list of factors.
If \code{by} is a vector of variable names,
they are extracted from \code{data} to define the groups for which
means are computed, while the variables for which the means are
computed are those not named in \code{by}.
If \code{by} is a data frame or a list of factors,
these are used to defined the groups for which means are computed,
while the variables for which the means are
computed are those not in \code{by}.
If \code{by} is a formula, its left-hand side determines the
variables of which means are computed, while its right-hand side
determines the factors that define the groups.
}
\item{weights}{an optional vector of weights, usually a variable in \code{data}.}
\item{subset}{an optional logical vector to select observations,
usually the result of an expression in variables from \code{data}.}
\item{default}{a default value used for empty cells without
observations.}
\item{se}{a logical value, indicates whether standard errors should be
computed.}
\item{ci}{a logical value, indicates whether limits of confidence
intervals should be computed.}
\item{ci.level}{a number, the confidence level of the confidence interval}
\item{counts}{a logical value, indicates whether numbers of valid
observations should be reported.}
\item{x}{for \code{as.data.frame()}, a result of \code{Means()}.}
\item{row.names}{an optional character vector. This argmument presently is
inconsequential and only included for reasons of compatiblity
with the standard methods of \code{\link[base]{as.data.frame}}.}
\item{optional}{an optional logical value. This argmument presently is
inconsequential and only included for reasons of compatiblity
with the standard methods of \code{\link[base]{as.data.frame}}.}
\item{drop}{a logical value, determines whether "empty cells" should
be dropped from the resulting data frame.}
\item{\dots}{other arguments, either ignored or passed on to other
methods where applicable.}
}
\value{
An array that inherits classes "means.table" and "table". If
\code{Means} was called with \code{se=TRUE} or \code{ci=TRUE}
then the result additionally inherits class "xmeans.table".
}
\examples{
# Preparing example data
USstates <- as.data.frame(state.x77)
USstates <- within(USstates,{
region <- state.region
name <- state.name
abb <- state.abb
division <- state.division
})
USstates$w <- sample(runif(n=6),size=nrow(USstates),replace=TRUE)
# Using the data frame method
Means(USstates[c("Murder","division","region")],by=c("division","region"))
Means(USstates[c("Murder","division","region")],by=USstates[c("division","region")])
Means(USstates[c("Murder")],1)
Means(USstates[c("Murder","region")],by=c("region"))
# Using the formula method
# One 'dependent' variable
Means(Murder~1, data=USstates)
Means(Murder~division, data=USstates)
Means(Murder~division, data=USstates,weights=w)
Means(Murder~division+region, data=USstates)
as.data.frame(Means(Murder~division+region, data=USstates))
# Standard errors and counts
Means(Murder~division, data=USstates, se=TRUE, counts=TRUE)
drop(Means(Murder~division, data=USstates, se=TRUE, counts=TRUE))
as.data.frame(Means(Murder~division, data=USstates, se=TRUE, counts=TRUE))
# Confidence intervals
Means(Murder~division, data=USstates, ci=TRUE)
drop(Means(Murder~division, data=USstates, ci=TRUE))
as.data.frame(Means(Murder~division, data=USstates, ci=TRUE))
# More than one dependent variable
Means(Murder+Illiteracy~division, data=USstates)
as.data.frame(Means(Murder+Illiteracy~division, data=USstates))
# Confidence intervals
Means(Murder+Illiteracy~division, data=USstates, ci=TRUE)
as.data.frame(Means(Murder+Illiteracy~division, data=USstates, ci=TRUE))
# Some 'non-standard' but still valid usages:
with(USstates,
Means(Murder~division+region,subset=region!="Northeast"))
with(USstates,
Means(Murder,by=list(division,region)))
}
|
