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
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
|
\name{coeftest}
\alias{coeftest}
\alias{coefci}
\alias{coeftest.default}
\alias{coeftest.survreg}
\alias{coeftest.glm}
\alias{coeftest.mlm}
\alias{coeftest.breakpointsfull}
\alias{print.coeftest}
\alias{confint.coeftest}
\alias{coef.coeftest}
\alias{df.residual.coeftest}
\alias{nobs.coeftest}
\alias{logLik.coeftest}
\alias{coefci.default}
\alias{coefci.survreg}
\alias{coefci.glm}
\alias{coefci.mlm}
\title{Inference for Estimated Coefficients}
\description{
\code{coeftest} is a generic function for performing
z and (quasi-)t Wald tests of estimated coefficients.
\code{coefci} computes the corresponding Wald confidence
intervals.
}
\usage{
coeftest(x, vcov. = NULL, df = NULL, \dots)
\method{coeftest}{default}(x, vcov. = NULL, df = NULL, \dots, save = FALSE)
coefci(x, parm = NULL, level = 0.95, vcov. = NULL, df = NULL, \dots)
}
\arguments{
\item{x}{an object (for details see below).}
\item{vcov.}{a specification of the covariance
matrix of the estimated coefficients. This can be
specified as a matrix or as a function yielding
a matrix when applied to \code{x}.}
\item{df}{the degrees of freedom to be used. If this
is a finite positive number a t test with \code{df}
degrees of freedom is performed. In all other cases,
a z test (using a normal approximation) is performed.
By default it tries to use \code{x$df.residual}
and performs a z test if this is \code{NULL}.}
\item{\dots}{further arguments passed to the methods
and to \code{vcov.} in the default method.}
\item{save}{logical. Should the object \code{x} itself be
saved as an attribute? (This may be useful for further
processing of \code{coeftest} objects, e.g., as part of
model summaries.)}
\item{parm}{a specification of which parameters are to be given
confidence intervals, either a vector of numbers or a vector
of names. If missing, all parameters are considered.}
\item{level}{the confidence level required.}
}
\details{
The generic function \code{coeftest} currently has a default
method (which works in particular for \code{"lm"} objects) and
dedicated methods for objects of class
\code{"glm"} (as computed by \code{\link[stats]{glm}}),
\code{"mlm"} (as computed by \code{\link[stats]{lm}} with multivariate responses),
\code{"survreg"} (as computed by \code{\link[survival]{survreg}}), and
\code{"breakpointsfull"} (as computed by \code{\link[strucchange]{breakpoints.formula}}).
The default method assumes that a \code{coef} methods exists,
such that \code{coef(x)} yields the estimated coefficients.
To specify the corresponding covariance matrix \code{vcov.} to be used, there
are three possibilities:
1. It is pre-computed and supplied in argument \code{vcov.}.
2. A function for extracting the covariance matrix from
\code{x} is supplied, e.g., \code{\link[sandwich]{sandwich}},
\code{\link[sandwich]{vcovHC}}, \code{\link[sandwich]{vcovCL}},
or \code{\link[sandwich]{vcovHAC}} from package \pkg{sandwich}.
3. \code{vcov.} is set to \code{NULL}, then it is assumed that
a \code{vcov} method exists, such that \code{vcov(x)} yields
a covariance matrix. Illustrations are provided in the examples below.
The degrees of freedom \code{df} determine whether a normal
approximation is used or a t distribution with \code{df} degrees
of freedom. The default method computes \code{df.residual(x)}
and if this is \code{NULL}, \code{0}, or \code{Inf} a z test is performed.
The method for \code{"glm"} objects always uses \code{df = Inf} (i.e., a z test).
The corresponding Wald confidence intervals can be computed either
by applying \code{coefci} to the original model or \code{\link[stats]{confint}}
to the output of \code{coeftest}. See below for examples.
Finally, \code{\link[stats]{nobs}} and \code{\link[stats]{logLik}}
methods are provided which work, provided that there are such methods
for the original object \code{x}. In that case, \code{"nobs"} and
\code{"logLik"} attributes are stored in the \code{coeftest} output
so that they can be still queried subsequently. If both methods are
available, \code{\link[stats]{AIC}} and \code{\link[stats]{BIC}}
can also be applied.
}
\value{
\code{coeftest} returns an object of class \code{"coeftest"} which
is essentially a coefficient matrix with columns containing the
estimates, associated standard errors, test statistics and p values.
Attributes for a \code{"method"} label, and the \code{"df"} are
added along with \code{"nobs"} and \code{"logLik"} (provided that
suitable extractor methods \code{\link[stats]{nobs}} and
\code{\link[stats]{logLik}} are available). Optionally, the full
object \code{x} can be \code{save}d in an attribute \code{"object"}
to facilitate further model summaries based on the \code{coeftest}
result.
\code{coefci} returns a matrix (or vector) with columns giving
lower and upper confidence limits for each parameter. These will
be labeled as (1-level)/2 and 1 - (1-level)/2 in percent.
}
\seealso{\code{\link{lm}}, \code{\link{waldtest}}}
\examples{
## load data and fit model
data("Mandible", package = "lmtest")
fm <- lm(length ~ age, data = Mandible, subset=(age <= 28))
## the following commands lead to the same tests:
summary(fm)
(ct <- coeftest(fm))
## a z test (instead of a t test) can be performed by
coeftest(fm, df = Inf)
## corresponding confidence intervals
confint(ct)
coefci(fm)
## which in this simple case is equivalent to
confint(fm)
## extract further model information either from
## the original model or from the coeftest output
nobs(fm)
nobs(ct)
logLik(fm)
logLik(ct)
AIC(fm, ct)
BIC(fm, ct)
if(require("sandwich")) {
## a different covariance matrix can be also used:
(ct <- coeftest(fm, df = Inf, vcov = vcovHC))
## the corresponding confidence interval can be computed either as
confint(ct)
## or based on the original model
coefci(fm, df = Inf, vcov = vcovHC)
## note that the degrees of freedom _actually used_ can be extracted
df.residual(ct)
## which differ here from
df.residual(fm)
## vcov can also be supplied as a function with additional arguments
coeftest(fm, df = Inf, vcov = vcovHC, type = "HC0")
## or as a matrix
coeftest(fm, df = Inf, vcov = vcovHC(fm, type = "HC0"))
}
}
\keyword{htest}
|
