The R Project SVN R

Rev

Rev 90146 | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 90146 Rev 90152
Line 25... Line 25...
25 
    You can specify just the initial letter.}
 25 
    You can specify just the initial letter.}
26 
  \item{conf.level}{confidence level for the returned confidence
 26 
  \item{conf.level}{confidence level for the returned confidence
27 
    interval.}
 27 
    interval.}
28 
  \item{two.sided.method}{a character string specifying the method for
 28 
  \item{two.sided.method}{a character string specifying the method for
29 
    computing two-sided p-values, must be one of \code{"minlike"}
 29 
    computing two-sided p-values, must be one of \code{"minlike"}
30 
    (default) or \code{"central"}, or an abbrevation thereof.
 30 
    (default) or \code{"central"}, or an abbreviation thereof.
31 
    See \sQuote{Details}.}
 31 
    See \sQuote{Details}.}
32 
}
 32 
}
33 
\details{
 33 
\details{
34 
  Confidence intervals are obtained by a procedure first given in
 34 
  Confidence intervals are obtained by a procedure first given in
35 
  \bibcitet{R:Clopper+Pearson:1934}.
 35 
  \bibcitet{R:Clopper+Pearson:1934}.
Line 44... Line 44...
44 
  two-sided alternative that \eqn{\theta \ne \theta_0}, the
44 
  two-sided alternative that \eqn{\theta \ne \theta_0}, the
45 
  \emph{central} p-value \bibcitep{R:Fay:2010} is
 45 
  \emph{central} p-value \bibcitep{R:Fay:2010} is
46 
  \deqn{\min(2 \min(P(T \le t | \theta_0), P(T \ge t | \theta_0)), 1)}
 46 
  \deqn{\min(2 \min(P(T \le t | \theta_0), P(T \ge t | \theta_0)), 1)}
47 
  (this is called the twice the smaller tail method in
 47 
  (this is called the twice the smaller tail method in
48 
  \bibcitet{R:Hirji:2005}).
 48 
  \bibcitet{R:Hirji:2005}).
49 
  The \emph{minlike} p-value is the probability of observing a value of
 49 
  The \emph{\I{minlike}} p-value is the probability of observing a value of
50 
  \eqn{T} not more likely than the one observed, i.e.,
50 
  \eqn{T} not more likely than the one observed, i.e.,
51 
  \deqn{\sum_{s: P(T = s | \theta_0) \le P(T = s | \theta_0)} P(T = s | \theta_0)}
 51 
  \deqn{\sum_{s: P(T = s | \theta_0) \le P(T = s | \theta_0)} P(T = s | \theta_0)}
52 
  (this is called the probability based method by Hirji).
 52 
  (this is called the probability based method by \I{Hirji}).
53
53
54 
  For the exact binomial test, \eqn{T} is the number of successes and
 54 
  For the exact binomial test, \eqn{T} is the number of successes and
55 
  \eqn{\theta = p} is the probability of success.
 55 
  \eqn{\theta = p} is the probability of success.
56 
 
 56 
 
57 
  Inverting the two-sided test which rejects when the central p-value is
 57 
  Inverting the two-sided test which rejects when the central p-value is
58 
  at most \eqn{\alpha} gives central confidence intervals with
 58 
  at most \eqn{\alpha} gives central confidence intervals with
59 
  confidence level at least \eqn{1 - \alpha} where both lower and upper
 59 
  confidence level at least \eqn{1 - \alpha} where both lower and upper
60 
  tails have probability at most \eqn{\alpha / 2}.  For the exact
 60 
  tails have probability at most \eqn{\alpha / 2}.  For the exact
61 
  binomial test, these are the Clopper-Pearson intervals.
 61 
  binomial test, these are the \I{Clopper}-\I{Pearson} intervals.
62 
 
 62 
 
63 
  It is intended to change the default method for computing two-sided
 63 
  It is intended to change the default method for computing two-sided
64 
  p-values from minlike to central to make these match the confidence
 64 
  p-values from \I{minlike} to central to make these match the confidence
65 
  intervals.
 65 
  intervals.
66 
}
 66 
}
67 
\value{
 67 
\value{
68 
  A list with class \code{"htest"} containing the following components:
 68 
  A list with class \code{"htest"} containing the following components:
69 
  \item{statistic}{the number of successes.}
 69 
  \item{statistic}{the number of successes.}