The R Project SVN R

Rev

Rev 90153 | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 90153 Rev 90259
Line 52... Line 52...
52 
  distributions with continuous distribution functions \eqn{F((t-m)/s)}
 52 
  distributions with continuous distribution functions \eqn{F((t-m)/s)}
53 
  and \eqn{F(t-m)} respectively, where \eqn{F(0) = 1/2}, \eqn{m} is an
 53 
  and \eqn{F(t-m)} respectively, where \eqn{F(0) = 1/2}, \eqn{m} is an
54 
  unknown nuisance parameter and
 54 
  unknown nuisance parameter and
55 
  \eqn{s}, the ratio of scales, is the parameter of interest.  The
 55 
  \eqn{s}, the ratio of scales, is the parameter of interest.  The
56 
  \I{Ansari}-\I{Bradley} test is used for testing the null that \eqn{s} equals
 56 
  \I{Ansari}-\I{Bradley} test is used for testing the null that \eqn{s} equals
57 
  1, the two-sided alternative being that \eqn{s \ne 1}{s != 1} (the
 57 
  1, the two-sided alternative being that \eqn{s \ne 1} (the
58 
  distributions differ only in variance), and the one-sided alternatives
 58 
  distributions differ only in variance), and the one-sided alternatives
59 
  being \eqn{s > 1} (the distribution underlying \code{x} has a larger
 59 
  being \eqn{s > 1} (the distribution underlying \code{x} has a larger
60 
  variance, \code{"greater"}) or \eqn{s < 1} (\code{"less"}).
 60 
  variance, \code{"greater"}) or \eqn{s < 1} (\code{"less"}).
61 
 
 61 
 
62 
  By default (if \code{exact} is not specified), an exact p-value
 62 
  By default (if \code{exact} is not specified), an exact p-value