The R Project SVN – R – Diff – /trunk/src/library/stats/man/p.adjust.Rd



}
\arguments{
  \item{p}{numeric vector of p-values (possibly with
    \code{\link{NA}}s).  Any other \R object is coerced by
    \code{\link{as.numeric}}.}
  \item{method}{correction method, a \code{\link{character}} string.
    Can be abbreviated.}
  \item{n}{number of comparisons, must be at least \code{length(p)};
    only set this (to non-default) when you know what you are doing!}
}
\details{
  The adjustment methods include the Bonferroni correction

  are independent or when they are non-negatively associated (Sarkar,
  1998; Sarkar and Chang, 1997).  Hommel's method is more powerful than
  Hochberg's, but the difference is usually small and the Hochberg
  p-values are faster to compute.
 
  The \code{"BH"} (aka \code{"fdr"}) and \code{"BY"} methods of
  Benjamini, Hochberg, and Yekutieli control the false discovery rate,
  the expected proportion of false discoveries amongst the rejected
  hypotheses.  The false discovery rate is a less stringent condition
  than the family-wise error rate, so these methods are more powerful
  than the others.

The R Project SVN R

(root)/trunk/src/library/stats/man/p.adjust.Rd – Rev 75151 → 76710

Rev 75151		Rev 76710
Line 18...		Line 18...
18	`}`	18	`}`
19	`\arguments{`	19	`\arguments{`
20	`\item{p}{numeric vector of p-values (possibly with`	20	`\item{p}{numeric vector of p-values (possibly with`
21	`\code{\link{NA}}s). Any other \R object is coerced by`	21	`\code{\link{NA}}s). Any other \R object is coerced by`
22	`\code{\link{as.numeric}}.}`	22	`\code{\link{as.numeric}}.}`
23	`\item{method}{correction method. Can be abbreviated.}`	23	`\item{method}{correction method, a \code{\link{character}} string.`
-		24	`Can be abbreviated.}`
24	`\item{n}{number of comparisons, must be at least \code{length(p)};`	25	`\item{n}{number of comparisons, must be at least \code{length(p)};`
25	`only set this (to non-default) when you know what you are doing!}`	26	`only set this (to non-default) when you know what you are doing!}`
26	`}`	27	`}`
27	`\details{`	28	`\details{`
28	`The adjustment methods include the Bonferroni correction`	29	`The adjustment methods include the Bonferroni correction`
Line 46...		Line 47...
46	`are independent or when they are non-negatively associated (Sarkar,`	47	`are independent or when they are non-negatively associated (Sarkar,`
47	`1998; Sarkar and Chang, 1997). Hommel's method is more powerful than`	48	`1998; Sarkar and Chang, 1997). Hommel's method is more powerful than`
48	`Hochberg's, but the difference is usually small and the Hochberg`	49	`Hochberg's, but the difference is usually small and the Hochberg`
49	`p-values are faster to compute.`	50	`p-values are faster to compute.`
50		51
51	`The \code{"BH"} (aka \code{"fdr"}) and \code{"BY"} method of`	52	`The \code{"BH"} (aka \code{"fdr"}) and \code{"BY"} methods of`
52	`Benjamini, Hochberg, and Yekutieli control the false discovery rate,`	53	`Benjamini, Hochberg, and Yekutieli control the false discovery rate,`
53	`the expected proportion of false discoveries amongst the rejected`	54	`the expected proportion of false discoveries amongst the rejected`
54	`hypotheses. The false discovery rate is a less stringent condition`	55	`hypotheses. The false discovery rate is a less stringent condition`
55	`than the family-wise error rate, so these methods are more powerful`	56	`than the family-wise error rate, so these methods are more powerful`
56	`than the others.`	57	`than the others.`