| Title: | Ryan-Holm Step-Down Bonferroni or Sidak Procedure |
|---|---|
| Description: | The Ryan-Holm step-down Bonferroni or Sidak procedure is to control the family-wise (experiment-wise) type I error rate in the multiple comparisons. This procedure provides the adjusting p-values and adjusting CIs. The methods used in this package are referenced from John Ludbrook (2000) <doi:10.1046/j.1440-1681.2000.03223.x>. |
| Authors: | Zhicheng Du Developer [aut, cre, cph], Hailin Feng Developer [aut] |
| Maintainer: | Zhicheng Du Developer <[email protected]> |
| License: | GPL-3 |
| Version: | 0.2.0 |
| Built: | 2025-02-11 03:56:13 UTC |
| Source: | https://github.com/cran/RHSDB |
This procedure provides the adjusting p-values and adjusting CIs.
rh.sd.bonferroni(p,effect,effect.se,df,type,sig,side,digits)rh.sd.bonferroni(p,effect,effect.se,df,type,sig,side,digits)
p |
the raw p values |
effect |
the effect size from the multiple comparisons, e.g. the mean difference from t test or paried t test |
effect.se |
the standard error of effect size from the multiple comparisons, e.g. the standard error of mean difference from t test or paried t test) |
df |
the degree of freedom of hypothesis test, e.g. n1+n2-2 for t test, n-1 for paried t test |
type |
the type of the effect size, default is "mean" |
sig |
the significance level, default is 0.05 |
side |
"one" or "two" sided hypothesis test |
digits |
the number of decimal digits |
p.adj |
the adjusted p value |
ci.adj.l |
the lower limit of adjusted confidence interval |
ci.adj.u |
the upper limit of adjusted confidence interval |
Please feel free to contact us, if you have any advice and find any bug!
Reference:
1. John Ludbrook (2000). MULTIPLE INFERENCES USING CONFIDENCE INTERVALS. Clinical and Experimental Pharmacology and Physiology. 27: 212-215.
Update:
Version 0.1.0: The first version.
Version 0.2.0: Fix the bug for maintaining monotonicity of the ranking p-values.
p=c(0.217,0.00028,0,0.001,0.024,0.719,0.00033) effect=c(16,74,-85,-38,29,5,91) effect.se=c(12,16,14,9,12,16,20) df=16 rh.sd.bonferroni(p,effect,effect.se,df)p=c(0.217,0.00028,0,0.001,0.024,0.719,0.00033) effect=c(16,74,-85,-38,29,5,91) effect.se=c(12,16,14,9,12,16,20) df=16 rh.sd.bonferroni(p,effect,effect.se,df)
This procedure provides the adjusting p-values and adjusting CIs.
rh.sd.sidak(p,effect,effect.se,df,type,sig,side,digits)rh.sd.sidak(p,effect,effect.se,df,type,sig,side,digits)
p |
the raw p values |
effect |
the effect size from the multiple comparisons, e.g. the mean difference from t test or paried t test |
effect.se |
the standard error of effect size from the multiple comparisons, e.g. the standard error of mean difference from t test or paried t test) |
df |
the degree of freedom of hypothesis test, e.g. n1+n2-2 for t test, n-1 for paried t test |
type |
the type of the effect size, default is "mean" |
sig |
the significance level, default is 0.05 |
side |
"one" or "two" sided hypothesis test |
digits |
the number of decimal digits |
p.adj |
the adjusted p value |
ci.adj.l |
the lower limit of adjusted confidence interval |
ci.adj.u |
the upper limit of adjusted confidence interval |
Please feel free to contact us, if you have any advice and find any bug!
Reference:
1. John Ludbrook (2000). MULTIPLE INFERENCES USING CONFIDENCE INTERVALS. Clinical and Experimental Pharmacology and Physiology. 27: 212-215.
Update:
Version 0.1.0: The first version.
Version 0.2.0: Fix the bug for maintaining monotonicity of the ranking p-values.
p=c(0.217,0.00028,0,0.001,0.024,0.719,0.00033) effect=c(16,74,-85,-38,29,5,91) effect.se=c(12,16,14,9,12,16,20) df=16 rh.sd.sidak(p,effect,effect.se,df)p=c(0.217,0.00028,0,0.001,0.024,0.719,0.00033) effect=c(16,74,-85,-38,29,5,91) effect.se=c(12,16,14,9,12,16,20) df=16 rh.sd.sidak(p,effect,effect.se,df)