pairw.oneway {asbio}

Welch tests controlled for simultaneous inference

Description

Conducts all possible pairwise Welch tests with adjustments to P-values using methods from p.adjust

Usage

pairw.oneway(y, x, conf = 0.95, digits = 5, method = "holm")

Arguments

y

Response variable

x

Explanatory variable

conf

Confidence level

digits

Number of digits in results

method

Generalized method for controlling family wise type one error. These must be methods from p.adjust, i.e., "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none". Names can be abbreviated.

Value

The function pairw.oneway and the confidence interval functions it calls return a list of class = "pairw".

1) the type of contrast (names are taken from levels in x),

2) the mean difference,

3) the lower confidence bound of the true mean difference,

4) the upper confidence bound of the true mean difference,

5) the hypothesis decision, given the prescribed significance level, and

6) the adjusted P-value.

Other invisible objects include:

cont

a vector of contrasts.

conf

The confidence level.

band

A two column matrix containing the lower and upper confidence bounds.

Note

Note that while P-values will be adjusted for simultaneous inference (unless method = "none"), confidence interval width are generally not adjusted. In particular, CI widths correspond to Welch SEs and Satterthwaite t degrees of freedoms Thus they control for heteroscedasticity, however they do not control for family-wise levels of \alpha unless method = "bonferroni", under which the restrictive confidence level 1 - (\alpha/2r) is used, where r is the number of comparisons.

Author(s)

Ken Aho and Peter Eckert

References

Kutner, M. H., Nachtsheim, C. J., Neter, J., and Li., W (2005) Applied Linear Statistical Models, 5th edition. McGraw-Hill, Boston.

See Also

p.adjust, pairw.anova

Examples

y <- rnorm(30)
x <- as.factor(c(rep(1,10), rep(2,10), rep(3, 10)))
p <- pairw.oneway(y,x)
p
plot(p)