pvalue.ellipse: Center and Radius for p-value Ellipses

In byandell/pda: Practial Data Analysis for Designed Experiments

Description

This is used by int.plot to construct ellipses for means at the level of the x factor that are not significantly different. Currently this does not handle mixed models properly.

Usage

pvalue.ellipse(plet, lsmeans, se, rdf,
   level=0.05, scale=0.3, full.width=TRUE, pset)

Arguments

plet	character vector with letter codes for non-significant differences
lsmeans	least squares means
se	standard errors for means
rdf	residual degrees of freedom
level	significance level
scale	scaling of ellipse around means
full.width	keep ellipses as full width to show how close means are if TRUE
pset	letters to use for means that are significantly different

Details

The ellipse will be the height of an LSD comparison interval if full.width is TRUE, or otherwise will be se*scale beyond the outer means of a comparison group.

Value

A 2*n matrix, where n is the number of distinct ellipses. The first row contains ellipse centers while the second row contains the radii.

Author(s)

Brian S. Yandell

See Also

lsmean, pletter, int.plot, ellipse

Examples

example( Oocyte )
int.plot( Oocyte.fit, bar.plot = "ellipse" )

byandell/pda documentation built on May 13, 2019, 9:27 a.m.