The refplot
function constructs a dot plot of effects along with a rweference distribution (either normal or simulated) to help in visually identifying active effects in a screening experiment.
refplot(effects, ref = TRUE, half = TRUE, method = "Zahn",
col = half, guides = FALSE, ID = FALSE, pch = 16, xlab, xlim, ...)
An environment that can be modified using dot.id
or dot.mod
.
Numeric vector of effects or contrasts to be explored.
Logical or character value. If TRUE
, a reference line or curve is added to the plot determined by method
. If FALSE
, no reference is added to the plot.
A character value matching "normal"
or "simulated"
is also permitted.
"normal"
is equivalent to ref = TRUE
. With "simulated"
,
, a kernel density estimate is displayed; it is obtained by scaling the simulated reference distribution for "method"
by its observed PSE.
Logical value. If TRUE
, a dot plot of the absolute effects is constructed. If FALSE
, the original signed effects are plotted.
Character value. When ref
is not false, the method to use in determining the PSE for scaling the reference curve (and also the method used for simulating the referemnce distribution when ref = "simulated"
). This must be the name of a provided pseudo-standard-error method (see PSE
), or a compatible user-supplied one.
Scalar or vector of colors; or a logical value. If logical, a value of TRUE
colors the positive effects blue, the negative effects red, and any zeros as black. A logical value of FALSE
colors them all black.
Logical value. If TRUE
, dotted lines are added that illustrate guide lines that could be used to draw the normal curve by hand.
Logical value. If logical and TRUE
, then after the plot is contructed, the dot.id
invoked so that the user may click on points to be labeled on the plot. If a numeric value is given, it is used as a threshold by which all effects greater than ID[1]
in absolute value are labeled.
Additional graphical parameters (see par
) passed to dot.plot
.
Russell V. Lenth
If the returned environment is saved, then dot.id
or dot.mod
may be used later as for dot.plot
results.
Other ways of assessing active effects include a half-normal plot (hnplot
), a Pareto plot of effects (see parplot
), and a tabular style of presenting effects and \(P\) values (see eff.test
). For more information on PSEs and methods, see PSE
and ref.dist
.
require(unrepx)
refplot(pdEff, ID = ME(pdEff))
if (FALSE) {
# Batman lives!
refplot(pdEff, ref = "sim", method = "Lenth", half = FALSE)
}
Run the code above in your browser using DataLab