Plot a (standardized) major axis fit, including the data and the fitted lines. There are many options for changing the appearance of the plot and generating high-quality publishable graphs.
# S3 method for sma
plot(
x,
which = c("default", "residual", "qq"),
use.null = FALSE,
add = FALSE,
type = "o",
xaxis = NULL,
yaxis = NULL,
xlab = NULL,
ylab = NULL,
pch = NULL,
col = NULL,
lty = NULL,
from = NULL,
to = NULL,
log = x$log,
frame.plot = TRUE,
tck = par("tck"),
p.lines.transparent = NA,
axes = TRUE,
...
)Object of class 'sma'.
If 'residual', plots a residual plot; if 'qq', plots a qq plot; otherwise an x-y plot.
Logical. If FALSE, plots the fitted lines (the default), otherwise those corresponding to the null hypothesis.
Logical. If TRUE, adds lines or points to an existing plot.
As in 'lm.default' : 'p' plots points only, 'l' only lines, and 'o' or 'b' plot both.
Special axis objects. See Details and examples.
Labels for X and Y axes.
Plotting symbols (see points).
Color of points and lines.
Line type (see lines).
Min and max X values for the lines (defaults to values given
by sma, which are the X values corresponding the maximum and
minimum fitted values in the data.).
One of 'x','y' or 'xy', to denote which axes to log-scale.
a logical indicating whether a box should be drawn around the plot, by default = TRUE.
The length of tick marks as a fraction of the smaller of the width or height of the plotting region. If tck >= 0.5 it is interpreted as a fraction of the relevant side, so if tck = 1 grid lines are drawn. By default set to current system defaults (tck = par("tck")).
Adjusts transparency level of fitted lines according to p-value of correlation between X and Y, via formula opacity = 1-p/p.lines.transparent). Setting to a value 0.1 means the line for any group with p = 0.1 would be fully transparent, while line for a group with p =0.05 would be 50 perecent transparent. by default set to NA, which means lines are fully visible.
If FALSE, suppress plotting of the axes (Default TRUE)
Further arguments passed to plot.default.
The plot.sma function produces one of three different types of plot,
depending on the which argument.
The default plot, which="default", produced a plot of y vs
x, with separate symbols for each group if appropriate, and MA
or SMA lines fitted through each group. The formula used in the sma
object that is input to the plot function determines whether there is
a group structure, whether fitted lines have common slope, etc.
A residual plot can be obtained via which="residual" - this is a plot
of residuals against fitted values. This can be used to check assumptions -
if assumptions are satisfied there should be no pattern.
A normal quantile plot can be obtained via which="qq" - this is a
normal quantile plot of residuals. This can be used to check the normality
assumption - if data are close to a straight line, normality is plausible.
Note that non-normality is only important to the validity of the test in
small samples. In larger samples, non-normality will not effect validity,
but strong non-normality will reduce the power of tests.
If use.null=TRUE then the lines added to the plot use the
coefficients estimated under the null hypothesis. For example, if the sma
object x was produced using a common slopes test (via
y~x*groups or similar) then use.null=TRUE will plot lines that
apply the common slope to each group.
The arguments pch, col, lty, from & to, are used
to modify characteristics of the plotted points and lines. If a vector of
values for anyone of these arguments is passed to plot.sma, then
successive values are applied to each group, provided group structure is
included in x and the vector length is at least as long as the number
of groups.
By default, plot.sma uses the default tick spacing given by
plot.default. To customise axes, users can pass special axis objects
to plot.sma, obtained using the defineAxis command as
in the example below. This enables high quality publishable plots to be
produced. See utils for more information.
if (FALSE) {
# Load leaf lifetime dataset:
data(leaflife)
# Only consider low-nutrient sites:
leaf.low.soilp <- subset(leaflife, soilp == 'low')
# Fit SMA's separately at each of high and low
# rainfall sites and test for common slope:
ft <- sma(longev~lma*rain, data=leaf.low.soilp, log="xy")
# Plot leaf longevity (longev) vs leaf mass per area (lma)
# separately for each of high and low rainfall:
plot(ft)
# As before but add lines which have a common slope:
plot(ft, use.null=TRUE)
#As above, but adding the common slope lines to an existing plot
plot(ft, type='p', col="black")
plot(ft, use.null=TRUE, add=TRUE, type='l')
# Plot with equally spaced tick marks:
plot(ft, xaxis=defineAxis(major.ticks=c(40,80,160,320,640)),
yaxis=defineAxis(major.ticks=c(0.5,1,2,4,8)) )
# Produce a residual plot to check assumptions:
plot(ft,which="res")
# Produce a normal quantile plot:
plot(ft,which="qq")
}