Mediation Figure

Plot simple mediation figures, maybe displaying indirect effects and confidence intervals of some sort, and potentially saving them.

Usage

medfig(mat, ci = NULL, ci.type = NULL, xlab = "x variable",
  ylab = "y variable", mlab = "mediating variable", title = FALSE,
  note = TRUE, box = TRUE, digits = 3, note.sig = TRUE,
  fin.length = 0.5, cords.box = list(), cords.arrow = list(),
  padding.box = 0.05, padding.arrow = 0.02, padding.number = 0.02,
  padding.title = 0.1, padding.note = 0.1, arrow.scale = 0.01,
  box.fill = NA, center.adj = 0.1, color = c(line = NA, box = NA, label =
  NA, number = NA, title = NA, note = NA), weight = c(line = NA, box = NA,
  label = NA, number = NA, title = 2, note = 0.9), save = FALSE,
  format = pdf, name = "mediation", ...)

Arguments

mat: Output from bspmed, or a matrix with two rows and 4-5 columns (see example).
ci: A vector of confidence intervals (one lower and one upper). If the entries are named, these will be read as numbers and added together for the displayed percent (see example).
ci.type: Text to be displayed in the note (i.e., ci.type confidence intervals).
xlab, ylab, mlab: Text to be displayed in each variable box.
title, note, box: Logical indicating whether to draw each element.
digits: Number of digits to round to.
note.sig: Logical indicating whether asterisks corresponding to p-values should be added to numbers.
fin.length: Length of the arrowhead fins, back from the line intersection (0 would be a flat base, for a triangular arrowhead).
cords.box: Center position of each box. Should be a list with at least one coordinate pair (a vector with values for y and x position). If entries aren't named, these x, y, and m are assigned as names; coordinates pairs can be entered positionally or named. Default is list(x = c(.1, .1), m = c(.5, .9), y = c(.9, .1)).
cords.arrow: Position of each arrow. This behaves in the same way as cords.box, but entries are names a, b, and c, and each set of coordinates should be 4 values specifying start and end positions. Any unspecified coordinates are calculated based on box positions.
padding.box, padding.arrow, padding.number, padding.title, padding.note: Specifies the padding around each of the elements; box adjust space around texts in each box, arrow adjust distance from boxes, and number adjusts space between numbers and lines.
arrow.scale: Size of each arrowhead.
box.fill: Color of the box backgrounds.
center.adj: Amount to shift the center text downward (positive values making center text closer to the bottom).
color: A vector with named entries corresponding to any of the elements (line, box, label, number, title, note). Box refers to the color of the line (whereas box.fill specifies the background). Single unnamed values are applied to all elements.
weight: The weight of arrow lines or box outlines, or size of numbers or texts. Behaves the same as color.
save, format, name: Information used when saving the figure to a file; if any are specified, an image is saved. Format refers to the graphics engine; default is pdf. Some elements may not show up properly using other engines (such as cairo_pdf). Other engines are included in the base link[grDevices]{grDevices} package. Another option is emf from the devEMF package (used to create the example image).
...: Additional arguments passed to par (parameters of the plot frame).

Examples

# by default, the bspmed function will show a figure
bspmed(disp, mpg, wt, scale(mtcars))
#> Effect of disp on mpg, by way of wt 
#> Processed in 0.99 seconds
#> 
#> Sobel
#>       a       b       c      c' indirect
#> b 0.888 -0.5440 -0.8476 -0.3645  -0.4831
#> p 0.000  0.0074  0.0000  0.0636   0.0055
#> 
#> Permutation confidence intervals
#>         2.5%      97.5% all <|> -0.4831
#> CI -1.132512 -0.3899162           FALSE


# you can also save the output from a bspmed call and plot a figure for it later:
med <- bspmed(disp, mpg, wt, scale(mtcars), figure = FALSE)
#> Effect of disp on mpg, by way of wt 
#> Processed in 0.99 seconds
#> 
#> Sobel
#>       a       b       c      c' indirect
#> b 0.888 -0.5440 -0.8476 -0.3645  -0.4831
#> p 0.000  0.0074  0.0000  0.0636   0.0055
#> 
#> Permutation confidence intervals
#>         2.5%      97.5% all <|> -0.4831
#> CI -1.173451 -0.3800334           FALSE
medfig(med)


# Or you can enter a matrix, with confidence intervals in the second position,
# and what type they are in the third.
mat <- matrix(
  c(.89, .0001, -.54, .001, -.85, .0001, -.36, .05, -.48, .001),
  2,
  dimnames = list(c("b", "p"), c("a", "b", "c", "c'", "i"))
)
medfig(mat, c("2.5" = -.38, "97.5" = .37), "Permutation")

Usage

Arguments

See also

Examples