Skip to contents

Funnel plot for a RoBMA object

Source: R/plots.R
funnel.Rd

funnel creates a funnel plot for a "RoBMA" object. Only available for normal-normal models estimated using the spike-and-slab algorithm (i.e., algorithm = "ss"). This function uses several simplifications to visualize the sampling distribution, see Details for more information.

Usage

funnel(
  x,
  conditional = FALSE,
  plot_type = "base",
  output_scale = NULL,
  incorporate_heterogeneity = TRUE,
  incorporate_publication_bias = TRUE,
  ...
)

Arguments

x

a fitted RoBMA object

conditional

whether conditional estimates should be plotted. Defaults to FALSE which plots the model-averaged estimates. Note that both "weightfunction" and "PET-PEESE" are always ignoring the other type of publication bias adjustment.

plot_type

whether to use a base plot "base" or ggplot2 "ggplot" for plotting. Defaults to "base".

output_scale

transform the effect sizes and the meta-analytic effect size estimate to a different scale. Defaults to NULL which returns the same scale as the model was estimated on.

incorporate_heterogeneity

Whether heterogeneity should be incorporated into the sampling distribution. Defaults to TRUE.

incorporate_publication_bias

Whether publication bias should be incorporated into the sampling distribution. Defaults to TRUE.

...

list of additional graphical arguments to be passed to the plotting function. Supported arguments are lwd, lty, col, col.fill, xlab, ylab, main, xlim, ylim to adjust the line thickness, line type, line color, fill color, x-label, y-label, title, x-axis range, and y-axis range respectively.

Details

The funnel function differs from the corresponding funnel function in two regards; 1) The heterogeneity is by default incorporated into the model and 2) the sampling distribution under publication bias is approximate by sampling from the estimated weighted normal distribution under the pooled effect. This approximation may distort the true sampling distribution (but that would be impossible) to visualize in the usual funnel plot style anyway?.

The sampling distribution is drawn under the mean effect size and heterogeneity estimates (the uncertainty about those values is not incorporated into the figure).