Skip to contents

Heterogeneity Forest Plot

heterogeneity_forest_plot.Rd

Accepts a data frame of subgroup estimates and produces a forest plot suitable for publication. Subgroups can be sorted by effect size. An optional overall estimate is shown as a diamond.

Usage

heterogeneity_forest_plot(
  estimates,
  subgroup = NULL,
  overall = NULL,
  xlab = "Estimated Effect",
  title = "Subgroup Analysis",
  ref_line = 0,
  sort_by = c("estimate", "none")
)

Arguments

estimates

A data frame with at minimum these columns:

label

Character. Subgroup label.

estimate

Numeric. Point estimate.

ci_lower

Numeric. Lower confidence interval bound.

ci_upper

Numeric. Upper confidence interval bound.

Optional columns: n (sample size, shown as text), p_value (used for significance colouring).

subgroup

Character or NULL. Column name in estimates to use as a grouping factor for colouring. Default NULL.

overall

A single-row data frame with columns estimate, ci_lower, ci_upper for the overall effect diamond. NULL to omit.

xlab

Character. x-axis label. Default "Estimated Effect".

title

Character. Plot title. Default "Subgroup Analysis".

ref_line

Numeric. x-intercept for the reference line. Default 0.

sort_by

Character. "estimate" to sort subgroups by effect size, "none" to preserve input order. Default "estimate".

Value

A ggplot2 object.

Details

Publication-Ready Forest Plot for Subgroup / CATE Analysis

Creates a publication-ready forest plot visualising treatment effect heterogeneity across subgroups or conditional average treatment effects (CATE). Confidence intervals are shown as horizontal whiskers, the point estimate as a filled dot colour-coded by statistical significance, and an optional overall-effect diamond is drawn at the bottom.

Examples

est_df <- data.frame(
  label    = c("Female", "Male", "Age < 40", "Age >= 40",
               "Low income", "High income"),
  estimate = c(0.18, 0.09, 0.22, 0.07, 0.25, 0.05),
  ci_lower = c(0.08, -0.01, 0.11, -0.04, 0.13, -0.06),
  ci_upper = c(0.28, 0.19, 0.33, 0.18, 0.37, 0.16),
  n        = c(150, 150, 140, 160, 130, 170),
  p_value  = c(0.001, 0.08, 0.001, 0.26, 0.001, 0.40),
  stringsAsFactors = FALSE
)

overall_df <- data.frame(estimate = 0.13, ci_lower = 0.07, ci_upper = 0.19)

heterogeneity_forest_plot(
  estimates = est_df,
  overall   = overall_df,
  title     = "Subgroup Treatment Effects",
  sort_by   = "estimate"
)