esci (version 1.0.7)

plot_meta: Generates a forest plot displaying results of a meta-analysis

Description

`plot_meta' returns a ggplot2 object visualizing the results of a meta-analysis, showing each study effect size and CI, the overall effect size and CI as a diamond, effect sizes estimated at each moderator level (if defined), and (optionally) prediction intervals for subsequent studies. This function requires as input an esci_estimate object generated by an esci meta-analysis function: meta_any(), meta_d1(), meta_d2(), meta_mdiff_two(), meta_mean(), meta_pdiff_two(), meta_proportion(), and meta_r().

Usage

plot_meta(
  estimate,
  mark_zero = TRUE,
  include_PIs = FALSE,
  report_CIs = FALSE,
  explain_DR = FALSE,
  meta_diamond_height = 0.35,
  ggtheme = ggplot2::theme_classic()
)

Value

Returns a ggplot object

Arguments

estimate

• an esci_estimate object generated by an esci meta_ function

mark_zero

• Boolean; defaults to TRUE to include a dotted line indicated no effect (effect_size = 0)

include_PIs

• Boolean; defaults to FALSE; set to TRUE to include prediction intervals for the overall effect and each moderator level (if defined)

report_CIs

• Boolean; defaults to FALSE; set to TRUE to include printed representation of each study effect size and CI along the right- hand of the figure

explain_DR

• Boolean; defaults to FALSE; set to TRUE if no moderator is defined to show both the RE and FE effect sizes to represent how the diamond ration measure of effect-size heterogeneity is calculated

meta_diamond_height

• Optional real number > 0 to indicate height that each meta-analytic diamond should be drawn; defaults to 0.35

ggtheme

• Optional ggplot2 theme object to control overall styling; defaults to ggplot2::theme_classic()

Details

This function was developed primarily for student use within jamovi when learning along with the text book Introduction to the New Statistics, 2nd edition (Cumming & Calin-Jageman, 2024).

Expect breaking changes as this function is improved for general use. Work still do be done includes:

• Revise to avoid deprecated ggplot features

• Revise for consistent ability to control aesthetics and consistent layer names

Examples

# Data set -- see Introduction to the New Statistics, 2nd edition
data("data_mccabemichael_brain")

# Meta-analysis: random effects, no moderator
estimate <- esci::meta_mdiff_two(
  data = esci::data_mccabemichael_brain,
  comparison_means = "M Brain",
  comparison_sds = "s Brain",
  comparison_ns = "n Brain",
  reference_means = "M No Brain",
  reference_sds = "s No Brain",
  reference_ns = "n No Brain",
  labels = "Study name",
  effect_label = "Brain Photo Rating - No Brain Photo Rating",
  assume_equal_variance = TRUE,
  random_effects = TRUE
)
# Forest plot
myplot_forest <- esci::plot_meta(estimate)


# Meta-analysis: random effects, moderator
estimate_moderator <- esci::meta_mdiff_two(
  data = esci::data_mccabemichael_brain,
  comparison_means = "M Brain",
  comparison_sds = "s Brain",
  comparison_ns = "n Brain",
  reference_means = "M No Brain",
  reference_sds = "s No Brain",
  reference_ns = "n No Brain",
  labels = "Study name",
  moderator = "Research group",
  effect_label = "Brain Photo Rating - No Brain Photo Rating",
  assume_equal_variance = TRUE,
  random_effects = TRUE
)
# Forest plot
myplot_forest_moderator <- esci::plot_meta(estimate_moderator)


# Meta-analysis: random effects, moderator, output d_s
estimate_moderator_d <- esci::meta_mdiff_two(
  data = esci::data_mccabemichael_brain,
  comparison_means = "M Brain",
  comparison_sds = "s Brain",
  comparison_ns = "n Brain",
  reference_means = "M No Brain",
  reference_sds = "s No Brain",
  reference_ns = "n No Brain",
  labels = "Study name",
  moderator = "Research group",
  effect_label = "Brain Photo Rating - No Brain Photo Rating",
  assume_equal_variance = TRUE,
  random_effects = TRUE
)
# Forest plot
myplot_forest_moderator_d <- esci::plot_meta(estimate_moderator_d)